Re: LCAS and GMPLS
Huub van Helvoort <hhelvoort@chello.nl> Sun, 31 July 2005 14:26 UTC
Received: from odin.ietf.org ([132.151.1.176] helo=ietf.org) by megatron.ietf.org with esmtp (Exim 4.32) id 1DzEmE-0006C7-Qc for ccamp-archive@megatron.ietf.org; Sun, 31 Jul 2005 10:26:54 -0400
Received: from ietf-mx.ietf.org (ietf-mx [132.151.6.1]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id KAA27630 for <ccamp-archive@ietf.org>; Sun, 31 Jul 2005 10:26:52 -0400 (EDT)
Received: from psg.com ([147.28.0.62] ident=mailnull) by ietf-mx.ietf.org with esmtp (Exim 4.43) id 1DzFI2-0004wL-Hd for ccamp-archive@ietf.org; Sun, 31 Jul 2005 11:00:00 -0400
Received: from majordom by psg.com with local (Exim 4.50 (FreeBSD)) id 1DzEhR-000NlO-3X for ccamp-data@psg.com; Sun, 31 Jul 2005 14:21:57 +0000
Received: from [213.46.243.17] (helo=amsfep12-int.chello.nl) by psg.com with esmtp (Exim 4.50 (FreeBSD)) id 1DzEhN-000Nky-SH for ccamp@ops.ietf.org; Sun, 31 Jul 2005 14:21:54 +0000
Received: from [192.168.1.4] (really [24.132.27.149]) by amsfep12-int.chello.nl (InterMail vM.6.01.04.04 201-2131-118-104-20050224) with ESMTP id <20050731142140.MHNB11566.amsfep12-int.chello.nl@[192.168.1.4]> for <ccamp@ops.ietf.org>; Sun, 31 Jul 2005 16:21:40 +0200
Message-ID: <42ECDE6F.5000903@chello.nl>
Date: Sun, 31 Jul 2005 16:21:35 +0200
From: Huub van Helvoort <hhelvoort@chello.nl>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax)
X-Accept-Language: en-us, en
MIME-Version: 1.0
To: ccamp <ccamp@ops.ietf.org>
Subject: Re: LCAS and GMPLS
Content-Type: text/plain; charset="us-ascii"; format="flowed"
Content-Transfer-Encoding: 7bit
X-Spam-Checker-Version: SpamAssassin 3.0.2 (2004-11-16) on psg.com
X-Spam-Status: No, score=-2.6 required=5.0 tests=BAYES_00 autolearn=ham version=3.0.2
Sender: owner-ccamp@ops.ietf.org
Precedence: bulk
X-Spam-Score: 0.1 (/)
X-Scan-Signature: 39bd8f8cbb76cae18b7e23f7cf6b2b9f
Content-Transfer-Encoding: 7bit
Hello Wataru, You wrote: > I forward a nice comment from Mr. Huub van Helvoort. Thank you for forwarding. (I hope I can send this message the list myself). > His comments indicates we need liason from ITU-T SG-15 regarding to > this issue. This is a good proposal, it will indicate to ITU-T SG15/Q11 that IETF/CCAMP supports this enhancement. Kind regards, Huub. --- from message 27-7-2005 --- >> >> One point for the problem statement could be that in order to >> guarantee a hitless removal of one of the members in a VCG >> (decrease of bandwidth) there is a mandatory sequence: >> First remove member at ingress node, wait for confirm and >> only then remove path and member at egress node. >> >> However, a solution was presented at the last ITU-T SG15/q11 >> meeting that removes this requirement. >> >> Cheers, Huub. -- ================================================================ http://members.chello.nl/hhelvoort/ ================================================================ Always remember that you are unique...just like everyone else... Envelope-to: ccamp-data@psg.com Delivery-date: Sun, 31 Jul 2005 14:22:51 +0000 Message-ID: <42ECDE6F.5000903@chello.nl> Date: Sun, 31 Jul 2005 16:21:35 +0200 From: Huub van Helvoort <hhelvoort@chello.nl> User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax) MIME-Version: 1.0 To: ccamp <ccamp@ops.ietf.org> Subject: Re: LCAS and GMPLS Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Hello Wataru, You wrote: > I forward a nice comment from Mr. Huub van Helvoort. Thank you for forwarding. (I hope I can send this message the list myself). > His comments indicates we need liason from ITU-T SG-15 regarding to > this issue. This is a good proposal, it will indicate to ITU-T SG15/Q11 that IETF/CCAMP supports this enhancement. Kind regards, Huub. --- from message 27-7-2005 --- >> >> One point for the problem statement could be that in order to >> guarantee a hitless removal of one of the members in a VCG >> (decrease of bandwidth) there is a mandatory sequence: >> First remove member at ingress node, wait for confirm and >> only then remove path and member at egress node. >> >> However, a solution was presented at the last ITU-T SG15/q11 >> meeting that removes this requirement. >> >> Cheers, Huub. -- ================================================================ http://members.chello.nl/hhelvoort/ ================================================================ Always remember that you are unique...just like everyone else... Envelope-to: ccamp-data@psg.com Delivery-date: Sun, 31 Jul 2005 11:17:31 +0000 Message-ID: <007e01c595c1$c0a47fc0$5e1cff56@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <ccamp@ops.ietf.org> Subject: Slides please Date: Sun, 31 Jul 2005 12:19:41 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi, Thanks to those who have already sent me slides. If the rest of you would like to send them, I will put them on line. Absolute deadline is 5pm Tuesday, Paris time. Thanks. Adrian Envelope-to: ccamp-data@psg.com Delivery-date: Sun, 31 Jul 2005 10:50:11 +0000 Message-ID: <006601c595bd$a6f4b620$5e1cff56@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: "Arthi Ayyangar" <arthi@juniper.net> Cc: "'Jean Philippe Vasseur'" <jvasseur@cisco.com>, <raymond.zhang@bt.infonet.com>, <ccamp@ops.ietf.org> Subject: Issue for CCAMP meeting [Was: Early chair review of draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt] Date: Sun, 31 Jul 2005 11:49:58 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi Arthi, [SNIPPED] > hop at least has IP reachability (via IGP or BGP). If the next-hop is > not reachable, then the path computation stops and the LSR sends back a > PathErr upstream. If the next-hop is reachable, then it finds an > ABR/ASBR to get to the next-hop. In the absence of an auto-discovery > mechanism, the ABR in the case of inter-area TE or the ASBR in the > next-hop AS in the case of inter-AS TE should be the loose next-hop in > the ERO and hence should be accessible via the TED, otherwise the path > computation for the inter-domain TE LSP will fail. > ## I would like you to make it VERY clear what you are doing here. > ## You are using the Routing Database to make TE routing decisions. > ## This may (or may not) be OK in people's minds > ## I will send separate mail to the CCAMP list because this is a BIG > ## DEAL in GMPLS networks. AA---> I noticed the mail that you sent out. Thanks! However, there have been no replies yet. But I think this ID does state that that if the next hop is not in the TED and does not have IP reachability, then such a computation would fail. Should we raise this point again either at the WG meeting or when we post the next rev ? This would be a good topic to close at the meeting. What you say here is OK: that is, if the hop is not in the TED *and* does not have IP reachability, then computation fails. Clearly, we also know what to do when the hop is in the TED (regardless of IP reachability). The issue we have to cover is what to do when the hop is *not* in the TED, but we do have IP reachability. What does this mean when we have a separate data and control plane? What does it mean when the data and control planes are shared (as in MPLS)? I think there may be some subtle operational issues hiding here and I feel a BCP coming on :-( If you could mention this in your slot, that would be valuable. If folks with opinions could rush to the mic (or the mailing list), that would also be helpful. Cheers, Adrian Envelope-to: ccamp-data@psg.com Delivery-date: Sat, 30 Jul 2005 09:32:29 +0000 From: "Adrian Farrel" <olddog@clara.co.uk> To: jeanlouis.leroux@francetelecom.com, Dimitri.Papadimitriou@alcatel.be Cc: arthi@juniper.net, ccamp@ops.ietf.org, jpv@cisco.com, adrian@olddog.co.uk Reply-To: adrian@olddog.co.uk Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Date: Sat, 30 Jul 2005 10:29:57 +0100 Mime-Version: 1.0 Content-Type: text/plain; format=flowed; charset="iso-8859-1" Content-Transfer-Encoding: 8bit Message-Id: <E1DynfJ-000MbM-K2@oceanus.uk.clara.net> Hi, Time for me to speak out about the omission of hitherto mandatory objects from signaling messages. I would MUCH prefer that you keep the objects (Upstream Label or Path, Label on Resv) and use a special label value (implicit null? some new reserved label value?) to mean "stitch this". For me this is more natural and is less of a significant change to the existing protocol and implementation. It also solves the bidirection problem. Opinions? Cheers, Adrian ----- Original Message ----- From: <Dimitri.Papadimitriou@alcatel.be> To: "LE ROUX Jean-Louis RD-CORE-LAN" <jeanlouis.leroux@francetelecom.com> Cc: "Arthi Ayyangar" <arthi@juniper.net>; <ccamp@ops.ietf.org>; <jpv@cisco.com> Sent: Friday, July 29, 2005 9:39 PM Subject: RE: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > > hi j-l, arthi > > > > > -In section 4.1.2 you partially describe bidirectional LSP > > stitching > > > procedure. You mention that an Upstream Label MUST NOT be > > allocated by > > > the end-to-end LSP on the LSP segment, which is OK. But > > then how the > > > LSP-Segment Egress will now that the end-to-end LSP is > > bidirectional? > > AA--------> Excellent point. > > > > > What about defining a flag in the Attributes Flags TLV of the > > > LSP_ATTRIBUTE object so as to indicate that the LSP is > > bidirectional? > > AA------> That would be a change to the processing that GMPLS > > nodes use > > AA------> to > > detect bidirectionality, isn't it ? Normally nodes look for > > the Upstream Label object to detect bidirectionality. > > > > So let us say that an LSP is bidirectional if a) an Upstream > > Label is present or b) no Upstream Label, but bit set in > > LSP_ATTRIBUTE or c) both However, reliance on an e2e > > attributes bit set by head end, means existing head ends will > > not be setting this bit, so that will be an issue (wrt compatibility). > > JLLR: I agree this may raise some backward compatibility issues. > This would require that head-ends be upgraded... > > > DP: the other issue is having two methods for indicating the same thing is also not advisable > > > > Could be nice if this signaling was just between the node > > doing the stitching and the end point of the LSP segment, > > since this is the hop that the bidirectionality information > > is lost. Let me think about this. > > JLLR: Yes, and what about defining a new MPLS label value, let say label 4 = "No label assigned", with a semantic distinct from label 3 semantic (=label pop). The UPSTREAM label object with this new label value would be included in the end-to-end Path message over the LSP-Segment... > > > DP: assuming an information has to be passed for the indicating this capability i do also think a method like passing an implicit indication in the upstream label value would be more appropriated (note that the initial issue is due to the fact one would allow for unidirectional e2e LSP making use of bidirectional LSP segments - see comment from J-L here below - and not maintaining a 1:1 relationship for the directionality between the e2e LSP and the LSP segment that can be retrieved with the procedure described in section 4.2.4 - and so the first question is it worth allowing this?) > > > > > Also the selection of the LSP segment in case of bidirectional LSP > > > should be detailed (e.g. If the end-to-end LSP is > > bidirecitonal then > > > the LSP-segment MUST be bidirectional. Also shall we allow that two > > > unidrectional end-to-end LSP use the same bidirectional LSP segment > > > (one in each direction)? > > AA----> Yes, this should be okay. IMO. > > > > > -At the end of section 4.2.5 you mention that LSP-Segment > > failure or > > > maintenance SHOULD be treated as a failure event for the > > end-to-end LSP. > > > I agree for LSP-Segment failure but not for LSP-Segment maintenance. > > > LSP-Segment maintenance should be treated as TE-link > > maintenance for > > > the end-to-end LSP, and procedures defined in GMPLS > > graceful TE-link > > > shutdown draft may be useful (Specific RSVP error code and TE-link > > > attribute)... > > AA---> Yes, I agree; we will mention this. Actually the graceful > > AA---> teardown > > sentence occurs few paragraphs before, but not in the right > > context. So we will clarify the above. > > > DP: i am not sure to understand the comment, the sentence speaks about deletion of the LSP segment due to failure/maintenance should be treated as a failure event for the e2e LSP, so i don't think the initial sentence was meant to translated the initial comment, anyway it would be of interest that you detail the error code > > DP: note also that section 4.2.5 indicates that for stitched LSPs Graceful deletion can be used - i perfectly agree - but the sequence should be detailed as part of this section > > > > > Hope this helps, > > AA---> Sure does. > > > > > > Thanks! > > > > -arthi > > > > > > > > > > > > -----Message d'origine----- > > > > De : owner-ccamp@ops.ietf.org > > > > [mailto:owner-ccamp@ops.ietf.org] De la part de > > > > Internet-Drafts@ietf.org Envoyé : vendredi 15 juillet > > 2005 21:50 À : > > > > i-d-announce@ietf.org Cc : ccamp@ops.ietf.org Objet : I-D > > > > ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > > > > > > > > A New Internet-Draft is available from the on-line > > Internet-Drafts > > > > directories. > > > > This draft is a work item of the Common Control and Measurement > > > > Plane Working Group of the IETF. > > > > > > > > Title: Label Switched Path Stitching with > > > > Generalized MPLS Traffic Engineering > > > > Author(s): A. Ayyangar, J. Vasseur > > > > Filename: draft-ietf-ccamp-lsp-stitching-01.txt > > > > Pages: 19 > > > > Date: 2005-7-15 > > > > > > > > In certain scenarios, there may be a need to combine together two > > > > different Generalized Multi-Protocol Label Switching > > (GMPLS) Label > > > > Switched Paths (LSPs) such that in the data plane, a > > single end-to- > > > > end (e2e) LSP is achieved and all traffic from one LSP > > is switched > > > > onto the other LSP. We will refer to this as "LSP > > stitching". > > > > This > > > > document covers cases where: a) the node performing > > the stitching > > > > does not require configuration of every LSP pair to be stitched > > > > together b) the node performing the stitching is not > > the egress of > > > > any of the LSPs c) LSP stitching not only results in > > an end-to-end > > > > LSP in the data plane, but there is also a > > corresponding end-to-end > > > > LSP (RSVP session) in the control plane. It might be > > possible to > > > > configure a GMPLS node to switch the traffic from an LSP for > > > > which it > > > > is the egress, to another LSP for which it is the > > ingress, without > > > > requiring any signaling or routing extensions whatsoever, > > > > completely > > > > transparent to other nodes. This will also result in > > LSP stitching > > > > in the data plane. However, this document does not cover this > > > > scenario of LSP stitching. > > > > > > > > A URL for this Internet-Draft is: > > > > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitc > > > hing-01.txt > > > > > > > > To remove yourself from the I-D Announcement list, send a > > message to > > > > i-d-announce-request@ietf.org with the word unsubscribe > > in the body > > > > of the message. > > > > You can also visit > > > > https://www1.ietf.org/mailman/listinfo/I-D-announce > > > > to change your subscription settings. > > > > > > > > > > > > Internet-Drafts are also available by anonymous FTP. > > Login with the > > > > username "anonymous" and a password of your e-mail address. After > > > > logging in, type "cd internet-drafts" and then > > > > "get draft-ietf-ccamp-lsp-stitching-01.txt". > > > > > > > > A list of Internet-Drafts directories can be found in > > > > http://www.ietf.org/shadow.html or > > > > ftp://ftp.ietf.org/ietf/1shadow-sites.txt > > > > > > > > > > > > Internet-Drafts can also be obtained by e-mail. > > > > > > > > Send a message to: > > > > mailserv@ietf.org. > > > > In the body type: > > > > "FILE /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt". > > > > > > > > NOTE:The mail server at ietf.org can return the document in > > > > MIME-encoded form by using the "mpack" utility. To use this > > > > feature, insert the command "ENCODING mime" before the "FILE" > > > > command. To decode the response(s), you will need "munpack" or > > > > a MIME-compliant mail reader. Different MIME-compliant mail > > > > readers > > > > exhibit different behavior, especially when dealing with > > > > "multipart" MIME messages (i.e. documents which have been split > > > > up into multiple messages), so check your local documentation on > > > > how to manipulate these messages. > > > > > > > > > > > > Below is the data which will enable a MIME compliant mail reader > > > > implementation to automatically retrieve the ASCII version of the > > > > Internet-Draft. > > > > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 29 Jul 2005 20:43:10 +0000 To: adrian@olddog.co.uk Cc: ccamp@ops.ietf.org, sob@harvard.edu, zinin@psg.com, fenner@research.att.com MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: Re: Proposed liaison to SG15 on RFC 4139 Date: Fri, 29 Jul 2005 22:42:18 +0200 Message-ID: <OF0DF74692.9D9F3A61-ONC125704D.0071BC36-C125704D.0071BCB3@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmhpIGFkcmlhbiAtIDwvRk9OVD48L1A+ PFA+Jm5ic3A7PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5pIHRoaW5rIHdv cnRoIGNvbnNpZGVyaW5nIHRoaXMgYWN0aW9uICZuYnNwOzwvRk9OVD48L1A+PFA+PEZPTlQgRkFD RT0iTW9ub3NwYWNlLENvdXJpZXIiPm5vdGU6IGRvIHlvdSB0aGluayB0aGlzIGxpYWlzb24gc2hv dWxkIGFsc28gaW5jbHVkZSBhbiBpbmZvcm1hdGlvbmFsIHN0YXRlbWVudCBjb25jZXJuaW5nIHRo ZSBzdWJzZXF1ZW50IGFjdGlvbnMgaW4gdGVybXMgb2Ygc2lnbmFsaW5nIHByb3RvY29sIHdvcmsg PzwvRk9OVD48L1A+PFA+Jm5ic3A7PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVy Ij50aGFua3MsPC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+LSBk aW1pdHJpLjwvRk9OVD48QlI+PEJSPjxGT05UIFNJWkU9Mj48Qj4mcXVvdDtBZHJpYW4gRmFycmVs JnF1b3Q7ICZsdDtvbGRkb2dAY2xhcmEuY28udWsmZ3Q7PC9CPjwvRk9OVD48QlI+PEZPTlQgU0la RT0yPlNlbnQgYnk6IG93bmVyLWNjYW1wQG9wcy5pZXRmLm9yZzwvRk9OVD48QlI+PEZPTlQgU0la RT0yPjA3LzI5LzIwMDUgMDg6MzI8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9Mj5QbGVhc2UgcmVzcG9u ZCB0byBhZHJpYW48L0ZPTlQ+PEJSPjxCUj4gPEZPTlQgU0laRT0yPlRvOjwvRk9OVD4gPEZPTlQg U0laRT0yPmNjYW1wQG9wcy5pZXRmLm9yZzwvRk9OVD48QlI+IDxGT05UIFNJWkU9Mj5jYzo8L0ZP TlQ+IDxGT05UIFNJWkU9Mj5zb2JAaGFydmFyZC5lZHUsIHppbmluQHBzZy5jb20sIGZlbm5lckBy ZXNlYXJjaC5hdHQuY29tPC9GT05UPjxCUj4gPEZPTlQgU0laRT0yPmJjYzo8L0ZPTlQ+IDxCUj4g PEZPTlQgU0laRT0yPlN1YmplY3Q6PC9GT05UPiA8Rk9OVCBTSVpFPTI+UHJvcG9zZWQgbGlhaXNv biB0byBTRzE1IG9uIFJGQyA0MTM5PC9GT05UPjxCUj4gPEJSPjxCUj48L1A+PFA+PEZPTlQgRkFD RT0iTW9ub3NwYWNlLENvdXJpZXIiPkhpLDxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9z cGFjZSxDb3VyaWVyIj5JIHByb3Bvc2UgdG8gc2VuZCB0aGlzIGxpYWlzb24gdG8gU3R1ZHkgR3Jv dXAgMTUgb2YgdGhlIElUVS1ULjxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxD b3VyaWVyIj5Db21tZW50cyBwbGVhc2UuPEJSPjwvRk9OVD48QlI+PEZPTlQgRkFDRT0iTW9ub3Nw YWNlLENvdXJpZXIiPlRoYW5rcyw8QlI+QWRyaWFuPEJSPjwvRk9OVD48QlI+PEZPTlQgRkFDRT0i TW9ub3NwYWNlLENvdXJpZXIiPj09PT09PT08QlI+PC9GT05UPjxCUj48Rk9OVCBGQUNFPSJNb25v c3BhY2UsQ291cmllciI+VG86IElUVS1UIFN0dWR5IEdyb3VwIDE1PEJSPkZyb206IElFVEYgQ0NB TVAgV29ya2luZyBHcm91cDxCUj5TdWJqZWN0OiBOZXcgUmVxdWVzdCBmb3IgQ29tbWVudCBQdWJs aXNoZWQ8QlI+Rm9yOiBJbmZvcm1hdGlvbjxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9z cGFjZSxDb3VyaWVyIj5UaGUgQ0NBTVAgV29ya2luZyBHcm91cCBvZiB0aGUgSUVURiB3aXNoZXMg dG8gaW5mb3JtIFN0dWR5IEdyb3VwIDE1IG9mIHRoZTxCUj5JVFUtVCBvZiB0aGUgcHVibGljYXRp b24gb2YgYSBuZXcgUmVxdWVzdCBmb3IgQ29tbWVudC48QlI+PC9GT05UPjxCUj48Rk9OVCBGQUNF PSJNb25vc3BhY2UsQ291cmllciI+UkZDIDQxMzkgJnF1b3Q7UmVxdWlyZW1lbnRzIGZvciBHZW5l cmFsaXplZCBNUExTIChHTVBMUykgU2lnbmFsaW5nIFVzYWdlIGFuZDxCUj5FeHRlbnNpb25zIGZv ciBBdXRvbWF0aWNhbGx5IFN3aXRjaGVkIE9wdGljYWwgTmV0d29yayAoQVNPTikmcXVvdDsgaXMg YXZhaWxhYmxlPEJSPmZyb20gdGhlIFJGQyBFZGl0b3IncyByZXBvc2l0b3J5IGF0PEJSPjxBIEhS RUY9ZnRwOi8vZnRwLnJmYy1lZGl0b3Iub3JnL2luLW5vdGVzL3JmYzQxMzkudHh0PmZ0cDovL2Z0 cC5yZmMtZWRpdG9yLm9yZy9pbi1ub3Rlcy9yZmM0MTM5LnR4dDwvQT48QlI+PC9GT05UPjxCUj48 Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+VGhlIEFic3RyYWN0IG9mIHRoaXMgUkZDIHJl YWRzIGFzIGZvbGxvd3MuPEJSPjwvRk9OVD48L1A+PFVMPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxD b3VyaWVyIj5UaGUgR2VuZXJhbGl6ZWQgTXVsdGktUHJvdG9jb2wgTGFiZWwgU3dpdGNoaW5nIChH TVBMUykgc3VpdGUgb2Y8QlI+cHJvdG9jb2xzIGhhcyBiZWVuIGRlZmluZWQgdG8gY29udHJvbCBk aWZmZXJlbnQgc3dpdGNoaW5nPEJSPnRlY2hub2xvZ2llcyBhbmQgZGlmZmVyZW50IGFwcGxpY2F0 aW9ucy4gJm5ic3A7VGhlc2UgaW5jbHVkZSBzdXBwb3J0PEJSPmZvciByZXF1ZXN0aW5nIFRpbWUg RGl2aXNpb24gTXVsdGlwbGV4aW5nIChURE0pIGNvbm5lY3Rpb25zLCBpbmNsdWRpbmc8QlI+U3lu Y2hyb25vdXMgT3B0aWNhbCBOZXR3b3JrIChTT05FVCkvU3luY2hyb25vdXMgRGlnaXRhbCBIaWVy YXJjaHk8QlI+KFNESCkgYW5kIE9wdGljYWwgVHJhbnNwb3J0IE5ldHdvcmtzIChPVE5zKS48L0ZP TlQ+PEJSPjxCUj48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+VGhpcyBkb2N1bWVudCBj b25jZW50cmF0ZXMgb24gdGhlIHNpZ25hbGluZyBhc3BlY3RzIG9mIHRoZSBHTVBMUzxCUj5zdWl0 ZSBvZiBwcm90b2NvbHMuICZuYnNwO0l0IGlkZW50aWZpZXMgdGhlIGZlYXR1cmVzIHRvIGJlIGNv dmVyZWQgYnkgdGhlPEJSPkdNUExTIHNpZ25hbGluZyBwcm90b2NvbCB0byBzdXBwb3J0IHRoZSBj YXBhYmlsaXRpZXMgb2YgYW48QlI+QXV0b21hdGljYWxseSBTd2l0Y2hlZCBPcHRpY2FsIE5ldHdv cmsgKEFTT04pLiAmbmJzcDtUaGlzIGRvY3VtZW50IHByb3ZpZGVzPEJSPmEgcHJvYmxlbSBzdGF0 ZW1lbnQgYW5kIGFkZGl0aW9uYWwgcmVxdWlyZW1lbnRzIGZvciB0aGUgR01QTFM8QlI+c2lnbmFs aW5nIHByb3RvY29sIHRvIHN1cHBvcnQgdGhlIEFTT04gZnVuY3Rpb25hbGl0eS48L0ZPTlQ+PEJS PjwvVUw+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPlRoZSBDQ0FNUCBXb3JraW5n IEdyb3VwIHdvdWxkIGxpa2UgdG8gdGhhbmsgdGhlIG1lbWJlcnMgYW5kIFF1ZXN0aW9ucyBvZjxC Uj5TdHVkeSBHcm91cCAxNSBmb3IgdGhlIGhlbHBmdWwgaW5wdXQgYW5kIHJldmlldyBmZWVkYmFj ayB0aGF0IHRoZXkgcHJvdmlkZWQ8QlI+d2hpbGUgdGhpcyBkb2N1bWVudCB3YXMgYmVpbmcgZGV2 ZWxvcGVkLjxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5SZWdh cmRzLDxCUj5BZHJpYW4gRmFycmVsIGFuZCBLaXJlZXRpIEtvbXBlbGxhPEJSPkNDQU1QIFdvcmtp bmcgR3JvdXAgQ28tY2hhaXJzPEJSPjwvRk9OVD48L1A+ Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 29 Jul 2005 20:41:41 +0000 To: "LE ROUX Jean-Louis RD-CORE-LAN" <jeanlouis.leroux@francetelecom.com> Cc: "Arthi Ayyangar" <arthi@juniper.net>, <ccamp@ops.ietf.org>, <jpv@cisco.com> MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: RE: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Date: Fri, 29 Jul 2005 22:39:04 +0200 Message-ID: <OF87E8B2A1.CD5437F6-ONC125704D.0071706F-C125704D.007170F8@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+Jm5ic3A7PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5oaSBqLWwsIGFydGhp PC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+PEJSPiZndDsgJmd0 OyAtSW4gc2VjdGlvbiA0LjEuMiB5b3UgcGFydGlhbGx5IGRlc2NyaWJlIGJpZGlyZWN0aW9uYWwg TFNQPEJSPiZndDsgc3RpdGNoaW5nPEJSPiZndDsgJmd0OyBwcm9jZWR1cmUuIFlvdSBtZW50aW9u IHRoYXQgYW4gVXBzdHJlYW0gTGFiZWwgTVVTVCBOT1QgYmU8QlI+Jmd0OyBhbGxvY2F0ZWQgYnk8 QlI+Jmd0OyAmZ3Q7IHRoZSBlbmQtdG8tZW5kIExTUCBvbiB0aGUgTFNQIHNlZ21lbnQsIHdoaWNo IGlzIE9LLiBCdXQ8QlI+Jmd0OyB0aGVuIGhvdyB0aGU8QlI+Jmd0OyAmZ3Q7IExTUC1TZWdtZW50 IEVncmVzcyB3aWxsIG5vdyB0aGF0IHRoZSBlbmQtdG8tZW5kIExTUCBpczxCUj4mZ3Q7IGJpZGly ZWN0aW9uYWw/PEJSPiZndDsgQUEtLS0tLS0tLSZndDsgRXhjZWxsZW50IHBvaW50LjxCUj4mZ3Q7 PEJSPiZndDsgJmd0OyBXaGF0IGFib3V0IGRlZmluaW5nIGEgZmxhZyBpbiB0aGUgQXR0cmlidXRl cyBGbGFncyBUTFYgb2YgdGhlPEJSPiZndDsgJmd0OyBMU1BfQVRUUklCVVRFIG9iamVjdCBzbyBh cyB0byBpbmRpY2F0ZSB0aGF0IHRoZSBMU1AgaXM8QlI+Jmd0OyBiaWRpcmVjdGlvbmFsPzxCUj4m Z3Q7IEFBLS0tLS0tJmd0OyBUaGF0IHdvdWxkIGJlIGEgY2hhbmdlIHRvIHRoZSBwcm9jZXNzaW5n IHRoYXQgR01QTFM8QlI+Jmd0OyBub2RlcyB1c2U8QlI+Jmd0OyBBQS0tLS0tLSZndDsgdG88QlI+ Jmd0OyBkZXRlY3QgYmlkaXJlY3Rpb25hbGl0eSwgaXNuJ3QgaXQgPyBOb3JtYWxseSBub2RlcyBs b29rIGZvcjxCUj4mZ3Q7IHRoZSBVcHN0cmVhbSBMYWJlbCBvYmplY3QgdG8gZGV0ZWN0IGJpZGly ZWN0aW9uYWxpdHkuPEJSPiZndDs8QlI+Jmd0OyBTbyBsZXQgdXMgc2F5IHRoYXQgYW4gTFNQIGlz IGJpZGlyZWN0aW9uYWwgaWYgYSkgYW4gVXBzdHJlYW08QlI+Jmd0OyBMYWJlbCBpcyBwcmVzZW50 IG9yIGIpIG5vIFVwc3RyZWFtIExhYmVsLCBidXQgYml0IHNldCBpbjxCUj4mZ3Q7IExTUF9BVFRS SUJVVEUgb3IgYykgYm90aCBIb3dldmVyLCByZWxpYW5jZSBvbiBhbiBlMmU8QlI+Jmd0OyBhdHRy aWJ1dGVzIGJpdCBzZXQgYnkgaGVhZCBlbmQsIG1lYW5zIGV4aXN0aW5nIGhlYWQgZW5kcyB3aWxs PEJSPiZndDsgbm90IGJlIHNldHRpbmcgdGhpcyBiaXQsIHNvIHRoYXQgd2lsbCBiZSBhbiBpc3N1 ZSAod3J0IGNvbXBhdGliaWxpdHkpLjxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFj ZSxDb3VyaWVyIj5KTExSOiBJIGFncmVlIHRoaXMgbWF5IHJhaXNlIHNvbWUgYmFja3dhcmQgY29t cGF0aWJpbGl0eSBpc3N1ZXMuPEJSPlRoaXMgd291bGQgcmVxdWlyZSB0aGF0IGhlYWQtZW5kcyBi ZSB1cGdyYWRlZC4uLjxCUj48L0ZPTlQ+PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3Vy aWVyIj5EUDogdGhlIG90aGVyIGlzc3VlIGlzIGhhdmluZyB0d28gbWV0aG9kcyBmb3IgaW5kaWNh dGluZyB0aGUgc2FtZSB0aGluZyBpcyBhbHNvIG5vdCBhZHZpc2FibGU8L0ZPTlQ+PC9QPjxQPjxG T05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj48QlI+Jmd0OyBDb3VsZCBiZSBuaWNlIGlmIHRo aXMgc2lnbmFsaW5nIHdhcyBqdXN0IGJldHdlZW4gdGhlIG5vZGU8QlI+Jmd0OyBkb2luZyB0aGUg c3RpdGNoaW5nIGFuZCB0aGUgZW5kIHBvaW50IG9mIHRoZSBMU1Agc2VnbWVudCw8QlI+Jmd0OyBz aW5jZSB0aGlzIGlzIHRoZSBob3AgdGhhdCB0aGUgYmlkaXJlY3Rpb25hbGl0eSBpbmZvcm1hdGlv bjxCUj4mZ3Q7IGlzIGxvc3QuIExldCBtZSB0aGluayBhYm91dCB0aGlzLjxCUj48L0ZPTlQ+PEJS PjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5KTExSOiBZZXMsIGFuZCB3aGF0IGFib3V0 IGRlZmluaW5nIGEgbmV3IE1QTFMgbGFiZWwgdmFsdWUsIGxldCBzYXkgbGFiZWwgNCA9ICZxdW90 O05vIGxhYmVsIGFzc2lnbmVkJnF1b3Q7LCB3aXRoIGEgc2VtYW50aWMgZGlzdGluY3QgZnJvbSBs YWJlbCAzIHNlbWFudGljICg9bGFiZWwgcG9wKS4gVGhlIFVQU1RSRUFNIGxhYmVsIG9iamVjdCB3 aXRoIHRoaXMgbmV3IGxhYmVsIHZhbHVlICZuYnNwO3dvdWxkIGJlIGluY2x1ZGVkIGluIHRoZSBl bmQtdG8tZW5kIFBhdGggbWVzc2FnZSBvdmVyIHRoZSBMU1AtU2VnbWVudC4uLjxCUj48L0ZPTlQ+ PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5EUDogYXNzdW1pbmcgYW4gaW5m b3JtYXRpb24gaGFzIHRvIGJlIHBhc3NlZCBmb3IgdGhlIGluZGljYXRpbmcgdGhpcyBjYXBhYmls aXR5IGkgZG8gYWxzbyB0aGluayBhIG1ldGhvZCBsaWtlIHBhc3NpbmcgYW4gaW1wbGljaXQgaW5k aWNhdGlvbiBpbiB0aGUgdXBzdHJlYW0gbGFiZWwgdmFsdWUgd291bGQgYmUgbW9yZSBhcHByb3By aWF0ZWQgKG5vdGUgdGhhdCB0aGUgaW5pdGlhbCBpc3N1ZSBpcyBkdWUgdG8gdGhlIGZhY3Qgb25l IHdvdWxkIGFsbG93IGZvciB1bmlkaXJlY3Rpb25hbCBlMmUgTFNQIG1ha2luZyB1c2Ugb2YgYmlk aXJlY3Rpb25hbCBMU1Agc2VnbWVudHMgLSBzZWUgY29tbWVudCBmcm9tIEotTCBoZXJlIGJlbG93 IC0gYW5kIG5vdCBtYWludGFpbmluZyBhIDE6MSByZWxhdGlvbnNoaXAgZm9yIHRoZSBkaXJlY3Rp b25hbGl0eSBiZXR3ZWVuIHRoZSBlMmUgTFNQIGFuZCB0aGUgTFNQIHNlZ21lbnQgdGhhdCBjYW4g YmUgcmV0cmlldmVkIHdpdGggdGhlIHByb2NlZHVyZSBkZXNjcmliZWQgaW4gc2VjdGlvbiA0LjIu NCAtIGFuZCBzbyB0aGUgZmlyc3QgcXVlc3Rpb24gaXMgaXQgd29ydGggYWxsb3dpbmcgdGhpcz8p IDwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPjxCUj4mZ3Q7ICZn dDsgQWxzbyB0aGUgc2VsZWN0aW9uIG9mIHRoZSBMU1Agc2VnbWVudCBpbiBjYXNlIG9mIGJpZGly ZWN0aW9uYWwgTFNQPEJSPiZndDsgJmd0OyBzaG91bGQgYmUgZGV0YWlsZWQgKGUuZy4gSWYgdGhl IGVuZC10by1lbmQgTFNQIGlzPEJSPiZndDsgYmlkaXJlY2l0b25hbCB0aGVuPEJSPiZndDsgJmd0 OyB0aGUgTFNQLXNlZ21lbnQgTVVTVCBiZSBiaWRpcmVjdGlvbmFsLiBBbHNvIHNoYWxsIHdlIGFs bG93IHRoYXQgdHdvPEJSPiZndDsgJmd0OyB1bmlkcmVjdGlvbmFsIGVuZC10by1lbmQgTFNQIHVz ZSB0aGUgc2FtZSBiaWRpcmVjdGlvbmFsIExTUCBzZWdtZW50PEJSPiZndDsgJmd0OyAob25lIGlu IGVhY2ggZGlyZWN0aW9uKT88QlI+Jmd0OyBBQS0tLS0mZ3Q7IFllcywgdGhpcyBzaG91bGQgYmUg b2theS4gSU1PLjxCUj4mZ3Q7PEJSPiZndDsgJmd0OyAtQXQgdGhlIGVuZCBvZiBzZWN0aW9uIDQu Mi41IHlvdSBtZW50aW9uIHRoYXQgTFNQLVNlZ21lbnQ8QlI+Jmd0OyBmYWlsdXJlIG9yPEJSPiZn dDsgJmd0OyBtYWludGVuYW5jZSBTSE9VTEQgYmUgdHJlYXRlZCBhcyBhIGZhaWx1cmUgZXZlbnQg Zm9yIHRoZTxCUj4mZ3Q7IGVuZC10by1lbmQgTFNQLjxCUj4mZ3Q7ICZndDsgSSBhZ3JlZSBmb3Ig TFNQLVNlZ21lbnQgZmFpbHVyZSBidXQgbm90IGZvciBMU1AtU2VnbWVudCBtYWludGVuYW5jZS48 QlI+Jmd0OyAmZ3Q7IExTUC1TZWdtZW50IG1haW50ZW5hbmNlIHNob3VsZCBiZSB0cmVhdGVkIGFz IFRFLWxpbms8QlI+Jmd0OyBtYWludGVuYW5jZSBmb3I8QlI+Jmd0OyAmZ3Q7IHRoZSBlbmQtdG8t ZW5kIExTUCwgYW5kIHByb2NlZHVyZXMgZGVmaW5lZCBpbiBHTVBMUzxCUj4mZ3Q7IGdyYWNlZnVs IFRFLWxpbms8QlI+Jmd0OyAmZ3Q7IHNodXRkb3duIGRyYWZ0IG1heSBiZSB1c2VmdWwgKFNwZWNp ZmljIFJTVlAgZXJyb3IgY29kZSBhbmQgVEUtbGluazxCUj4mZ3Q7ICZndDsgYXR0cmlidXRlKS4u LjxCUj4mZ3Q7IEFBLS0tJmd0OyBZZXMsIEkgYWdyZWU7IHdlIHdpbGwgbWVudGlvbiB0aGlzLiBB Y3R1YWxseSB0aGUgZ3JhY2VmdWw8QlI+Jmd0OyBBQS0tLSZndDsgdGVhcmRvd248QlI+Jmd0OyBz ZW50ZW5jZSBvY2N1cnMgZmV3IHBhcmFncmFwaHMgYmVmb3JlLCBidXQgbm90IGluIHRoZSByaWdo dDxCUj4mZ3Q7IGNvbnRleHQuIFNvIHdlIHdpbGwgY2xhcmlmeSB0aGUgYWJvdmUuPEJSPjwvRk9O VD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPkRQOiBpIGFtIG5vdCBzdXJl IHRvIHVuZGVyc3RhbmQgdGhlIGNvbW1lbnQsIHRoZSBzZW50ZW5jZSBzcGVha3MgYWJvdXQgZGVs ZXRpb24gb2YgdGhlIExTUCBzZWdtZW50IGR1ZSB0byBmYWlsdXJlL21haW50ZW5hbmNlIHNob3Vs ZCBiZSB0cmVhdGVkIGFzIGEgZmFpbHVyZSBldmVudCBmb3IgdGhlIGUyZSBMU1AsIHNvIGkgZG9u J3QgdGhpbmsgdGhlIGluaXRpYWwgc2VudGVuY2Ugd2FzIG1lYW50IHRvIHRyYW5zbGF0ZWQgdGhl IGluaXRpYWwgY29tbWVudCwgYW55d2F5IGl0IHdvdWxkIGJlIG9mIGludGVyZXN0IHRoYXQgeW91 IGRldGFpbCB0aGUgZXJyb3IgY29kZTwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNl LENvdXJpZXIiPkRQOiBub3RlIGFsc28gdGhhdCBzZWN0aW9uIDQuMi41IGluZGljYXRlcyB0aGF0 IGZvciBzdGl0Y2hlZCBMU1BzIEdyYWNlZnVsIGRlbGV0aW9uIGNhbiBiZSB1c2VkIC0gaSBwZXJm ZWN0bHkgYWdyZWUgLSBidXQgdGhlIHNlcXVlbmNlIHNob3VsZCBiZSBkZXRhaWxlZCBhcyBwYXJ0 IG9mIHRoaXMgc2VjdGlvbiA8L0ZPTlQ+PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3Vy aWVyIj48QlI+Jmd0OyAmZ3Q7IEhvcGUgdGhpcyBoZWxwcyw8QlI+Jmd0OyBBQS0tLSZndDsgJm5i c3A7U3VyZSBkb2VzLjxCUj4mZ3Q7PEJSPiZndDs8QlI+Jmd0OyBUaGFua3MhPEJSPiZndDs8QlI+ Jmd0OyAtYXJ0aGk8QlI+Jmd0OzxCUj4mZ3Q7ICZndDs8QlI+Jmd0OyAmZ3Q7PEJSPiZndDsgJmd0 OyAmZ3Q7IC0tLS0tTWVzc2FnZSBkJ29yaWdpbmUtLS0tLTxCUj4mZ3Q7ICZndDsgJmd0OyBEZSA6 IG93bmVyLWNjYW1wQG9wcy5pZXRmLm9yZzxCUj4mZ3Q7ICZndDsgJmd0OyBbbWFpbHRvOm93bmVy LWNjYW1wQG9wcy5pZXRmLm9yZ10gRGUgbGEgcGFydCBkZTxCUj4mZ3Q7ICZndDsgJmd0OyBJbnRl cm5ldC1EcmFmdHNAaWV0Zi5vcmcgRW52b3kmZWFjdXRlOyA6IHZlbmRyZWRpIDE1IGp1aWxsZXQ8 QlI+Jmd0OyAyMDA1IDIxOjUwICZBZ3JhdmU7IDo8QlI+Jmd0OyAmZ3Q7ICZndDsgaS1kLWFubm91 bmNlQGlldGYub3JnIENjIDogY2NhbXBAb3BzLmlldGYub3JnIE9iamV0IDogSS1EPEJSPiZndDsg Jmd0OyAmZ3Q7IEFDVElPTjpkcmFmdC1pZXRmLWNjYW1wLWxzcC1zdGl0Y2hpbmctMDEudHh0PEJS PiZndDsgJmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyAmZ3Q7IEEgTmV3IEludGVybmV0LURyYWZ0IGlz IGF2YWlsYWJsZSBmcm9tIHRoZSBvbi1saW5lPEJSPiZndDsgSW50ZXJuZXQtRHJhZnRzPEJSPiZn dDsgJmd0OyAmZ3Q7IGRpcmVjdG9yaWVzLjxCUj4mZ3Q7ICZndDsgJmd0OyBUaGlzIGRyYWZ0IGlz IGEgd29yayBpdGVtIG9mIHRoZSBDb21tb24gQ29udHJvbCBhbmQgTWVhc3VyZW1lbnQ8QlI+Jmd0 OyAmZ3Q7ICZndDsgUGxhbmUgV29ya2luZyBHcm91cCBvZiB0aGUgSUVURi48QlI+Jmd0OyAmZ3Q7 ICZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsgVGl0bGU6IExhYmVsIFN3aXRjaGVkIFBhdGggU3RpdGNo aW5nIHdpdGg8QlI+Jmd0OyAmZ3Q7ICZndDsgR2VuZXJhbGl6ZWQgTVBMUyBUcmFmZmljIEVuZ2lu ZWVyaW5nPEJSPiZndDsgJmd0OyAmZ3Q7IEF1dGhvcihzKTogQS4gQXl5YW5nYXIsIEouIFZhc3Nl dXI8QlI+Jmd0OyAmZ3Q7ICZndDsgRmlsZW5hbWU6IGRyYWZ0LWlldGYtY2NhbXAtbHNwLXN0aXRj aGluZy0wMS50eHQ8QlI+Jmd0OyAmZ3Q7ICZndDsgUGFnZXM6IDE5PEJSPiZndDsgJmd0OyAmZ3Q7 IERhdGU6IDIwMDUtNy0xNTxCUj4mZ3Q7ICZndDsgJmd0OzxCUj4mZ3Q7ICZndDsgJmd0OyBJbiBj ZXJ0YWluIHNjZW5hcmlvcywgdGhlcmUgbWF5IGJlIGEgbmVlZCB0byBjb21iaW5lIHRvZ2V0aGVy IHR3bzxCUj4mZ3Q7ICZndDsgJmd0OyAmbmJzcDsgJm5ic3A7ZGlmZmVyZW50IEdlbmVyYWxpemVk IE11bHRpLVByb3RvY29sIExhYmVsIFN3aXRjaGluZzxCUj4mZ3Q7IChHTVBMUykgTGFiZWw8QlI+ Jmd0OyAmZ3Q7ICZndDsgJm5ic3A7ICZuYnNwO1N3aXRjaGVkIFBhdGhzIChMU1BzKSBzdWNoIHRo YXQgaW4gdGhlIGRhdGEgcGxhbmUsIGE8QlI+Jmd0OyBzaW5nbGUgZW5kLXRvLTxCUj4mZ3Q7ICZn dDsgJmd0OyAmbmJzcDsgJm5ic3A7ZW5kIChlMmUpIExTUCBpcyBhY2hpZXZlZCBhbmQgYWxsIHRy YWZmaWMgZnJvbSBvbmUgTFNQPEJSPiZndDsgaXMgc3dpdGNoZWQ8QlI+Jmd0OyAmZ3Q7ICZndDsg Jm5ic3A7ICZuYnNwO29udG8gdGhlIG90aGVyIExTUC4gJm5ic3A7V2Ugd2lsbCByZWZlciB0byB0 aGlzIGFzICZxdW90O0xTUDxCUj4mZ3Q7IHN0aXRjaGluZyZxdW90Oy48QlI+Jmd0OyAmZ3Q7ICZn dDsgVGhpczxCUj4mZ3Q7ICZndDsgJmd0OyAmbmJzcDsgJm5ic3A7ZG9jdW1lbnQgY292ZXJzIGNh c2VzIHdoZXJlOiBhKSB0aGUgbm9kZSBwZXJmb3JtaW5nPEJSPiZndDsgdGhlIHN0aXRjaGluZzxC Uj4mZ3Q7ICZndDsgJmd0OyAmbmJzcDsgJm5ic3A7ZG9lcyBub3QgcmVxdWlyZSBjb25maWd1cmF0 aW9uIG9mIGV2ZXJ5IExTUCBwYWlyIHRvIGJlIHN0aXRjaGVkPEJSPiZndDsgJmd0OyAmZ3Q7ICZu YnNwOyAmbmJzcDt0b2dldGhlciBiKSB0aGUgbm9kZSBwZXJmb3JtaW5nIHRoZSBzdGl0Y2hpbmcg aXMgbm90PEJSPiZndDsgdGhlIGVncmVzcyBvZjxCUj4mZ3Q7ICZndDsgJmd0OyAmbmJzcDsgJm5i c3A7YW55IG9mIHRoZSBMU1BzIGMpIExTUCBzdGl0Y2hpbmcgbm90IG9ubHkgcmVzdWx0cyBpbjxC Uj4mZ3Q7IGFuIGVuZC10by1lbmQ8QlI+Jmd0OyAmZ3Q7ICZndDsgJm5ic3A7ICZuYnNwO0xTUCBp biB0aGUgZGF0YSBwbGFuZSwgYnV0IHRoZXJlIGlzIGFsc28gYTxCUj4mZ3Q7IGNvcnJlc3BvbmRp bmcgZW5kLXRvLWVuZDxCUj4mZ3Q7ICZndDsgJmd0OyAmbmJzcDsgJm5ic3A7TFNQIChSU1ZQIHNl c3Npb24pIGluIHRoZSBjb250cm9sIHBsYW5lLiAmbmJzcDtJdCBtaWdodCBiZTxCUj4mZ3Q7IHBv c3NpYmxlIHRvPEJSPiZndDsgJmd0OyAmZ3Q7ICZuYnNwOyAmbmJzcDtjb25maWd1cmUgYSBHTVBM UyBub2RlIHRvIHN3aXRjaCB0aGUgdHJhZmZpYyBmcm9tIGFuIExTUCBmb3I8QlI+Jmd0OyAmZ3Q7 ICZndDsgd2hpY2ggaXQ8QlI+Jmd0OyAmZ3Q7ICZndDsgJm5ic3A7ICZuYnNwO2lzIHRoZSBlZ3Jl c3MsIHRvIGFub3RoZXIgTFNQIGZvciB3aGljaCBpdCBpcyB0aGU8QlI+Jmd0OyBpbmdyZXNzLCB3 aXRob3V0PEJSPiZndDsgJmd0OyAmZ3Q7ICZuYnNwOyAmbmJzcDtyZXF1aXJpbmcgYW55IHNpZ25h bGluZyBvciByb3V0aW5nIGV4dGVuc2lvbnMgd2hhdHNvZXZlciw8QlI+Jmd0OyAmZ3Q7ICZndDsg Y29tcGxldGVseTxCUj4mZ3Q7ICZndDsgJmd0OyAmbmJzcDsgJm5ic3A7dHJhbnNwYXJlbnQgdG8g b3RoZXIgbm9kZXMuICZuYnNwO1RoaXMgd2lsbCBhbHNvIHJlc3VsdCBpbjxCUj4mZ3Q7IExTUCBz dGl0Y2hpbmc8QlI+Jmd0OyAmZ3Q7ICZndDsgJm5ic3A7ICZuYnNwO2luIHRoZSBkYXRhIHBsYW5l LiAmbmJzcDtIb3dldmVyLCB0aGlzIGRvY3VtZW50IGRvZXMgbm90IGNvdmVyIHRoaXM8QlI+Jmd0 OyAmZ3Q7ICZndDsgJm5ic3A7ICZuYnNwO3NjZW5hcmlvIG9mIExTUCBzdGl0Y2hpbmcuPEJSPiZn dDsgJmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyAmZ3Q7IEEgVVJMIGZvciB0aGlzIEludGVybmV0LURy YWZ0IGlzOjxCUj4mZ3Q7ICZndDsgJmd0OyA8QSBIUkVGPWh0dHA6Ly93d3cuaWV0Zi5vcmcvaW50 ZXJuZXQtZHJhZnRzL2RyYWZ0LWlldGYtY2NhbXAtbHNwLXN0aXRjPmh0dHA6Ly93d3cuaWV0Zi5v cmcvaW50ZXJuZXQtZHJhZnRzL2RyYWZ0LWlldGYtY2NhbXAtbHNwLXN0aXRjPC9BPjxCUj4mZ3Q7 ICZndDsgaGluZy0wMS50eHQ8QlI+Jmd0OyAmZ3Q7ICZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsgVG8g cmVtb3ZlIHlvdXJzZWxmIGZyb20gdGhlIEktRCBBbm5vdW5jZW1lbnQgbGlzdCwgc2VuZCBhPEJS PiZndDsgbWVzc2FnZSB0bzxCUj4mZ3Q7ICZndDsgJmd0OyBpLWQtYW5ub3VuY2UtcmVxdWVzdEBp ZXRmLm9yZyB3aXRoIHRoZSB3b3JkIHVuc3Vic2NyaWJlPEJSPiZndDsgaW4gdGhlIGJvZHk8QlI+ Jmd0OyAmZ3Q7ICZndDsgb2YgdGhlIG1lc3NhZ2UuPEJSPiZndDsgJmd0OyAmZ3Q7IFlvdSBjYW4g YWxzbyB2aXNpdDxCUj4mZ3Q7ICZndDsgJmd0OyA8QSBIUkVGPWh0dHBzOi8vd3d3MS5pZXRmLm9y Zy9tYWlsbWFuL2xpc3RpbmZvL0ktRC1hbm5vdW5jZT5odHRwczovL3d3dzEuaWV0Zi5vcmcvbWFp bG1hbi9saXN0aW5mby9JLUQtYW5ub3VuY2U8L0E+PEJSPiZndDsgJmd0OyAmZ3Q7IHRvIGNoYW5n ZSB5b3VyIHN1YnNjcmlwdGlvbiBzZXR0aW5ncy48QlI+Jmd0OyAmZ3Q7ICZndDs8QlI+Jmd0OyAm Z3Q7ICZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsgSW50ZXJuZXQtRHJhZnRzIGFyZSBhbHNvIGF2YWls YWJsZSBieSBhbm9ueW1vdXMgRlRQLjxCUj4mZ3Q7IExvZ2luIHdpdGggdGhlPEJSPiZndDsgJmd0 OyAmZ3Q7IHVzZXJuYW1lICZxdW90O2Fub255bW91cyZxdW90OyBhbmQgYSBwYXNzd29yZCBvZiB5 b3VyIGUtbWFpbCBhZGRyZXNzLiBBZnRlcjxCUj4mZ3Q7ICZndDsgJmd0OyBsb2dnaW5nIGluLCB0 eXBlICZxdW90O2NkIGludGVybmV0LWRyYWZ0cyZxdW90OyBhbmQgdGhlbjxCUj4mZ3Q7ICZndDsg Jmd0OyAmcXVvdDtnZXQgZHJhZnQtaWV0Zi1jY2FtcC1sc3Atc3RpdGNoaW5nLTAxLnR4dCZxdW90 Oy48QlI+Jmd0OyAmZ3Q7ICZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsgQSBsaXN0IG9mIEludGVybmV0 LURyYWZ0cyBkaXJlY3RvcmllcyBjYW4gYmUgZm91bmQgaW48QlI+Jmd0OyAmZ3Q7ICZndDsgPEEg SFJFRj1odHRwOi8vd3d3LmlldGYub3JnL3NoYWRvdy5odG1sPmh0dHA6Ly93d3cuaWV0Zi5vcmcv c2hhZG93Lmh0bWw8L0E+IG9yPEJSPiZndDsgJmd0OyAmZ3Q7IDxBIEhSRUY9ZnRwOi8vZnRwLmll dGYub3JnL2lldGYvMXNoYWRvdy1zaXRlcy50eHQ+ZnRwOi8vZnRwLmlldGYub3JnL2lldGYvMXNo YWRvdy1zaXRlcy50eHQ8L0E+PEJSPiZndDsgJmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyAmZ3Q7PEJS PiZndDsgJmd0OyAmZ3Q7IEludGVybmV0LURyYWZ0cyBjYW4gYWxzbyBiZSBvYnRhaW5lZCBieSBl LW1haWwuPEJSPiZndDsgJmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyAmZ3Q7IFNlbmQgYSBtZXNzYWdl IHRvOjxCUj4mZ3Q7ICZndDsgJmd0OyBtYWlsc2VydkBpZXRmLm9yZy48QlI+Jmd0OyAmZ3Q7ICZn dDsgSW4gdGhlIGJvZHkgdHlwZTo8QlI+Jmd0OyAmZ3Q7ICZndDsgJnF1b3Q7RklMRSAvaW50ZXJu ZXQtZHJhZnRzL2RyYWZ0LWlldGYtY2NhbXAtbHNwLXN0aXRjaGluZy0wMS50eHQmcXVvdDsuPEJS PiZndDsgJmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyAmZ3Q7IE5PVEU6VGhlIG1haWwgc2VydmVyIGF0 IGlldGYub3JnIGNhbiByZXR1cm4gdGhlIGRvY3VtZW50IGluPEJSPiZndDsgJmd0OyAmZ3Q7IE1J TUUtZW5jb2RlZCBmb3JtIGJ5IHVzaW5nIHRoZSAmcXVvdDttcGFjayZxdW90OyB1dGlsaXR5LiAm bmJzcDtUbyB1c2UgdGhpczxCUj4mZ3Q7ICZndDsgJmd0OyBmZWF0dXJlLCBpbnNlcnQgdGhlIGNv bW1hbmQgJnF1b3Q7RU5DT0RJTkcgbWltZSZxdW90OyBiZWZvcmUgdGhlICZxdW90O0ZJTEUmcXVv dDs8QlI+Jmd0OyAmZ3Q7ICZndDsgY29tbWFuZC4gJm5ic3A7VG8gZGVjb2RlIHRoZSByZXNwb25z ZShzKSwgeW91IHdpbGwgbmVlZCAmcXVvdDttdW5wYWNrJnF1b3Q7IG9yPEJSPiZndDsgJmd0OyAm Z3Q7IGEgTUlNRS1jb21wbGlhbnQgbWFpbCByZWFkZXIuICZuYnNwO0RpZmZlcmVudCBNSU1FLWNv bXBsaWFudCBtYWlsPEJSPiZndDsgJmd0OyAmZ3Q7IHJlYWRlcnM8QlI+Jmd0OyAmZ3Q7ICZndDsg ZXhoaWJpdCBkaWZmZXJlbnQgYmVoYXZpb3IsIGVzcGVjaWFsbHkgd2hlbiBkZWFsaW5nIHdpdGg8 QlI+Jmd0OyAmZ3Q7ICZndDsgJnF1b3Q7bXVsdGlwYXJ0JnF1b3Q7IE1JTUUgbWVzc2FnZXMgKGku ZS4gZG9jdW1lbnRzIHdoaWNoIGhhdmUgYmVlbiBzcGxpdDxCUj4mZ3Q7ICZndDsgJmd0OyB1cCBp bnRvIG11bHRpcGxlIG1lc3NhZ2VzKSwgc28gY2hlY2sgeW91ciBsb2NhbCBkb2N1bWVudGF0aW9u IG9uPEJSPiZndDsgJmd0OyAmZ3Q7IGhvdyB0byBtYW5pcHVsYXRlIHRoZXNlIG1lc3NhZ2VzLjxC Uj4mZ3Q7ICZndDsgJmd0OzxCUj4mZ3Q7ICZndDsgJmd0OzxCUj4mZ3Q7ICZndDsgJmd0OyBCZWxv dyBpcyB0aGUgZGF0YSB3aGljaCB3aWxsIGVuYWJsZSBhIE1JTUUgY29tcGxpYW50IG1haWwgcmVh ZGVyPEJSPiZndDsgJmd0OyAmZ3Q7IGltcGxlbWVudGF0aW9uIHRvIGF1dG9tYXRpY2FsbHkgcmV0 cmlldmUgdGhlIEFTQ0lJIHZlcnNpb24gb2YgdGhlPEJSPiZndDsgJmd0OyAmZ3Q7IEludGVybmV0 LURyYWZ0LjxCUj4mZ3Q7ICZndDsgJmd0OzxCUj4mZ3Q7ICZndDs8QlI+Jmd0OzwvRk9OVD48L1A+ Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 29 Jul 2005 15:27:29 +0000 Message-ID: <42EA4A1E.8030901@lab.ntt.co.jp> Date: Sat, 30 Jul 2005 00:24:14 +0900 From: Kohei Shiomoto <shiomoto.kohei@lab.ntt.co.jp> User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; ja-JP; rv:1.7.7) Gecko/20050414 MIME-Version: 1.0 To: LE ROUX Jean-Louis RD-CORE-LAN <jeanlouis.leroux@francetelecom.com> CC: rpapneja@isocore.com, Richard Rabbat <richard.rabbat@us.fujitsu.com>, ccamp@ops.ietf.org Subject: Re: I-D ACTION:draft-ietf-ccamp-gmpls-addressing-01.txt Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: quoted-printable Hi Jean-Louis Thank you for your comments. Your point sounds important when we=20 consider the case IPv4 and IPv6 coexist. But I feel that the IPv4 and IPv6 interwork seems a little bit future=20 topic in a (G)MPLS context, in particular, from the control plane=20 perspective. I would be happy if you could provide your experience with the actual=20 network operation and contribution. Best regards, -- Kohei LE ROUX Jean-Louis RD-CORE-LAN wrote: > Hi Kohei, Rajiv, Richard, >=20 > I have on comment related to your GMPLS addressing ID: > In section 5.2 you recommend, that in the session object, the IP tunnel= end-point be set to the destination TE router id and that the extended t= unnel id be set to the source TE router id, which makes sense. But as per= 3209, a session object can either encode IPv4 or IPv6 addresses but not = both types, and there may be an issue if source TE router id is IPv6 and = destination TE router id is IPv4 (or vice versa). This may happen for ins= tance with inter-AS/SP LSPs. IMHO we should address this point. >=20 > Best Regards, >=20 > JL >=20 >=20 >=20 >=20 >=20 >>-----Message d'origine----- >>De : owner-ccamp@ops.ietf.org=20 >>[mailto:owner-ccamp@ops.ietf.org] De la part de=20 >>Internet-Drafts@ietf.org >>Envoy=EF=BF=BD : mardi 21 juin 2005 21:50 >>=EF=BF=BD : i-d-announce@ietf.org >>Cc : ccamp@ops.ietf.org >>Objet : I-D ACTION:draft-ietf-ccamp-gmpls-addressing-01.txt=20 >> >>A New Internet-Draft is available from the on-line=20 >>Internet-Drafts directories. >>This draft is a work item of the Common Control and=20 >>Measurement Plane Working Group of the IETF. >> >> Title : Use of Addresses in Generalized Multi-Protocol >> Label Switching (GMPLS) Networks >> Author(s) : K. Shiomoto, et al. >> Filename : draft-ietf-ccamp-gmpls-addressing-01.txt >> Pages : 23 >> Date : 2005-6-21 >>=09 >>This document explains and clarifies the use of addresses in=20 >> Generalized Multi-Protocol Label Switching (GMPLS)=20 >>networks. The aim=20 >> of this document is to facilitate and ensure better=20 >>interworking of=20 >> GMPLS-capable Label Switching Routers (LSRs) based on experience=20 >> gained in deployments and interoperability testing and proper=20 >> interpretation of published RFCs.=20 >> =20 >> The document recommends a proper approach for the=20 >>interpretation and=20 >> choice of address and identifier fields within GMPLS protocols and=20 >> references specific control plane usage models. It also examines=20 >> some common GMPLS Resource Reservation Protocol-Traffic=20 >>Engineering=20 >> (RSVP-TE) signaling message processing issues and recommends=20 >> solutions. It finally discusses how to handle IPv6 sources and=20 >> destinations in the MPLS and GMPLS TE (Traffic Engineering) MIB=20 >> (Management Information Base) modules. >> >>A URL for this Internet-Draft is: >>http://www.ietf.org/internet-drafts/draft-ietf-ccamp-gmpls-add >=20 > ressing-01.txt >=20 >>To remove yourself from the I-D Announcement list, send a=20 >>message to i-d-announce-request@ietf.org with the word=20 >>unsubscribe in the body of the message. =20 >>You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce >>to change your subscription settings. >> >> >>Internet-Drafts are also available by anonymous FTP. Login=20 >>with the username "anonymous" and a password of your e-mail=20 >>address. After logging in, type "cd internet-drafts" and then >> "get draft-ietf-ccamp-gmpls-addressing-01.txt". >> >>A list of Internet-Drafts directories can be found in=20 >>http://www.ietf.org/shadow.html or=20 >>ftp://ftp.ietf.org/ietf/1shadow-sites.txt >> >> >>Internet-Drafts can also be obtained by e-mail. >> >>Send a message to: >> mailserv@ietf.org. >>In the body type: >> "FILE=20 >>/internet-drafts/draft-ietf-ccamp-gmpls-addressing-01.txt". >>=09 >>NOTE: The mail server at ietf.org can return the document in >> MIME-encoded form by using the "mpack" utility. To use this >> feature, insert the command "ENCODING mime" before the "FILE" >> command. To decode the response(s), you will need "munpack" or >> a MIME-compliant mail reader. Different MIME-compliant=20 >>mail readers >> exhibit different behavior, especially when dealing with >> "multipart" MIME messages (i.e. documents which have been split >> up into multiple messages), so check your local documentation on >> how to manipulate these messages. >> =09 >> =09 >>Below is the data which will enable a MIME compliant mail reader >>implementation to automatically retrieve the ASCII version of the >>Internet-Draft. >> >=20 >=20 >=20 >=20 --=20 Kohei Shiomoto, Ph.D Senior Research Engineer, Supervisor NTT Network Service Systems Laboratories 3-9-11 Midori, Musashino, Tokyo 180-8585, Japan Phone +81 422 59 4402 Fax +81 422 59 3787 Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 29 Jul 2005 09:20:51 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: ARE: I-D ACTION:draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt Date: Fri, 29 Jul 2005 11:18:49 +0200 Message-ID: <D109C8C97C15294495117745780657AE02F23552@ftrdmel1.rd.francetelecom.fr> Thread-Topic: ARE: I-D ACTION:draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt Thread-Index: AcWTmolU8tLIPi2FQB6cHsE87A63/AAgyNng From: "LE ROUX Jean-Louis RD-CORE-LAN" <jeanlouis.leroux@francetelecom.com> To: "Arthi Ayyangar" <arthi@juniper.net> Cc: "JP Vasseur" <jvasseur@cisco.com>, <Raymond_Zhang@infonet.com>, <ccamp@ops.ietf.org> Hi Arthi > AA---------> Actually the section on Assumptions where we introduce > Auto-discovery does say that it would rely on IGP and/or BGP=20 > information. > Would you just like us to expand that to include use of local=20 > policies in conjuction with the above information when there=20 > are sevaral downstream candidates ? Yes that would be good, as actually, in an inter-SP context, such = policies based on the so called "TE peering contracts" might be of huge = importance... Thanks, JL > -----Message d'origine----- > De : Arthi Ayyangar [mailto:arthi@juniper.net]=20 > Envoy=E9 : jeudi 28 juillet 2005 19:34 > =C0 : LE ROUX Jean-Louis RD-CORE-LAN > Cc : JP Vasseur; Raymond_Zhang@infonet.com; ccamp@ops.ietf.org > Objet : Re: ARE: I-D=20 > ACTION:draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt >=20 > Hi JL, >=20 > (You seem to be on a roll...14 IDs down, 4 to go... :-)) :-) >=20 > > A few comments related to your inter-domain pd draft > AA---> Thanks for the comments. >=20 > > -Section: 1) common assumptions, Why not adding and addressing the=20 > > case where a path is made of a set of abstract nodes? For=20 > instance, in=20 > > the inter-AS case, an ERO may list a set of ASes. This may require=20 > > specific procedures when computing the path and selecting=20 > the downstream ASBR... > AA-------> True. This needs to be addressed. The procedures would=20 > AA-------> require > interaction with BGP to find the ASBR exit point for the nexthop AS. >=20 > > -Section 4: 1) Even if this is not really in the scope I=20 > think that=20 > > the auto-discovery mechanism would require some more words. This is=20 > > partially addressed latter in the draft(4.2.1), but you=20 > should mention=20 > > here that such auto-discovery may rely on IGP information=20 > and/or BGP=20 > > information and also on local policies (particularly in inter-AS/SP=20 > > case). In the inter-AS/SP case there may be several candidate=20 > > downstream ASBRs, pertaining to distinct ASes/SPs, and the=20 > downstream=20 > > AS selection may require appying policies related to=20 > specific "TE peering contract" > > negociated with neigbors Ases/SPs... > AA---------> Actually the section on Assumptions where we introduce > Auto-discovery does say that it would rely on IGP and/or BGP=20 > information. > Would you just like us to expand that to include use of local=20 > policies in conjuction with the above information when there=20 > are sevaral downstream candidates ? Yes that would be good, as actually in an inter-SP context these = policies might be of huge importance >=20 >=20 > > -Section 4 2) b) ii) What do you exactly mean by "the=20 > boundary LSR is=20 > > not a FA-LSP/LSP segment candidate"? Does it mean in this=20 > case, that a=20 > > contigous LSP is created?, this would not be consistent with b) > AA----> This is what happens when drafts are merged, split,=20 > re-named; etc. > :-). One of the versions of the draft actually explained what=20 > "candidate" > meant...Well... > This means that if the LSR has no local policy for nesting=20 > or stitching the inter-domain LSP, then it would simply set=20 > up a contiguous LSP. > Also, I think we need to fix the assumption that FA LSPs=20 > or LSP segments can exist only between domain boundary LSRs.=20 > It might be the case, but it is not necessary. >=20 > > -Top of Section 4.1.1: You should remove the first sentence=20 > after the=20 > > three path examples, "At least the set", as this sounds an=20 > ambiguous=20 > > repetition of what is just stated above > AA-----> Okay. >=20 > > -Section 6.2: "If the Inter-AS TE related Information is flooded in=20 > > the IGP, each ASBR is capable of computing the path..." > > > > Actually there may be some topologies where an ASBR cannot=20 > compute a=20 > > link protection path. > > > > See the following example > > > > AS1--> <---AS2---- > > ASBR1----ASBR2 > > | | > > | | > > ASBR3----ASBR4 > > > > ASBR1 is aware of links ASBR1->ASBR2, ASBR1->ASBR3 and ASBR3->ASBR4=20 > > but not the link ASBR2->ASBR4 as there is no flooding=20 > between AS1 and AS2. > > Hence ASBR1 cannot compute a path to protect link ASBR1-ASBR2. > > > > Actually this case is similar to node protection issue, the path=20 > > should be computed according to the pd computation methods=20 > defined in the doc. > AA--------> I agree completely. >=20 > > > > Also it would be useful to mention that pd inter-domain backup path > may requires XRO incusion in RSVP messages... > AA-----> That's fine. This is useful for any computation=20 > where loose hop > expansion occurs in different domains. >=20 > > -Section 7.2 > > > > "Note that stiching or nesting rely on local=20 > optimization...hence the=20 > > reoptimization is not global and... Procedures describes in=20 > loose hop=20 > > MUST be used... " It would be good to mention that global PCE-based=20 > > reoptimization (with potential change of border LSRs) also apply to=20 > > stiching/nesting. > AA-------> I couldn't more. >=20 > thanks, > -arthi >=20 > In both cases, contiguous or stiching/nesting, PCE or > > any offline computation is required to achieve end-to-end=20 > > reoptimization. >=20 >=20 > > > > > > > > > > > -----Message d'origine----- > > > De : owner-ccamp@ops.ietf.org > > > [mailto:owner-ccamp@ops.ietf.org] De la part de=20 > > > Internet-Drafts@ietf.org Envoy=E9 : vendredi 15 avril 2005=20 > 21:34 =C0 :=20 > > > i-d-announce@ietf.org Cc : ccamp@ops.ietf.org Objet : I-D=20 > > > ACTION:draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt > > > > > > A New Internet-Draft is available from the on-line=20 > Internet-Drafts=20 > > > directories. > > > This draft is a work item of the Common Control and Measurement=20 > > > Plane Working Group of the IETF. > > > > > > Title : A Per-domain path computation method for > > > computing Inter-domain Traffic > > > Engineering (TE) > > > Label Switched Path (LSP) > > > Author(s) : J. Vasseur, et al. > > > Filename : > > > draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt > > > Pages : 0 > > > Date : 2005-4-15 > > > > > > This document specifies a per-domain path computation method for=20 > > > computing inter-domain Traffic Engineering (TE)=20 > Multiprotocol Label=20 > > > Switching (MPLS) and Generalized MPLS > > > (GMPLS) Label Switched (LSP) paths. In this document a domain is=20 > > > referred to as a collection of network elements within a common=20 > > > sphere of address management or path computational responsibility=20 > > > such as IGP areas and Autonomous Systems. > > > > > > A URL for this Internet-Draft is: > > > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-inter-dom > > > ain-pd-path-comp-00.txt > > > > > > To remove yourself from the I-D Announcement list, send a=20 > message to=20 > > > i-d-announce-request@ietf.org with the word unsubscribe=20 > in the body=20 > > > of the message. > > > You can also visit=20 > > > https://www1.ietf.org/mailman/listinfo/I-D-announce > > > to change your subscription settings. > > > > > > > > > Internet-Drafts are also available by anonymous FTP.=20 > Login with the=20 > > > username "anonymous" and a password of your e-mail address. After=20 > > > logging in, type "cd internet-drafts" and then > > > "get draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt". > > > > > > A list of Internet-Drafts directories can be found in=20 > > > http://www.ietf.org/shadow.html or=20 > > > ftp://ftp.ietf.org/ietf/1shadow-sites.txt > > > > > > > > > Internet-Drafts can also be obtained by e-mail. > > > > > > Send a message to: > > > mailserv@ietf.org. > > > In the body type: > > > "FILE > > >=20 > /internet-drafts/draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt". > > > > > > NOTE: The mail server at ietf.org can return the document in > > > MIME-encoded form by using the "mpack" utility. To use this > > > feature, insert the command "ENCODING mime" before the "FILE" > > > command. To decode the response(s), you will need "munpack" or > > > a MIME-compliant mail reader. Different MIME-compliant mail=20 > > > readers > > > exhibit different behavior, especially when dealing with > > > "multipart" MIME messages (i.e. documents which have been split > > > up into multiple messages), so check your local documentation on > > > how to manipulate these messages. > > > > > > > > > Below is the data which will enable a MIME compliant mail reader=20 > > > implementation to automatically retrieve the ASCII version of the=20 > > > Internet-Draft. > > > > > >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 29 Jul 2005 09:08:40 +0000 Mime-Version: 1.0 (Apple Message framework v730) Content-Type: text/plain; charset=ISO-8859-1; delsp=yes; format=flowed Message-Id: <28E5C340-FC30-4A35-BB7E-97EC4AD1E0C9@cisco.com> Cc: <ccamp@ops.ietf.org>, "Arthi Ayyangar" <arthi@juniper.net>, <jpv@cisco.com> Content-Transfer-Encoding: quoted-printable From: JP Vasseur <jvasseur@cisco.com> Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Date: Fri, 29 Jul 2005 11:08:19 +0200 To: LE ROUX Jean-Louis RD-CORE-LAN <jeanlouis.leroux@francetelecom.com> Hi Jean-Louis, Thanks for the useful comments. We plan to submit a new revision =20 (including answer to Adrian's comments also) right after this IETF. Cheers JP. On Jul 28, 2005, at 10:00 AM, LE ROUX Jean-Louis RD-CORE-LAN wrote: > Hi Arthi, JP > > Thanks for this new revision that sounds really good > > I have two commments: > > -In section 4.1.2 you partially describe bidirectional LSP =20 > stitching procedure. > You mention that an Upstream Label MUST NOT be allocated by the end-=20= > to-end LSP on the LSP segment, which is OK. > But then how the LSP-Segment Egress will now that the end-to-end =20 > LSP is bidirectional? What about defining a flag in the Attributes =20 > Flags TLV of the LSP_ATTRIBUTE object so as to indicate that the =20 > LSP is bidirectional? > Also the selection of the LSP segment in case of bidirectional LSP =20 > should be detailed (e.g. If the end-to-end LSP is bidirecitonal then > the LSP-segment MUST be bidirectional. Also shall we allow that two =20= > unidrectional end-to-end LSP use the same bidirectional LSP segment =20= > (one in each direction)? > > -At the end of section 4.2.5 you mention that LSP-Segment failure =20 > or maintenance SHOULD be treated as a failure event for the end-to-=20 > end LSP. I agree for LSP-Segment failure but not for LSP-Segment =20 > maintenance. LSP-Segment maintenance should be treated as TE-link =20 > maintenance for the end-to-end LSP, and procedures defined in GMPLS =20= > graceful TE-link shutdown draft may be useful (Specific RSVP error =20 > code and TE-link attribute)... > > Hope this helps, > > Best Regards > > JL > > > >> -----Message d'origine----- >> De : owner-ccamp@ops.ietf.org >> [mailto:owner-ccamp@ops.ietf.org] De la part de >> Internet-Drafts@ietf.org >> Envoy=E9 : vendredi 15 juillet 2005 21:50 >> =C0 : i-d-announce@ietf.org >> Cc : ccamp@ops.ietf.org >> Objet : I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt >> >> A New Internet-Draft is available from the on-line >> Internet-Drafts directories. >> This draft is a work item of the Common Control and >> Measurement Plane Working Group of the IETF. >> >> Title : Label Switched Path Stitching with >> Generalized MPLS Traffic Engineering >> Author(s) : A. Ayyangar, J. Vasseur >> Filename : draft-ietf-ccamp-lsp-stitching-01.txt >> Pages : 19 >> Date : 2005-7-15 >> >> In certain scenarios, there may be a need to combine together two >> different Generalized Multi-Protocol Label Switching (GMPLS) Label >> Switched Paths (LSPs) such that in the data plane, a single end-=20= >> to- >> end (e2e) LSP is achieved and all traffic from one LSP is switched >> onto the other LSP. We will refer to this as "LSP >> stitching". This >> document covers cases where: a) the node performing the stitching >> does not require configuration of every LSP pair to be stitched >> together b) the node performing the stitching is not the egress of >> any of the LSPs c) LSP stitching not only results in an end-to-end >> LSP in the data plane, but there is also a corresponding end-to-=20= >> end >> LSP (RSVP session) in the control plane. It might be possible to >> configure a GMPLS node to switch the traffic from an LSP >> for which it >> is the egress, to another LSP for which it is the ingress, without >> requiring any signaling or routing extensions whatsoever, >> completely >> transparent to other nodes. This will also result in LSP =20 >> stitching >> in the data plane. However, this document does not cover this >> scenario of LSP stitching. >> >> A URL for this Internet-Draft is: >> http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitc >> > hing-01.txt > >> >> To remove yourself from the I-D Announcement list, send a >> message to i-d-announce-request@ietf.org with the word >> unsubscribe in the body of the message. >> You can also visit https://www1.ietf.org/mailman/listinfo/I-D-=20 >> announce >> to change your subscription settings. >> >> >> Internet-Drafts are also available by anonymous FTP. Login >> with the username "anonymous" and a password of your e-mail >> address. After logging in, type "cd internet-drafts" and then >> "get draft-ietf-ccamp-lsp-stitching-01.txt". >> >> A list of Internet-Drafts directories can be found in >> http://www.ietf.org/shadow.html or >> ftp://ftp.ietf.org/ietf/1shadow-sites.txt >> >> >> Internet-Drafts can also be obtained by e-mail. >> >> Send a message to: >> mailserv@ietf.org. >> In the body type: >> "FILE /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt". >> >> NOTE: The mail server at ietf.org can return the document in >> MIME-encoded form by using the "mpack" utility. To use this >> feature, insert the command "ENCODING mime" before the "FILE" >> command. To decode the response(s), you will need "munpack" or >> a MIME-compliant mail reader. Different MIME-compliant >> mail readers >> exhibit different behavior, especially when dealing with >> "multipart" MIME messages (i.e. documents which have been split >> up into multiple messages), so check your local documentation on >> how to manipulate these messages. >> >> >> Below is the data which will enable a MIME compliant mail reader >> implementation to automatically retrieve the ASCII version of the >> Internet-Draft. >> > Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 29 Jul 2005 09:05:51 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Date: Fri, 29 Jul 2005 11:02:17 +0200 Message-ID: <D109C8C97C15294495117745780657AE02F23511@ftrdmel1.rd.francetelecom.fr> Thread-Topic: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Thread-Index: AcWTj8tKH7/XEuqLSBuSqDAy1UCmmAAiq39A From: "LE ROUX Jean-Louis RD-CORE-LAN" <jeanlouis.leroux@francetelecom.com> To: "Arthi Ayyangar" <arthi@juniper.net> Cc: <ccamp@ops.ietf.org>, <jpv@cisco.com> Hi Arthi, Please see inline=20 > -----Message d'origine----- > De : Arthi Ayyangar [mailto:arthi@juniper.net]=20 > Envoy=E9 : jeudi 28 juillet 2005 18:17 > =C0 : LE ROUX Jean-Louis RD-CORE-LAN > Cc : ccamp@ops.ietf.org; jpv@cisco.com > Objet : RE: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt=20 >=20 > Hi JL, >=20 > Thanks for the comments. >=20 > > -In section 4.1.2 you partially describe bidirectional LSP=20 > stitching=20 > > procedure. You mention that an Upstream Label MUST NOT be=20 > allocated by=20 > > the end-to-end LSP on the LSP segment, which is OK. But=20 > then how the=20 > > LSP-Segment Egress will now that the end-to-end LSP is=20 > bidirectional? > AA--------> Excellent point. >=20 > > What about defining a flag in the Attributes Flags TLV of the=20 > > LSP_ATTRIBUTE object so as to indicate that the LSP is=20 > bidirectional? > AA------> That would be a change to the processing that GMPLS=20 > nodes use=20 > AA------> to > detect bidirectionality, isn't it ? Normally nodes look for=20 > the Upstream Label object to detect bidirectionality. >=20 > So let us say that an LSP is bidirectional if a) an Upstream=20 > Label is present or b) no Upstream Label, but bit set in=20 > LSP_ATTRIBUTE or c) both However, reliance on an e2e=20 > attributes bit set by head end, means existing head ends will=20 > not be setting this bit, so that will be an issue (wrt compatibility). JLLR: I agree this may raise some backward compatibility issues. This would require that head-ends be upgraded... >=20 > Could be nice if this signaling was just between the node=20 > doing the stitching and the end point of the LSP segment,=20 > since this is the hop that the bidirectionality information=20 > is lost. Let me think about this. JLLR: Yes, and what about defining a new MPLS label value, let say label = 4 =3D "No label assigned", with a semantic distinct from label 3 = semantic (=3Dlabel pop). The UPSTREAM label object with this new label = value would be included in the end-to-end Path message over the = LSP-Segment... Let discuss this point during Paris meeting, Regards, JL >=20 >=20 > > Also the selection of the LSP segment in case of bidirectional LSP=20 > > should be detailed (e.g. If the end-to-end LSP is=20 > bidirecitonal then=20 > > the LSP-segment MUST be bidirectional. Also shall we allow that two=20 > > unidrectional end-to-end LSP use the same bidirectional LSP segment=20 > > (one in each direction)? > AA----> Yes, this should be okay. IMO. >=20 > > -At the end of section 4.2.5 you mention that LSP-Segment=20 > failure or=20 > > maintenance SHOULD be treated as a failure event for the=20 > end-to-end LSP. > > I agree for LSP-Segment failure but not for LSP-Segment maintenance. > > LSP-Segment maintenance should be treated as TE-link=20 > maintenance for=20 > > the end-to-end LSP, and procedures defined in GMPLS=20 > graceful TE-link=20 > > shutdown draft may be useful (Specific RSVP error code and TE-link=20 > > attribute)... > AA---> Yes, I agree; we will mention this. Actually the graceful=20 > AA---> teardown > sentence occurs few paragraphs before, but not in the right=20 > context. So we will clarify the above. >=20 > > Hope this helps, > AA---> Sure does. >=20 >=20 > Thanks! >=20 > -arthi >=20 > > > > > > > -----Message d'origine----- > > > De : owner-ccamp@ops.ietf.org > > > [mailto:owner-ccamp@ops.ietf.org] De la part de=20 > > > Internet-Drafts@ietf.org Envoy=E9 : vendredi 15 juillet=20 > 2005 21:50 =C0 :=20 > > > i-d-announce@ietf.org Cc : ccamp@ops.ietf.org Objet : I-D=20 > > > ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > > > > > > A New Internet-Draft is available from the on-line=20 > Internet-Drafts=20 > > > directories. > > > This draft is a work item of the Common Control and Measurement=20 > > > Plane Working Group of the IETF. > > > > > > Title : Label Switched Path Stitching with > > > Generalized MPLS Traffic Engineering > > > Author(s) : A. Ayyangar, J. Vasseur > > > Filename : draft-ietf-ccamp-lsp-stitching-01.txt > > > Pages : 19 > > > Date : 2005-7-15 > > > > > > In certain scenarios, there may be a need to combine together two > > > different Generalized Multi-Protocol Label Switching=20 > (GMPLS) Label > > > Switched Paths (LSPs) such that in the data plane, a=20 > single end-to- > > > end (e2e) LSP is achieved and all traffic from one LSP=20 > is switched > > > onto the other LSP. We will refer to this as "LSP=20 > stitching". =20 > > > This > > > document covers cases where: a) the node performing=20 > the stitching > > > does not require configuration of every LSP pair to be stitched > > > together b) the node performing the stitching is not=20 > the egress of > > > any of the LSPs c) LSP stitching not only results in=20 > an end-to-end > > > LSP in the data plane, but there is also a=20 > corresponding end-to-end > > > LSP (RSVP session) in the control plane. It might be=20 > possible to > > > configure a GMPLS node to switch the traffic from an LSP for=20 > > > which it > > > is the egress, to another LSP for which it is the=20 > ingress, without > > > requiring any signaling or routing extensions whatsoever,=20 > > > completely > > > transparent to other nodes. This will also result in=20 > LSP stitching > > > in the data plane. However, this document does not cover this > > > scenario of LSP stitching. > > > > > > A URL for this Internet-Draft is: > > > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitc > > hing-01.txt > > > > > > To remove yourself from the I-D Announcement list, send a=20 > message to=20 > > > i-d-announce-request@ietf.org with the word unsubscribe=20 > in the body=20 > > > of the message. > > > You can also visit=20 > > > https://www1.ietf.org/mailman/listinfo/I-D-announce > > > to change your subscription settings. > > > > > > > > > Internet-Drafts are also available by anonymous FTP.=20 > Login with the=20 > > > username "anonymous" and a password of your e-mail address. After=20 > > > logging in, type "cd internet-drafts" and then > > > "get draft-ietf-ccamp-lsp-stitching-01.txt". > > > > > > A list of Internet-Drafts directories can be found in=20 > > > http://www.ietf.org/shadow.html or=20 > > > ftp://ftp.ietf.org/ietf/1shadow-sites.txt > > > > > > > > > Internet-Drafts can also be obtained by e-mail. > > > > > > Send a message to: > > > mailserv@ietf.org. > > > In the body type: > > > "FILE /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt". > > > > > > NOTE: The mail server at ietf.org can return the document in > > > MIME-encoded form by using the "mpack" utility. To use this > > > feature, insert the command "ENCODING mime" before the "FILE" > > > command. To decode the response(s), you will need "munpack" or > > > a MIME-compliant mail reader. Different MIME-compliant mail=20 > > > readers > > > exhibit different behavior, especially when dealing with > > > "multipart" MIME messages (i.e. documents which have been split > > > up into multiple messages), so check your local documentation on > > > how to manipulate these messages. > > > > > > > > > Below is the data which will enable a MIME compliant mail reader=20 > > > implementation to automatically retrieve the ASCII version of the=20 > > > Internet-Draft. > > > > > >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 29 Jul 2005 07:33:39 +0000 From: "Adrian Farrel" <olddog@clara.co.uk> To: ccamp@ops.ietf.org Cc: sob@harvard.edu, zinin@psg.com, fenner@research.att.com Reply-To: adrian@olddog.co.uk Subject: Proposed liaison to SG15 on RFC 4139 Date: Fri, 29 Jul 2005 08:32:54 +0100 Mime-Version: 1.0 Content-Type: text/plain; format=flowed; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Message-Id: <E1DyPMV-000PNO-57@oceanus.uk.clara.net> Hi, I propose to send this liaison to Study Group 15 of the ITU-T. Comments please. Thanks, Adrian ======= To: ITU-T Study Group 15 From: IETF CCAMP Working Group Subject: New Request for Comment Published For: Information The CCAMP Working Group of the IETF wishes to inform Study Group 15 of the ITU-T of the publication of a new Request for Comment. RFC 4139 "Requirements for Generalized MPLS (GMPLS) Signaling Usage and Extensions for Automatically Switched Optical Network (ASON)" is available from the RFC Editor's repository at ftp://ftp.rfc-editor.org/in-notes/rfc4139.txt The Abstract of this RFC reads as follows. The Generalized Multi-Protocol Label Switching (GMPLS) suite of protocols has been defined to control different switching technologies and different applications. These include support for requesting Time Division Multiplexing (TDM) connections, including Synchronous Optical Network (SONET)/Synchronous Digital Hierarchy (SDH) and Optical Transport Networks (OTNs). This document concentrates on the signaling aspects of the GMPLS suite of protocols. It identifies the features to be covered by the GMPLS signaling protocol to support the capabilities of an Automatically Switched Optical Network (ASON). This document provides a problem statement and additional requirements for the GMPLS signaling protocol to support the ASON functionality. The CCAMP Working Group would like to thank the members and Questions of Study Group 15 for the helpful input and review feedback that they provided while this document was being developed. Regards, Adrian Farrel and Kireeti Kompella CCAMP Working Group Co-chairs Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 29 Jul 2005 07:33:27 +0000 From: "Adrian Farrel" <olddog@clara.co.uk> To: ccamp@ops.ietf.org Reply-To: adrian@olddog.co.uk Subject: New CCAMP RFC Date: Fri, 29 Jul 2005 08:31:24 +0100 Mime-Version: 1.0 Content-Type: text/plain; format=flowed; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Message-Id: <E1DyPL2-000PHp-RZ@oceanus.uk.clara.net> Thanks to everyone involved. Not quite sure how this RFC jumped the queue, but good to have it published. Adrian ----- Original Message ----- From: <rfc-editor@rfc-editor.org> To: <ietf-announce@ietf.org> Cc: <ccamp@ops.ietf.org>; <rfc-editor@rfc-editor.org> Sent: Wednesday, July 27, 2005 11:37 PM Subject: RFC 4139 on Requirements for Generalized MPLS (GMPLS) SignalingUsage and Extensions for Automatically Switched OpticalNetwork (ASON) > > A new Request for Comments is now available in online RFC libraries. > > > RFC 4139 > > Title: Requirements for Generalized MPLS (GMPLS) > Signaling Usage and Extensions for Automatically > Switched Optical Network (ASON) > Author(s): D. Papadimitriou, J. Drake, J. Ash, A. Farrel, > L. Ong > Status: Informational > Date: July 2005 > Mailbox: dimitri.papadimitriou@alcatel.be, > John.E.Drake2@boeing.com, gash@att.com, > adrian@olddog.co.uk, lyong@ciena.com > Pages: 16 > Characters: 36660 > Updates/Obsoletes/SeeAlso: None > > I-D Tag: draft-ietf-ccamp-gmpls-ason-reqts-07.txt > > URL: ftp://ftp.rfc-editor.org/in-notes/rfc4139.txt > > > The Generalized Multi-Protocol Label Switching (GMPLS) suite of > protocols has been defined to control different switching > technologies and different applications. These include support > for requesting Time Division Multiplexing (TDM) connections, including > Synchronous Optical Network (SONET)/Synchronous Digital Hierarchy > (SDH) and Optical Transport Networks (OTNs). > > This document concentrates on the signaling aspects of the GMPLS > suite of protocols. It identifies the features to be covered by the > GMPLS signaling protocol to support the capabilities of an > Automatically Switched Optical Network (ASON). This document provides > a problem statement and additional requirements for the GMPLS > signaling protocol to support the ASON functionality. > > This document is a product of the Common Control and Measurement Plane > Working Group of the IETF. > > This memo provides information for the Internet community. It does > not specify an Internet standard of any kind. Distribution of this > memo is unlimited. > > This announcement is sent to the IETF list and the RFC-DIST list. > Requests to be added to or deleted from the IETF distribution list > should be sent to IETF-REQUEST@IETF.ORG. Requests to be > added to or deleted from the RFC-DIST distribution list should > be sent to RFC-DIST-REQUEST@RFC-EDITOR.ORG. > > Details on obtaining RFCs via FTP or EMAIL may be obtained by sending > an EMAIL message to rfc-info@RFC-EDITOR.ORG with the message body > help: ways_to_get_rfcs. For example: > > To: rfc-info@RFC-EDITOR.ORG > Subject: getting rfcs > > help: ways_to_get_rfcs > > Requests for special distribution should be addressed to either the > author of the RFC in question, or to RFC-Manager@RFC-EDITOR.ORG. Unless > specifically noted otherwise on the RFC itself, all RFCs are for > unlimited distribution. > > Submissions for Requests for Comments should be sent to > RFC-EDITOR@RFC-EDITOR.ORG. Please consult RFC 2223, Instructions to RFC > Authors, for further information. > > > Joyce K. Reynolds and Sandy Ginoza > USC/Information Sciences Institute Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 29 Jul 2005 01:20:15 +0000 Message-ID: <42E983B8.6000309@jp.fujitsu.com> Date: Fri, 29 Jul 2005 10:17:44 +0900 From: "K. Miyazaki" <miyazaki.keiji@jp.fujitsu.com> User-Agent: Mozilla Thunderbird 1.0.2 (Windows/20050317) MIME-Version: 1.0 To: ccamp@ops.ietf.org Subject: Re: draft-otani-ccamp-interas-gmpls-te-03.txt Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Hi everyone, I think following draft is important requirement in GMPLS inter-domain and it should be WG document asap. Regards, Keiji Miyazaki Tomohiro Otani wrote: > Hi everyone, > > We updated the GMPLS inter-domain TE req. draft. Since last meeting, > we believe this document has been already stable and well-recognized, > however, per the last meeting discussion, we have to wait for the > charter being revised, in order to be a WG document (This must be > initiated by chairs). > > Until then, we continue to update the contents and indeed welcome your > feedback and comments through this list or during the meeting in Paris. > > Regards, > > tomo > > > > > ----------------------------------------------------------------------------- > A New Internet-Draft is available from the on-line Internet-Drafts directories. > > > Title : GMPLS Inter-domain Traffic Engineering Requirements > Author(s) : T. Otani, et al. > Filename : draft-otani-ccamp-interas-gmpls-te-03.txt > Pages : 15 > Date : 2005-7-20 > > This draft provides requirements for the support of generalized > multi-protocol label switching (GMPLS) inter-domain traffic > engineering (TE). Its main objective is to present the differences > between MPLS inter-domain TE and GMPLS inter-domain TE. This draft > covers not only GMPLS Inter-domain architecture but also functional > requirements in terms of GMPLS signaling and routing in order to > specify these in a GMPLS Inter-domain environment. > > A URL for this Internet-Draft is: > http://www.ietf.org/internet-drafts/draft-otani-ccamp-interas-gmpls-te-03.txt > > To remove yourself from the I-D Announcement list, send a message to > i-d-announce-request at ietf.org with the word unsubscribe in the body of the message. > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > to change your subscription settings. > > > Internet-Drafts are also available by anonymous FTP. Login with the username > "anonymous" and a password of your e-mail address. After logging in, > type "cd internet-drafts" and then > "get draft-otani-ccamp-interas-gmpls-te-03.txt". > > A list of Internet-Drafts directories can be found in > http://www.ietf.org/shadow.html > or ftp://ftp.ietf.org/ietf/1shadow-sites.txt > Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 28 Jul 2005 22:55:03 +0000 Mime-Version: 1.0 (Apple Message framework v730) Content-Type: text/plain; charset=ISO-8859-1; delsp=yes; format=flowed Message-Id: <566BB9E4-2104-41CE-B7EA-B36BF5EFDE03@cisco.com> Cc: Adrian Farrel <adrian@olddog.co.uk>, raymond.zhang@bt.infonet.com, ccamp@ops.ietf.org, Arthi Ayyangar <arthi@juniper.net> Content-Transfer-Encoding: quoted-printable From: JP Vasseur <jvasseur@cisco.com> Subject: Re: Early chair review of draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt Date: Fri, 29 Jul 2005 00:53:23 +0200 To: Adrian Farrel <adrian@olddog.co.uk> Thanks for your comments Adrian. I hadn't missed it, just did not =20 find the time to reply yet ... More after this IETF. Thanks. JP. On Jul 28, 2005, at 9:18 PM, Arthi Ayyangar wrote: > Hi Adrian, > > Thanks for your exhaustive review and comments. > Looks like I missed this mail. Sorry about that. > > Please see inline for my replies to some questions/issues raised. For > formatting issues, I will turn to JP so that we can address them in =20= > the > next rev. Thanks! > > >> I have not caught all of the typos, I am sure. Also worried that the >> number of changes needed may have obscured some technical issues, =20 >> so we >> will have to go around this at least once more. >> > AA-----> Fair enough. > > >> ## The next paragraph is a bit unclear. Why not it with... >> ## Per-domain computation applies where the full route to an >> ## inter-domain LSP cannot be or is not determined at the ingress >> ## point of the LSP, and is not signaled across domain boundaries. >> ## This is most likely to arrise owing to TE visibilty limitations. >> ## The signaling message indicates the destination and nodes up to >> ## the next domain boundary. It may also indicate further domain >> ## boundaries or domain identifiers. The route through each domain, >> ## possibly including the choice of exit point from the domain, >> ## must be determined within the domain. >> > AA-------> Sounds better to me. > > >> The principle of per-domain path computation specified in this =20 >> document >> comprises of a GMPLS or MPLS node along an inter-domain LSP path, >> computing a path up to the last reachable hop within its domain. It >> covers cases where this last hop (domain exit point) may already be >> specified in the signaling message as well the case where this =20 >> last hop >> may need to be determined by the GMPLS node. >> >> > > >> ABR Routers: routers used to connect two IGP areas (areas in OSPF or >> L1/L2 in IS-IS) >> ##replace "L1/L2" with "levels" >> > AA----> Yep. > > >> ## Missing ASBR. Just point to Interconnect routers >> > AA----> Okay. > > > >> Boundary LSR: a boundary LSR is either an ABR in the context of =20 >> inter- >> area MPLS TE or an ASBR in the context of inter-AS MPLS TE. >> >> Bypass Tunnel: an LSP that is used to protect a set of LSPs passing >> over a common facility. >> >> CSPF: Constraint-based Shortest Path First. >> ## I searched for the uses of this acronym and I thought all of them >> ## should be less specific (that is allowing any path computation >> ## technique). In which case, you will not need this in the =20 >> terminology >> ## section. >> > AA---> That is true. > > <snipped> > >> >> ## The next paragpraph is true but irrelevant to this I-D. >> > AA---> Agree. > >> There are different ways in which inter-domain TE LSP path =20 >> computation >> may be performed. For example, if the requirement is to get an end-=20= >> to- >> end constraint-based shortest path across multiple domains, then a >> mechanism using one or more distributed PCEs could be used to compute >> the shortest path across different domains. Alternatively, one could >> also use some static or discovery mechanisms to determine the next >> boundary LSR per domain as the inter-domain TE LSP is being signaled. >> Other offline mechanisms for path computation are not precluded =20 >> either. >> Depending on the Service Provider=C6s requirements, one may adopt =20 >> either >> of these techniques for inter-domain path computation. >> >> > <snipped> > > >> ## Why do you mention CSPF? The path computation algorithm in use is >> ## surely not relevant to this I-D. >> > AA----> Agree. > > >> Note that PCE-based path computation may be mentioned in this =20 >> document >> for the sake of reference but such techniques are outside the =20 >> scope of >> this document. >> ## Then don't mention them. >> > AA----> :-) > > >> ## Do you need to say that you also support the case where area2 is >> ## not a spearate area, but is part of area 1? >> ## >> ## Do you support routing across (i.e. into and out of) an area >> ## that is not area zero? >> > AA----> Don't see why not. > > >> - The various ASBRs are BGP peers, without any IGP running on the >> single hop links interconnecting the ASBRs and also referred to as >> inter-ASBR links, >> ## Whoah! I hope you will explain why you *need* to run BGP, and >> ## how you will manage non-TE distribution of reachability >> ## information when the destination of a TE-LSP is a TE address. >> ## Will you perhaps be mandating (stronger than recommending) that >> ## the destination of an inter-AS TE LSP is the TE Router ID of >> ## the egress in order to be sure that this address is one of the >> ## reachable addresses advertised by BGP? >> > AA--------> See below. > > > >> hop at least has IP reachability (via IGP or BGP). If the next-hop is >> not reachable, then the path computation stops and the LSR sends =20 >> back a >> PathErr upstream. If the next-hop is reachable, then it finds an >> ABR/ASBR to get to the next-hop. In the absence of an auto-discovery >> mechanism, the ABR in the case of inter-area TE or the ASBR in the >> next-hop AS in the case of inter-AS TE should be the loose next-=20 >> hop in >> the ERO and hence should be accessible via the TED, otherwise the =20 >> path >> computation for the inter-domain TE LSP will fail. >> ## I would like you to make it VERY clear what you are doing here. >> ## You are using the Routing Database to make TE routing decisions. >> ## This may (or may not) be OK in people's minds >> ## I will send separate mail to the CCAMP list because this is a BIG >> ## DEAL in GMPLS networks. >> > AA---> I noticed the mail that you sent out. Thanks! However, there =20= > have > been no replies yet. But I think this ID does state that that if =20 > the next > hop is not in the TED and does not have IP reachability, then such a > computation would fail. Should we raise this point again either at the > WG meeting or when we post the next rev ? > > >> ## In any case, can you please clarify that this process MUST NOT be >> ## used when the next-hop is within the same domain but does not >> ## appear in the TED, or does not have TE connectivity available. >> > AA-------> Okay, this will be clarified. > > <snipped> > >> or LSP segment(s) already exist, then it SHOULD try to >> select from among those intra-area/AS LSPs. Depending >> on local policy, it MAY signal a new FA-LSP/LSP segment >> if this selection fails. If the FA-LSP/LSP segment is >> successfully signaled or selected, it propagates the >> inter-domain Path message to the next-hop following the >> procedures described in [LSP-HIER]. If, for some reason >> the dynamic FA-LSP/LSP segment setup to the next-hop >> boundary LSR fails, the path computation stops and a >> ## Is it the path computation that stops? >> ## Is there no scope for retires or fall-back to ordinary routing? >> > AA-------> This is the case which JL mentioned as well in his email. > There could be multiple exit points and if one fails, then another =20 > could > be tried based on routing information and local policies. So I =20 > agree that > a retry is possible and if all attempts fail then an error is =20 > propagated. > > > <snipped> > >> 4.1. Example with an inter-area TE LSP >> >> 4.1.1. Case 1: T1 is a contiguous TE LSP >> >> ## I think your example should start at the ingress! It seems you >> ## cut back to that after the first paragraph. >> > AA---> Okay. > > > >> ## What does the user supply? What does the ingress compute? >> ## How is ABR1 selected? (I like ABR2: it is a nicer color. Why >> ## can't I have ABR2?) >> ## There is also some repeated text in this section. >> ## Also, why are you considering Example 1 since it is out of scope >> ## for you? >> > AA--------> Not clear to me what is out of scope and why ? > > >> ## Presumably, Example 3 only applies if areas 1 and 2 are actually >> ## part of the same area, so you should say so. >> >> When the path message reaches ABR1, it first determines the egress =20= >> LSR >> from its area 0 along the LSP path (say ABR=C61), either directly = from >> the ERO (if for example the next hop ABR is specified as a loose =20 >> hop in >> the ERO) or by using some constraint-aware auto-discovery =20 >> mechanism. In >> ## Wow! "A constraint-aware auto-discovery mechanism." I borrowed >> ## one of those from Batman once, but I could never get it to work. >> > AA---------> :-) > > >> ## What are you talking about? >> ## Yes, I see this in section 3, 1) and in section 4, 1) but it does >> ## not explain what you are suggesting may exist. I must say it =20 >> sounds >> ## suspisciously like TE aggregation advertisement using BGP. >> > AA-------> I don't think "constraint-aware" stuff is relevant to this > document and I am okay with getting rid of it. We need discovery of =20= > exit > points...but that's about it. > > > <snipped> > >> >> ## Can you explain why the TE information pertaining to the inter-AS >> ## links *needs* to be flooded within the ASs? >> > AA-----> Hmm...I don't think we say that TE information for =20 > inter_AS links > *needs* to be flooded. See below for answer to why ? > > It is surely enough that > >> ## the ASBRs have access to the TE information about their own TE =20 >> links. >> > AA----> Yes they do, but the previous hop LSR in this AS that picked > this ASBR as the exit point doesn't have that information. So, if =20 > the LSR > selecting this ASBR had that information (and used it in path > computation), then if the ASBR-ASBR link was a bottleneck, then it =20 > could > have chosen another exit ASBR instead. > > >> ## After all, you have defined that the ASBRs in the *next* AS can do >> ## hop resolution and crankback. Why should the ASBRs in *this* AS =20= >> also >> ## not perform that operation if necessary? >> > AA---> Sure it can. > > >> ## You are reducing the chance of call setup failure at the =20 >> expense of >> ## increasing the scope of the TE information flooded. But we know >> ## how that works! >> > AA-----> How many ASBR-ASBR links are we talking about ? > > <snipped> > >> 4.2.2. Case 2: T1 is a stitched or nested TE LSP >> >> The signaling procedures are very similar to the inter-area LSP setup >> case described earlier. In this case, the FA-LSPs or LSP segments =20 >> will >> only be originated by the ASBRs at the entry to the AS. >> ## This is a really important point that you should bring out a bit! >> ## An FA cannot exist out of the IGP instance that contains its >> ## component TE links. >> ## What about non-FAs? Could I operate a stitched segment for just =20= >> the >> ## link between ASBRs? I think so. Indeed, I might want to. >> ## Could I run a hierarchical LSP for just the link between ASBRs? =20= >> Yes >> ## although I am not convinced it adds value, but maybe there is some >> ## administrative benefit when crossing a trust boundary (for =20 >> example, >> ## all I have to do is count the packets with one label). >> > AA-------> Is your point that FA-LSPs and LSP segments can =20 > originate and > terminate on any a) boundary node or b) any other node that has local > configuration/policy to do so (for originating) ? If yes, then I =20 > agree. I > think this ID needs to be fixed to explicitly state that. > > <snipped> > >> >> 6. MPLS Traffic Engineering Fast Reroute for inter-domain TE LSPs >> ## The following section is valuable in the context of the popularity >> ## of RFC 4090, but you cannot produce this I-D in CCAMP without also >> ## discussing the use of Segment Protection to achieve the same >> ## function in the data plane. >> > AA-------> Agree. > > > <snipped> > >> 7.1. Contiguous TE LSPs >> >> ## Need to generalize this section for areas, etc. >> >> After an inter-AS TE LSP has been set up, a more optimal route might >> appear in the various traversed ASes. Then in this case, it is >> desirable to get the ability to reroute an inter-AS TE LSP in a non- >> disruptive fashion (making use of the so called Make Before Break >> procedure) to follow this more optimal path. This is a known as a TE >> LSP reoptimization procedure. >> >> ## Is there any requirement to redescribe the procedures of >> ## [LOOSE-PATH-REOPT]? Why not simply state that... >> ## [LOOSE-PATH-REOPT] describes a mechanism to control discovery >> ## or more optimal paths, and to signal new paths using make before >> ## break. >> > AA----> Nope, no need at all. > > >> #### BEGIN DELETE >> # >> # >> >> [LOOSE-REOPT] proposes a mechanisms allowing: >> ## s/[LOOSE-REOPT]/[LOOSE-PATH-REOPT]/ -- change throughout the doc >> ## s/proposes/defines/ >> >> >> Vasseur, Ayyangar and Zhang 14 >> >> >> >> draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt April =20= >> 2005 >> >> >> - The head-end LSR to trigger on every LSR whose next hop is a >> loose hop the re evaluation of the current path in order to >> ## s/re evaluation/re-evaluation/ >> detect a potentially more optimal path. This is done via >> explicit signaling request: the head-end LSR sets the =F4ERO >> Expansion request=F6 bit of the SESSION-ATTRIBUTE object = carried >> in the RSVP Path message. >> >> - An LSR whose next hop is a loose-hop to signal to the head- >> end LSR that a better path exists. This is performed by sending >> an RSVP Path Error Notify message (ERROR-CODE =3D 25), sub-code = 6 >> (Better path exists). >> >> This indication may be sent either: >> >> - In response to a query sent by the head-end LSR, >> - Spontaneously by any LSR having detected a more >> optimal path >> >> Such a mechanism allows for the reoptimization of a TE LSP if and =20 >> only >> if a better path is some downstream area/AS is detected. >> >> The reoptimization event can either be timer or event-driven based (a >> link UP event for instance). >> >> Note that the reoptimization MUST always be performed in a non- >> disruptive fashion. >> >> Once the head-end LSR is informed of the existence of a more optimal >> path either in its head-end area/AS or in another AS, the inter-AS TE >> Path computation is triggered using the same set of mechanisms as =20 >> when >> the TE LSP is first set up. Then the inter-AS TE LSP is set up >> following the more optimal path, making use of the make before break >> procedure. In case of a contiguous LSP, the reoptimization process is >> strictly controlled by the head-end LSR which triggers the make-=20 >> before- >> break procedure, regardless of the location where the more optimal =20= >> path >> is. >> >> Note that in the case of loose hop reoptimization, the TE LSP may >> follow a preferable path within one or more domain(s) whereas in the >> case of PCE-based path computation techniques, the reoptimization >> process may lead to following a completely different inter-domain =20 >> path >> (including a different set of ABRs/ASBRs) since end-to-end shortest >> path is computed. >> >> # >> # >> #### END DELETE >> >> 7.2. Stitched or nested (non-contiguous) TE LSPs >> >> ## This section contains quite a lot of repetition. >> >> In the case of a stitched or nested inter-domain TE LSP, the re- >> optimization process is treated as a local matter to any Area/AS. The >> main reason is that the inter-domain TE LSP is a different LSP (and >> therefore different RSVP session) from the intra-domain LSP =20 >> segment or >> >> ## Your causality is broken. Being a different LSP does not imply >> ## being a different session. However, the key point *is* that this >> ## is a different session (with different ingress and egress). >> > AA-----> We'll simply highlight that this is a different session. > > <snipped> > > >> 8. Security Considerations >> >> ## This section seems very lightweight given the importance of >> ## inter-AS trust boundaries. >> > AA----> I agree. We should add details about policy control, trust > relationship; etc . They may be an overlap with the signaling > ID, but that's okay. Also, need to highlight how a boundary node =20 > (or any > other node) would handle the case where it is being overwhelmed =20 > with path > computation requests for an inter-domain LSP from some other domain. > > thanks, > -arthi > Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 28 Jul 2005 19:20:58 +0000 Date: Thu, 28 Jul 2005 12:18:49 -0700 (PDT) From: Arthi Ayyangar <arthi@juniper.net> To: Adrian Farrel <adrian@olddog.co.uk> cc: "'Jean Philippe Vasseur'" <jvasseur@cisco.com>, raymond.zhang@bt.infonet.com, ccamp@ops.ietf.org Subject: Re: Early chair review of draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt Message-ID: <20050728103500.V12097@zircon.juniper.net> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=X-UNKNOWN Content-Transfer-Encoding: QUOTED-PRINTABLE Hi Adrian, Thanks for your exhaustive review and comments. Looks like I missed this mail. Sorry about that. Please see inline for my replies to some questions/issues raised. For formatting issues, I will turn to JP so that we can address them in the next rev. Thanks! > I have not caught all of the typos, I am sure. Also worried that the > number of changes needed may have obscured some technical issues, so we > will have to go around this at least once more. AA-----> Fair enough. > ## The next paragraph is a bit unclear. Why not it with... > ## Per-domain computation applies where the full route to an > ## inter-domain LSP cannot be or is not determined at the ingress > ## point of the LSP, and is not signaled across domain boundaries. > ## This is most likely to arrise owing to TE visibilty limitations. > ## The signaling message indicates the destination and nodes up to > ## the next domain boundary. It may also indicate further domain > ## boundaries or domain identifiers. The route through each domain, > ## possibly including the choice of exit point from the domain, > ## must be determined within the domain. AA-------> Sounds better to me. > The principle of per-domain path computation specified in this document > comprises of a GMPLS or MPLS node along an inter-domain LSP path, > computing a path up to the last reachable hop within its domain. It > covers cases where this last hop (domain exit point) may already be > specified in the signaling message as well the case where this last hop > may need to be determined by the GMPLS node. > > ABR Routers: routers used to connect two IGP areas (areas in OSPF or > L1/L2 in IS-IS) > ##replace "L1/L2" with "levels" AA----> Yep. > ## Missing ASBR. Just point to Interconnect routers AA----> Okay. > Boundary LSR: a boundary LSR is either an ABR in the context of inter- > area MPLS TE or an ASBR in the context of inter-AS MPLS TE. > > Bypass Tunnel: an LSP that is used to protect a set of LSPs passing > over a common facility. > > CSPF: Constraint-based Shortest Path First. > ## I searched for the uses of this acronym and I thought all of them > ## should be less specific (that is allowing any path computation > ## technique). In which case, you will not need this in the terminology > ## section. AA---> That is true. <snipped> > > ## The next paragpraph is true but irrelevant to this I-D. AA---> Agree. > There are different ways in which inter-domain TE LSP path computation > may be performed. For example, if the requirement is to get an end-to- > end constraint-based shortest path across multiple domains, then a > mechanism using one or more distributed PCEs could be used to compute > the shortest path across different domains. Alternatively, one could > also use some static or discovery mechanisms to determine the next > boundary LSR per domain as the inter-domain TE LSP is being signaled. > Other offline mechanisms for path computation are not precluded either. > Depending on the Service Provider=C6s requirements, one may adopt either > of these techniques for inter-domain path computation. > <snipped> > ## Why do you mention CSPF? The path computation algorithm in use is > ## surely not relevant to this I-D. AA----> Agree. > Note that PCE-based path computation may be mentioned in this document > for the sake of reference but such techniques are outside the scope of > this document. > ## Then don't mention them. AA----> :-) > ## Do you need to say that you also support the case where area2 is > ## not a spearate area, but is part of area 1? > ## > ## Do you support routing across (i.e. into and out of) an area > ## that is not area zero? AA----> Don't see why not. > - The various ASBRs are BGP peers, without any IGP running on the > single hop links interconnecting the ASBRs and also referred to as > inter-ASBR links, > ## Whoah! I hope you will explain why you *need* to run BGP, and > ## how you will manage non-TE distribution of reachability > ## information when the destination of a TE-LSP is a TE address. > ## Will you perhaps be mandating (stronger than recommending) that > ## the destination of an inter-AS TE LSP is the TE Router ID of > ## the egress in order to be sure that this address is one of the > ## reachable addresses advertised by BGP? AA--------> See below. > hop at least has IP reachability (via IGP or BGP). If the next-hop is > not reachable, then the path computation stops and the LSR sends back a > PathErr upstream. If the next-hop is reachable, then it finds an > ABR/ASBR to get to the next-hop. In the absence of an auto-discovery > mechanism, the ABR in the case of inter-area TE or the ASBR in the > next-hop AS in the case of inter-AS TE should be the loose next-hop in > the ERO and hence should be accessible via the TED, otherwise the path > computation for the inter-domain TE LSP will fail. > ## I would like you to make it VERY clear what you are doing here. > ## You are using the Routing Database to make TE routing decisions. > ## This may (or may not) be OK in people's minds > ## I will send separate mail to the CCAMP list because this is a BIG > ## DEAL in GMPLS networks. AA---> I noticed the mail that you sent out. Thanks! However, there have been no replies yet. But I think this ID does state that that if the next hop is not in the TED and does not have IP reachability, then such a computation would fail. Should we raise this point again either at the WG meeting or when we post the next rev ? > ## In any case, can you please clarify that this process MUST NOT be > ## used when the next-hop is within the same domain but does not > ## appear in the TED, or does not have TE connectivity available. AA-------> Okay, this will be clarified. <snipped> > or LSP segment(s) already exist, then it SHOULD try to > select from among those intra-area/AS LSPs. Depending > on local policy, it MAY signal a new FA-LSP/LSP segment > if this selection fails. If the FA-LSP/LSP segment is > successfully signaled or selected, it propagates the > inter-domain Path message to the next-hop following the > procedures described in [LSP-HIER]. If, for some reason > the dynamic FA-LSP/LSP segment setup to the next-hop > boundary LSR fails, the path computation stops and a > ## Is it the path computation that stops? > ## Is there no scope for retires or fall-back to ordinary routing? AA-------> This is the case which JL mentioned as well in his email. There could be multiple exit points and if one fails, then another could be tried based on routing information and local policies. So I agree that a retry is possible and if all attempts fail then an error is propagated. <snipped> > 4.1. Example with an inter-area TE LSP > > 4.1.1. Case 1: T1 is a contiguous TE LSP > > ## I think your example should start at the ingress! It seems you > ## cut back to that after the first paragraph. AA---> Okay. > ## What does the user supply? What does the ingress compute? > ## How is ABR1 selected? (I like ABR2: it is a nicer color. Why > ## can't I have ABR2?) > ## There is also some repeated text in this section. > ## Also, why are you considering Example 1 since it is out of scope > ## for you? AA--------> Not clear to me what is out of scope and why ? > ## Presumably, Example 3 only applies if areas 1 and 2 are actually > ## part of the same area, so you should say so. > > When the path message reaches ABR1, it first determines the egress LSR > from its area 0 along the LSP path (say ABR=C61), either directly from > the ERO (if for example the next hop ABR is specified as a loose hop in > the ERO) or by using some constraint-aware auto-discovery mechanism. In > ## Wow! "A constraint-aware auto-discovery mechanism." I borrowed > ## one of those from Batman once, but I could never get it to work. AA---------> :-) > ## What are you talking about? > ## Yes, I see this in section 3, 1) and in section 4, 1) but it does > ## not explain what you are suggesting may exist. I must say it sounds > ## suspisciously like TE aggregation advertisement using BGP. AA-------> I don't think "constraint-aware" stuff is relevant to this document and I am okay with getting rid of it. We need discovery of exit points...but that's about it. <snipped> > > ## Can you explain why the TE information pertaining to the inter-AS > ## links *needs* to be flooded within the ASs? AA-----> Hmm...I don't think we say that TE information for inter_AS links *needs* to be flooded. See below for answer to why ? It is surely enough that > ## the ASBRs have access to the TE information about their own TE links. AA----> Yes they do, but the previous hop LSR in this AS that picked this ASBR as the exit point doesn't have that information. So, if the LSR selecting this ASBR had that information (and used it in path computation), then if the ASBR-ASBR link was a bottleneck, then it could have chosen another exit ASBR instead. > ## After all, you have defined that the ASBRs in the *next* AS can do > ## hop resolution and crankback. Why should the ASBRs in *this* AS also > ## not perform that operation if necessary? AA---> Sure it can. > ## You are reducing the chance of call setup failure at the expense of > ## increasing the scope of the TE information flooded. But we know > ## how that works! AA-----> How many ASBR-ASBR links are we talking about ? <snipped> > 4.2.2. Case 2: T1 is a stitched or nested TE LSP > > The signaling procedures are very similar to the inter-area LSP setup > case described earlier. In this case, the FA-LSPs or LSP segments will > only be originated by the ASBRs at the entry to the AS. > ## This is a really important point that you should bring out a bit! > ## An FA cannot exist out of the IGP instance that contains its > ## component TE links. > ## What about non-FAs? Could I operate a stitched segment for just the > ## link between ASBRs? I think so. Indeed, I might want to. > ## Could I run a hierarchical LSP for just the link between ASBRs? Yes > ## although I am not convinced it adds value, but maybe there is some > ## administrative benefit when crossing a trust boundary (for example, > ## all I have to do is count the packets with one label). AA-------> Is your point that FA-LSPs and LSP segments can originate and terminate on any a) boundary node or b) any other node that has local configuration/policy to do so (for originating) ? If yes, then I agree. I think this ID needs to be fixed to explicitly state that. <snipped> > > 6. MPLS Traffic Engineering Fast Reroute for inter-domain TE LSPs > ## The following section is valuable in the context of the popularity > ## of RFC 4090, but you cannot produce this I-D in CCAMP without also > ## discussing the use of Segment Protection to achieve the same > ## function in the data plane. AA-------> Agree. <snipped> > 7.1. Contiguous TE LSPs > > ## Need to generalize this section for areas, etc. > > After an inter-AS TE LSP has been set up, a more optimal route might > appear in the various traversed ASes. Then in this case, it is > desirable to get the ability to reroute an inter-AS TE LSP in a non- > disruptive fashion (making use of the so called Make Before Break > procedure) to follow this more optimal path. This is a known as a TE > LSP reoptimization procedure. > > ## Is there any requirement to redescribe the procedures of > ## [LOOSE-PATH-REOPT]? Why not simply state that... > ## [LOOSE-PATH-REOPT] describes a mechanism to control discovery > ## or more optimal paths, and to signal new paths using make before > ## break. AA----> Nope, no need at all. > #### BEGIN DELETE > # > # > > [LOOSE-REOPT] proposes a mechanisms allowing: > ## s/[LOOSE-REOPT]/[LOOSE-PATH-REOPT]/ -- change throughout the doc > ## s/proposes/defines/ > > > Vasseur, Ayyangar and Zhang 14 > > > > draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt April 2005 > > > - The head-end LSR to trigger on every LSR whose next hop is a > loose hop the re evaluation of the current path in order to > ## s/re evaluation/re-evaluation/ > detect a potentially more optimal path. This is done via > explicit signaling request: the head-end LSR sets the =F4ERO > Expansion request=F6 bit of the SESSION-ATTRIBUTE object carried > in the RSVP Path message. > > - An LSR whose next hop is a loose-hop to signal to the head- > end LSR that a better path exists. This is performed by sending > an RSVP Path Error Notify message (ERROR-CODE =3D 25), sub-code 6 > (Better path exists). > > This indication may be sent either: > > - In response to a query sent by the head-end LSR, > - Spontaneously by any LSR having detected a more > optimal path > > Such a mechanism allows for the reoptimization of a TE LSP if and only > if a better path is some downstream area/AS is detected. > > The reoptimization event can either be timer or event-driven based (a > link UP event for instance). > > Note that the reoptimization MUST always be performed in a non- > disruptive fashion. > > Once the head-end LSR is informed of the existence of a more optimal > path either in its head-end area/AS or in another AS, the inter-AS TE > Path computation is triggered using the same set of mechanisms as when > the TE LSP is first set up. Then the inter-AS TE LSP is set up > following the more optimal path, making use of the make before break > procedure. In case of a contiguous LSP, the reoptimization process is > strictly controlled by the head-end LSR which triggers the make-before- > break procedure, regardless of the location where the more optimal path > is. > > Note that in the case of loose hop reoptimization, the TE LSP may > follow a preferable path within one or more domain(s) whereas in the > case of PCE-based path computation techniques, the reoptimization > process may lead to following a completely different inter-domain path > (including a different set of ABRs/ASBRs) since end-to-end shortest > path is computed. > > # > # > #### END DELETE > > 7.2. Stitched or nested (non-contiguous) TE LSPs > > ## This section contains quite a lot of repetition. > > In the case of a stitched or nested inter-domain TE LSP, the re- > optimization process is treated as a local matter to any Area/AS. The > main reason is that the inter-domain TE LSP is a different LSP (and > therefore different RSVP session) from the intra-domain LSP segment or > > ## Your causality is broken. Being a different LSP does not imply > ## being a different session. However, the key point *is* that this > ## is a different session (with different ingress and egress). AA-----> We'll simply highlight that this is a different session. <snipped> > 8. Security Considerations > > ## This section seems very lightweight given the importance of > ## inter-AS trust boundaries. AA----> I agree. We should add details about policy control, trust relationship; etc . They may be an overlap with the signaling ID, but that's okay. Also, need to highlight how a boundary node (or any other node) would handle the case where it is being overwhelmed with path computation requests for an inter-domain LSP from some other domain. thanks, -arthi Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 28 Jul 2005 17:40:16 +0000 Date: Thu, 28 Jul 2005 10:39:23 -0700 (PDT) From: Arthi Ayyangar <arthi@juniper.net> To: LE ROUX Jean-Louis RD-CORE-LAN <jeanlouis.leroux@francetelecom.com> cc: JP Vasseur <jvasseur@cisco.com>, Raymond_Zhang@infonet.com, ccamp@ops.ietf.org Subject: Re: ARE: I-D ACTION:draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt Message-ID: <20050728103821.J12097@zircon.juniper.net> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII > > "Note that stiching or nesting rely on local optimization...hence the > > reoptimization is not global and... Procedures describes in loose hop > > MUST be used... " It would be good to mention that global PCE-based > > reoptimization (with potential change of border LSRs) also apply to > > stiching/nesting. agree > AA-------> I couldn't ^ more. -arthi Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 28 Jul 2005 17:34:49 +0000 Date: Thu, 28 Jul 2005 10:33:36 -0700 (PDT) From: Arthi Ayyangar <arthi@juniper.net> To: LE ROUX Jean-Louis RD-CORE-LAN <jeanlouis.leroux@francetelecom.com> cc: JP Vasseur <jvasseur@cisco.com>, Raymond_Zhang@infonet.com, ccamp@ops.ietf.org Subject: Re: ARE: I-D ACTION:draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt Message-ID: <20050728100528.N12097@zircon.juniper.net> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Hi JL, (You seem to be on a roll...14 IDs down, 4 to go... :-)) > A few comments related to your inter-domain pd draft AA---> Thanks for the comments. > -Section: 1) common assumptions, Why not adding and addressing the case > where a path is made of a set of abstract nodes? For instance, in the > inter-AS case, an ERO may list a set of ASes. This may require specific > procedures when computing the path and selecting the downstream ASBR... AA-------> True. This needs to be addressed. The procedures would require interaction with BGP to find the ASBR exit point for the nexthop AS. > -Section 4: 1) Even if this is not really in the scope I think that the > auto-discovery mechanism would require some more words. This is > partially addressed latter in the draft(4.2.1), but you should mention > here that such auto-discovery may rely on IGP information and/or BGP > information and also on local policies (particularly in inter-AS/SP > case). In the inter-AS/SP case there may be several candidate downstream > ASBRs, pertaining to distinct ASes/SPs, and the downstream AS selection > may require appying policies related to specific "TE peering contract" > negociated with neigbors Ases/SPs... AA---------> Actually the section on Assumptions where we introduce Auto-discovery does say that it would rely on IGP and/or BGP information. Would you just like us to expand that to include use of local policies in conjuction with the above information when there are sevaral downstream candidates ? > -Section 4 2) b) ii) What do you exactly mean by "the boundary LSR is > not a FA-LSP/LSP segment candidate"? Does it mean in this case, that a > contigous LSP is created?, this would not be consistent with b) AA----> This is what happens when drafts are merged, split, re-named; etc. :-). One of the versions of the draft actually explained what "candidate" meant...Well... This means that if the LSR has no local policy for nesting or stitching the inter-domain LSP, then it would simply set up a contiguous LSP. Also, I think we need to fix the assumption that FA LSPs or LSP segments can exist only between domain boundary LSRs. It might be the case, but it is not necessary. > -Top of Section 4.1.1: You should remove the first sentence after the > three path examples, "At least the set", as this sounds an ambiguous > repetition of what is just stated above AA-----> Okay. > -Section 6.2: "If the Inter-AS TE related Information is flooded in the > IGP, each ASBR is capable of computing the path..." > > Actually there may be some topologies where an ASBR cannot compute a > link protection path. > > See the following example > > AS1--> <---AS2---- > ASBR1----ASBR2 > | | > | | > ASBR3----ASBR4 > > ASBR1 is aware of links ASBR1->ASBR2, ASBR1->ASBR3 and ASBR3->ASBR4 but > not the link ASBR2->ASBR4 as there is no flooding between AS1 and AS2. > Hence ASBR1 cannot compute a path to protect link ASBR1-ASBR2. > > Actually this case is similar to node protection issue, the path should > be computed according to the pd computation methods defined in the doc. AA--------> I agree completely. > > Also it would be useful to mention that pd inter-domain backup path may requires XRO incusion in RSVP messages... AA-----> That's fine. This is useful for any computation where loose hop expansion occurs in different domains. > -Section 7.2 > > "Note that stiching or nesting rely on local optimization...hence the > reoptimization is not global and... Procedures describes in loose hop > MUST be used... " It would be good to mention that global PCE-based > reoptimization (with potential change of border LSRs) also apply to > stiching/nesting. AA-------> I couldn't more. thanks, -arthi In both cases, contiguous or stiching/nesting, PCE or > any offline computation is required to achieve end-to-end > reoptimization. > > > > > > -----Message d'origine----- > > De : owner-ccamp@ops.ietf.org > > [mailto:owner-ccamp@ops.ietf.org] De la part de > > Internet-Drafts@ietf.org > > Envoy=E9 : vendredi 15 avril 2005 21:34 > > =C0 : i-d-announce@ietf.org > > Cc : ccamp@ops.ietf.org > > Objet : I-D ACTION:draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt > > > > A New Internet-Draft is available from the on-line > > Internet-Drafts directories. > > This draft is a work item of the Common Control and > > Measurement Plane Working Group of the IETF. > > > > =09Title=09=09: A Per-domain path computation method for > > =09=09=09 computing Inter-domain Traffic > > Engineering (TE) > > =09=09=09 Label Switched Path (LSP) > > =09Author(s)=09: J. Vasseur, et al. > > =09Filename=09: > > draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt > > =09Pages=09=09: 0 > > =09Date=09=09: 2005-4-15 > > > > This document specifies a per-domain path computation method > > for computing inter-domain Traffic Engineering (TE) > > Multiprotocol Label Switching (MPLS) and Generalized MPLS > > (GMPLS) Label Switched (LSP) paths. In this document a domain > > is referred to as a collection of network elements within a > > common sphere of address management or path computational > > responsibility such as IGP areas and Autonomous Systems. > > > > A URL for this Internet-Draft is: > > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-inter-dom > > ain-pd-path-comp-00.txt > > > > To remove yourself from the I-D Announcement list, send a > > message to i-d-announce-request@ietf.org with the word > > unsubscribe in the body of the message. > > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > > to change your subscription settings. > > > > > > Internet-Drafts are also available by anonymous FTP. Login > > with the username "anonymous" and a password of your e-mail > > address. After logging in, type "cd internet-drafts" and then > > =09"get draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt". > > > > A list of Internet-Drafts directories can be found in > > http://www.ietf.org/shadow.html or > > ftp://ftp.ietf.org/ietf/1shadow-sites.txt > > > > > > Internet-Drafts can also be obtained by e-mail. > > > > Send a message to: > > =09mailserv@ietf.org. > > In the body type: > > =09"FILE > > /internet-drafts/draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt". > > > > NOTE:=09The mail server at ietf.org can return the document in > > =09MIME-encoded form by using the "mpack" utility. To use this > > =09feature, insert the command "ENCODING mime" before the "FILE" > > =09command. To decode the response(s), you will need "munpack" or > > =09a MIME-compliant mail reader. Different MIME-compliant > > mail readers > > =09exhibit different behavior, especially when dealing with > > =09"multipart" MIME messages (i.e. documents which have been split > > =09up into multiple messages), so check your local documentation on > > =09how to manipulate these messages. > > > > > > Below is the data which will enable a MIME compliant mail reader > > implementation to automatically retrieve the ASCII version of the > > Internet-Draft. > > > Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 28 Jul 2005 17:04:15 +0000 Date: Thu, 28 Jul 2005 10:03:18 -0700 (PDT) From: Arthi Ayyangar <arthi@juniper.net> To: Igor Bryskin <i_bryskin@yahoo.com> cc: Dimitri.Papadimitriou@alcatel.be, Igor Bryskin <ibryskin@movaz.com>, ccamp@ops.ietf.org Subject: Re: LSP stitching questions for discussion Message-ID: <20050726120537.L91159@zircon.juniper.net> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Thanks Dimitri, Igor and Adrian for posting your answers. There seems to be agreement that the procedures are applicable to both PSC and non-PSC LSPs. Good. As expected, the main disagreement is whether stitching should or should not be allowed when region boundaries (SC based) are traversed. I tend to agree with Dimitri and Igor on this. Just because it is possible, doesn't mean it should be allowed, especially given that one can use hierarchical LSPs for this purpose. However, to Adrian's point about "declaring a specific hardware deployment illegal in the control plane", it would be useful to understand this scenario wrt what is done in data plane and in control plane and is there any benefit of this behavior. I will try to capture this discussion in the presentation at the WG meeting. thanks, -arthi > [dp] ok - last try for an end-to-end PSC-1 LSP of 2MB you could make use > of a PSC-1 segment of 4MB > > Ah, Ok. I see your point and agree. > > Igor > > > You are right: whether PSC1 and PSC2 are separate network layers or not > is a topic of a separate (painful :=)) discussion. However, I just > wanted to point out here that a PSC2 LSP is a provisioned as such for a > reason: it is meant to *nest* PSC1 LSPs rather than to stitch them. From > the VNT management point of view PSC2 and PSC1 are separate topological > layers. that is PSC2 produces topology for PSC1, just like VC4 produces > topology for VC12. The only difference is that in the latter case this > relationship is forced by the hardware, and in the former case is not > (just logical). Take for example C and C++ programming. Certain things > you simply forced not to do in C++, whereas in C you could but > shouldn't. > > See you in Paris. > Igor > > > > > "Igor Bryskin" <ibryskin@movaz.com> > Sent by: owner-ccamp@ops.ietf.org > 07/20/2005 13:25 AST > > To: "Arthi Ayyangar" <arthi@juniper.net>, Dimitri > PAPADIMITRIOU/BE/ALCATEL@ALCATEL cc: <ccamp@ops.ietf.org> bcc: > Subject: Re: LSP stitching questions for discussion > > > Dimitri, > > It seems to me that I mostly agree(?!) with what you wrote in this mail. > I'd still want to confirm your views on the following questions: > > > a) Is it possible (should we allow) to stitch PSC1 e2e LSP by a LSC LSP > segment ? > > b) Is it possible (should we allow) to stitch VC12 e2e LSP by a VC4 LSP > segment ? c) Is it possible (should we allow) to stitch PSC1 e2e LSP by > a PSC2 LSP segment ? > > My answers are NO to all three questions. You can sticth PSC1 with PSC1, > VC12 with VC12, etc. In all other cases you need, first, to setup a > hierarchical LSP (H-LSP) (whether it is FA-LSP (single domain) or not > (multi-domain)) in order to create a data link of the e2e LSP switching > capability and only after that you can also use a stitching segment on > top of this H-LSP). > > See also in-line. > > Igor > > > ----- Original Message ----- > From: Dimitri.Papadimitriou@alcatel.be > To: Arthi Ayyangar > Cc: Igor Bryskin ; ccamp@ops.ietf.org > Sent: Wednesday, July 20, 2005 4:46 AM > Subject: Re: LSP stitching questions for discussion > > arthi, > > > > first a recap: this item on stitching was initiated in order to allow > per "domain" control of end-to-end (contiguous) LSP, as part of a > separate document, and not to extend its scope (unduly, in this case, > see below) but mainly in order to have the set of all related protocol > mechanisms within a specific document - > > > > from this initial working assumption, it is clear that the segment must > have the properties of the incoming end-to-end LSP which are determined > among other by its switching type e.g. PSC-3 or TDM, the first > requirement of LSP stitching is contiguity of the end-to-end LSP at the > data plane level (after stitching operation) - and this by definition of > the initial working assumption - > > IB>> Agree > > note: the only latitude which is left (for a given switching type e.g. > PSC-2) is due to the properties of packet (PSC) networks that allows you > stitch a segment having a larger Max Reservable Bandwidth and Unreserved > Bandwidth (we would have equivalent latitude for other attributes but > this is a local policy decision - which at the end is the sole purpose > of this approach) > > IB>> Disagree. Conceptually PCS1 and PSC2 are different layers and the > latter cannot stitch the former, just like in case VC12 and VC4. You > should nest PSC1 within PSC2 instead. > > > > Here are some of the questions related to LSP stitching to start a > discussion on the usage of LSP stitching. > > >1) Are the control plane signaling procedures for LSP stitching as > >described in this ID, (explicit request for LSP stitching on > >LSP segment, different label allocation rules; etc) REQUIRED to stitch > >packet LSPs in data plane > > > -> yes (beside manual config operation) > > IB>> Agree > > > >2) Are the control plane signaling procedures for LSP stitching as > >described in this ID, (explicit request for LSP stitching on > >LSP segment, different label allocation rules; etc) REQUIRED to stitch > >non-packet LSPs in data plane ? > > > -> yes (beside manual config operation) > > IB>> Agree > > > >3) If an e2e LSP crosses a region boundary (based on different TE > >link switching capabilities), then can I use LSP stitching > >a) control plane procedures for packet LSPs > >b) data plane procedures for packet LSPs > >c) control plane procedures for non-packet LSPs > >d) data plane procedures for non-packet LSPs > > > -> no, when crossing a region boundary, you simply use a document named > "draft-mpls-lsp-hierarchy-08.txt" which is in the RFC Editor queue, but > this does not prevent from creating an LSP segment on top of an FA-LSP > (see below) > > IB>> Agree with the caviat that it should be H-LSP instead of FA-LSP to > cover single and multi-domain scenarios > > > >4) If the switching type of an e2e LSP is different from that of an LSP > >segment, can I use LSP stitching > >a) control plane procedures for packet LSPs > >b) data plane procedures for packet LSPs > >c) control plane procedures for non-packet LSPs > >d) data plane procedures for non-packet LSPs > > > -> no, in this case progression of the end-to-end LSP establishment, > implies that the LSP segment must fulfil a set of constraints carried as > part of the incoming end-to-end LSP request, in this case one has to > create an LSP segment (with a switching type equal to the incoming > end-to-end LSP switching type) over that FA (i.e. referred here above to > the LSP segment with a different Switching Type) > > > IB>> Agree with the caviat that it should be H-LSP instead of FA-LSP to > cover single and multi-domain scenarios > > Cheers, > > Igor > > >Note that although 3) and 4) appear to be the same, the reason I have > >them as different questions is that in one case the TE link switching > >capability is examined and in the other case the Generalized Label Request > >is examined. > > > > -> what it appears to me is that this discussion mixes the issue of the > operation describing the stitching operations at the domain boundaries > from the LSP segment creation itself i.e. an LSP segment can indeed be > an FA link inherited from the creation of an (intra-domain) FA-LSP - > > > __________________________________________________ > Do You Yahoo!? > Tired of spam? Yahoo! Mail has the best spam protection around > http://mail.yahoo.com Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 28 Jul 2005 16:57:13 +0000 Date: Thu, 28 Jul 2005 09:55:48 -0700 (PDT) From: Arthi Ayyangar <arthi@juniper.net> To: LE ROUX Jean-Louis RD-CORE-LAN <jeanlouis.leroux@francetelecom.com> cc: jpv@cisco.com, ccamp@ops.ietf.org Subject: RE: I-D ACTION:draft-ietf-ccamp-inter-domain-rsvp-te-01.txt Message-ID: <20050728093810.R12097@zircon.juniper.net> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Hi JL, Thanks for the comments. > > -In section 5.2, I'm not sure that path p3 is a valid one: > "RO-X1-ASBR1(loose)-ASBR4(strict)-ASBR7(strict)-ASBR9(loose)-R6. What > does ERO sub-objects for ASBR4 and ASBR5 represent? Actually this cannot > be an interface or router id, because the ERO would be invalid (cannot > be strict as ASBR4 and ASBR7 ar not directly connected). The only way to > proceed in this way is to include ERO-sub-objects corresponding to > FA-LSP or LSP-Segment TE-links (numbered or unnumbered tunnel interface > id). One could assume that RO would obtain such ERO including FA-LSP or > LSP-segment as a result of a PCE based compuation... On receip of a Path > message the ERO subobject would allow to directly determine the > underlying FA-LSP or LSP-Segment... AA----------> The intention was actually to show a case where ERO contains FA-LSP or LSP segment TE link address (as you have described above) in which case, contiguous LSP would not be setup. Looks like the part stating this got edited out somewhere along the line. :-( Also, I agree with you that in that case the ERO would either contain numbered or unnumbered TE link address corresponding to FA-LSP or LSP segment. So we will fix the ERO in the above case to explain this scenario. Thanks for pointing this out. > By the way it would also be good to ilusrate the case of abstract nodes, > e.g. a path made of AS number ERO subobjects... AA-----> You mean in the ERO processing discussion, right ? Will do. > -It would be good to add some text related to failure detection, at the > end of 6.1.3. In case of stitching or nesting, the PLR could rely on > BFD-MPLS so as to rapidliy detect FA-LSP/LSP-Segment tail-end failure. > This would help ensuring fast recovery... AA--------> Makes sense. > -In section 6.2 the FRR considerations described in 6.1.3 for stiching > and nesting, actually also apply to GMPLS Segment Recovery. AA-------> Okay...but mechanisms are different. > -Section 7 would benefif from some words on global reoptimization (ie > with change of border routers) Even if this is more a path computation > issue this may have some impact on signaling... AA------> Since, as you said it is a path computation issue and it doesn't directly cover any signaling extension, it was deliberately left out of this ID. It is mentioned in the pd-path-comp ID. > -And two minor edits Section 5.1.1: issue with B3 path, in the figure > ASBR6 and ASBR7 are not connnected. Section 8: [INTER-AS-TE-REQ] > instead of [INTER-AREA-TE-REQS] AA----> Okay. Will fix. thanks, -arthi > Hope this helps, > > Best Regards > > JL > > > > -----Message d'origine----- > > De : owner-ccamp@ops.ietf.org > > [mailto:owner-ccamp@ops.ietf.org] De la part de > > Internet-Drafts@ietf.org > > Envoy=E9 : mardi 19 juillet 2005 21:50 > > =C0 : i-d-announce@ietf.org > > Cc : ccamp@ops.ietf.org > > Objet : I-D ACTION:draft-ietf-ccamp-inter-domain-rsvp-te-01.txt > > > > A New Internet-Draft is available from the on-line > > Internet-Drafts directories. > > This draft is a work item of the Common Control and > > Measurement Plane Working Group of the IETF. > > > > =09Title=09=09: Inter domain GMPLS Traffic > > Engineering - RSVP-TE extensions > > =09Author(s)=09: A. Ayyangar, J. Vasseur > > =09Filename=09: draft-ietf-ccamp-inter-domain-rsvp-te-01.txt > > =09Pages=09=09: 17 > > =09Date=09=09: 2005-7-19 > > > > This document describes extensions to Generalized > > Multi-Protocol Label Switching (GMPLS) Resource ReserVation > > Protocol - Traffic Engineering > > (RSVP-TE) signaling required to support mechanisms for the > > establishment and maintenance of GMPLS Traffic Engineering > > (TE) Label Switched Paths (LSPs), both packet and non-packet, > > that traverse multiple domains. For the purpose of this > > document, a domain is considered to be any collection of > > network elements within a common realm of address space or > > path computation responsibility. Examples of such domains > > include Autonomous Systems, IGP areas and GMPLS overlay networks. > > > > A URL for this Internet-Draft is: > > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-inter-dom > > ain-rsvp-te-01.txt > > > > To remove yourself from the I-D Announcement list, send a > > message to i-d-announce-request@ietf.org with the word > > unsubscribe in the body of the message. > > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > > to change your subscription settings. > > > > > > Internet-Drafts are also available by anonymous FTP. Login > > with the username "anonymous" and a password of your e-mail > > address. After logging in, type "cd internet-drafts" and then > > =09"get draft-ietf-ccamp-inter-domain-rsvp-te-01.txt". > > > > A list of Internet-Drafts directories can be found in > > http://www.ietf.org/shadow.html or > > ftp://ftp.ietf.org/ietf/1shadow-sites.txt > > > > > > Internet-Drafts can also be obtained by e-mail. > > > > Send a message to: > > =09mailserv@ietf.org. > > In the body type: > > =09"FILE > > /internet-drafts/draft-ietf-ccamp-inter-domain-rsvp-te-01.txt". > > > > NOTE:=09The mail server at ietf.org can return the document in > > =09MIME-encoded form by using the "mpack" utility. To use this > > =09feature, insert the command "ENCODING mime" before the "FILE" > > =09command. To decode the response(s), you will need "munpack" or > > =09a MIME-compliant mail reader. Different MIME-compliant > > mail readers > > =09exhibit different behavior, especially when dealing with > > =09"multipart" MIME messages (i.e. documents which have been split > > =09up into multiple messages), so check your local documentation on > > =09how to manipulate these messages. > > > > > > Below is the data which will enable a MIME compliant mail reader > > implementation to automatically retrieve the ASCII version of the > > Internet-Draft. > > > Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 28 Jul 2005 16:18:10 +0000 Date: Thu, 28 Jul 2005 09:16:56 -0700 (PDT) From: Arthi Ayyangar <arthi@juniper.net> To: LE ROUX Jean-Louis RD-CORE-LAN <jeanlouis.leroux@francetelecom.com> cc: ccamp@ops.ietf.org, jpv@cisco.com Subject: RE: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Message-ID: <20050728085618.M12097@zircon.juniper.net> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=X-UNKNOWN Content-Transfer-Encoding: QUOTED-PRINTABLE Hi JL, Thanks for the comments. > -In section 4.1.2 you partially describe bidirectional LSP stitching > procedure. You mention that an Upstream Label MUST NOT be allocated by > the end-to-end LSP on the LSP segment, which is OK. But then how the > LSP-Segment Egress will now that the end-to-end LSP is bidirectional? AA--------> Excellent point. > What about defining a flag in the Attributes Flags TLV of the > LSP_ATTRIBUTE object so as to indicate that the LSP is bidirectional? AA------> That would be a change to the processing that GMPLS nodes use to detect bidirectionality, isn't it ? Normally nodes look for the Upstream Label object to detect bidirectionality. So let us say that an LSP is bidirectional if a) an Upstream Label is present or b) no Upstream Label, but bit set in LSP_ATTRIBUTE or c) both However, reliance on an e2e attributes bit set by head end, means existing head ends will not be setting this bit, so that will be an issue (wrt compatibility). Could be nice if this signaling was just between the node doing the stitching and the end point of the LSP segment, since this is the hop that the bidirectionality information is lost. Let me think about this. > Also the selection of the LSP segment in case of bidirectional LSP > should be detailed (e.g. If the end-to-end LSP is bidirecitonal then the > LSP-segment MUST be bidirectional. Also shall we allow that two > unidrectional end-to-end LSP use the same bidirectional LSP segment (one > in each direction)? AA----> Yes, this should be okay. IMO. > -At the end of section 4.2.5 you mention that LSP-Segment failure or > maintenance SHOULD be treated as a failure event for the end-to-end LSP. > I agree for LSP-Segment failure but not for LSP-Segment maintenance. > LSP-Segment maintenance should be treated as TE-link maintenance for the > end-to-end LSP, and procedures defined in GMPLS graceful TE-link > shutdown draft may be useful (Specific RSVP error code and TE-link > attribute)... AA---> Yes, I agree; we will mention this. Actually the graceful teardown sentence occurs few paragraphs before, but not in the right context. So we will clarify the above. > Hope this helps, AA---> Sure does. Thanks! -arthi > > > > -----Message d'origine----- > > De : owner-ccamp@ops.ietf.org > > [mailto:owner-ccamp@ops.ietf.org] De la part de > > Internet-Drafts@ietf.org > > Envoy=E9 : vendredi 15 juillet 2005 21:50 > > =C0 : i-d-announce@ietf.org > > Cc : ccamp@ops.ietf.org > > Objet : I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > > > > A New Internet-Draft is available from the on-line > > Internet-Drafts directories. > > This draft is a work item of the Common Control and > > Measurement Plane Working Group of the IETF. > > > > =09Title=09=09: Label Switched Path Stitching with > > Generalized MPLS Traffic Engineering > > =09Author(s)=09: A. Ayyangar, J. Vasseur > > =09Filename=09: draft-ietf-ccamp-lsp-stitching-01.txt > > =09Pages=09=09: 19 > > =09Date=09=09: 2005-7-15 > > > > In certain scenarios, there may be a need to combine together two > > different Generalized Multi-Protocol Label Switching (GMPLS) Label > > Switched Paths (LSPs) such that in the data plane, a single end-to- > > end (e2e) LSP is achieved and all traffic from one LSP is switched > > onto the other LSP. We will refer to this as "LSP > > stitching". This > > document covers cases where: a) the node performing the stitching > > does not require configuration of every LSP pair to be stitched > > together b) the node performing the stitching is not the egress of > > any of the LSPs c) LSP stitching not only results in an end-to-end > > LSP in the data plane, but there is also a corresponding end-to-end > > LSP (RSVP session) in the control plane. It might be possible to > > configure a GMPLS node to switch the traffic from an LSP > > for which it > > is the egress, to another LSP for which it is the ingress, without > > requiring any signaling or routing extensions whatsoever, > > completely > > transparent to other nodes. This will also result in LSP stitching > > in the data plane. However, this document does not cover this > > scenario of LSP stitching. > > > > A URL for this Internet-Draft is: > > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitc > hing-01.txt > > > > To remove yourself from the I-D Announcement list, send a > > message to i-d-announce-request@ietf.org with the word > > unsubscribe in the body of the message. > > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > > to change your subscription settings. > > > > > > Internet-Drafts are also available by anonymous FTP. Login > > with the username "anonymous" and a password of your e-mail > > address. After logging in, type "cd internet-drafts" and then > > =09"get draft-ietf-ccamp-lsp-stitching-01.txt". > > > > A list of Internet-Drafts directories can be found in > > http://www.ietf.org/shadow.html or > > ftp://ftp.ietf.org/ietf/1shadow-sites.txt > > > > > > Internet-Drafts can also be obtained by e-mail. > > > > Send a message to: > > =09mailserv@ietf.org. > > In the body type: > > =09"FILE /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt". > > > > NOTE:=09The mail server at ietf.org can return the document in > > =09MIME-encoded form by using the "mpack" utility. To use this > > =09feature, insert the command "ENCODING mime" before the "FILE" > > =09command. To decode the response(s), you will need "munpack" or > > =09a MIME-compliant mail reader. Different MIME-compliant > > mail readers > > =09exhibit different behavior, especially when dealing with > > =09"multipart" MIME messages (i.e. documents which have been split > > =09up into multiple messages), so check your local documentation on > > =09how to manipulate these messages. > > > > > > Below is the data which will enable a MIME compliant mail reader > > implementation to automatically retrieve the ASCII version of the > > Internet-Draft. > > > Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 28 Jul 2005 15:50:27 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: I-D ACTION:draft-ietf-ccamp-gmpls-addressing-01.txt Date: Thu, 28 Jul 2005 17:48:10 +0200 Message-ID: <D109C8C97C15294495117745780657AE02EDBF93@ftrdmel1.rd.francetelecom.fr> Thread-Topic: I-D ACTION:draft-ietf-ccamp-gmpls-addressing-01.txt Thread-Index: AcV2m77Ol7fv6cWXQFmcF4CWX3tWXQc7fv6g From: "LE ROUX Jean-Louis RD-CORE-LAN" <jeanlouis.leroux@francetelecom.com> To: <shiomoto.kohei@lab.ntt.co.jp>, <rpapneja@isocore.com>, "Richard Rabbat" <richard.rabbat@us.fujitsu.com> Cc: <ccamp@ops.ietf.org> Hi Kohei, Rajiv, Richard, I have on comment related to your GMPLS addressing ID: In section 5.2 you recommend, that in the session object, the IP tunnel = end-point be set to the destination TE router id and that the extended = tunnel id be set to the source TE router id, which makes sense. But as = per 3209, a session object can either encode IPv4 or IPv6 addresses but = not both types, and there may be an issue if source TE router id is IPv6 = and destination TE router id is IPv4 (or vice versa). This may happen = for instance with inter-AS/SP LSPs. IMHO we should address this point. Best Regards, JL > -----Message d'origine----- > De : owner-ccamp@ops.ietf.org=20 > [mailto:owner-ccamp@ops.ietf.org] De la part de=20 > Internet-Drafts@ietf.org > Envoy=E9 : mardi 21 juin 2005 21:50 > =C0 : i-d-announce@ietf.org > Cc : ccamp@ops.ietf.org > Objet : I-D ACTION:draft-ietf-ccamp-gmpls-addressing-01.txt=20 >=20 > A New Internet-Draft is available from the on-line=20 > Internet-Drafts directories. > This draft is a work item of the Common Control and=20 > Measurement Plane Working Group of the IETF. >=20 > Title : Use of Addresses in Generalized Multi-Protocol > Label Switching (GMPLS) Networks > Author(s) : K. Shiomoto, et al. > Filename : draft-ietf-ccamp-gmpls-addressing-01.txt > Pages : 23 > Date : 2005-6-21 > =09 > This document explains and clarifies the use of addresses in=20 > Generalized Multi-Protocol Label Switching (GMPLS)=20 > networks. The aim=20 > of this document is to facilitate and ensure better=20 > interworking of=20 > GMPLS-capable Label Switching Routers (LSRs) based on experience=20 > gained in deployments and interoperability testing and proper=20 > interpretation of published RFCs.=20 > =20 > The document recommends a proper approach for the=20 > interpretation and=20 > choice of address and identifier fields within GMPLS protocols and=20 > references specific control plane usage models. It also examines=20 > some common GMPLS Resource Reservation Protocol-Traffic=20 > Engineering=20 > (RSVP-TE) signaling message processing issues and recommends=20 > solutions. It finally discusses how to handle IPv6 sources and=20 > destinations in the MPLS and GMPLS TE (Traffic Engineering) MIB=20 > (Management Information Base) modules. >=20 > A URL for this Internet-Draft is: > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-gmpls-add ressing-01.txt >=20 > To remove yourself from the I-D Announcement list, send a=20 > message to i-d-announce-request@ietf.org with the word=20 > unsubscribe in the body of the message. =20 > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > to change your subscription settings. >=20 >=20 > Internet-Drafts are also available by anonymous FTP. Login=20 > with the username "anonymous" and a password of your e-mail=20 > address. After logging in, type "cd internet-drafts" and then > "get draft-ietf-ccamp-gmpls-addressing-01.txt". >=20 > A list of Internet-Drafts directories can be found in=20 > http://www.ietf.org/shadow.html or=20 > ftp://ftp.ietf.org/ietf/1shadow-sites.txt >=20 >=20 > Internet-Drafts can also be obtained by e-mail. >=20 > Send a message to: > mailserv@ietf.org. > In the body type: > "FILE=20 > /internet-drafts/draft-ietf-ccamp-gmpls-addressing-01.txt". > =09 > NOTE: The mail server at ietf.org can return the document in > MIME-encoded form by using the "mpack" utility. To use this > feature, insert the command "ENCODING mime" before the "FILE" > command. To decode the response(s), you will need "munpack" or > a MIME-compliant mail reader. Different MIME-compliant=20 > mail readers > exhibit different behavior, especially when dealing with > "multipart" MIME messages (i.e. documents which have been split > up into multiple messages), so check your local documentation on > how to manipulate these messages. > =09 > =09 > Below is the data which will enable a MIME compliant mail reader > implementation to automatically retrieve the ASCII version of the > Internet-Draft. >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 28 Jul 2005 13:56:02 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: ARE: I-D ACTION:draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt Date: Thu, 28 Jul 2005 15:53:03 +0200 Message-ID: <D109C8C97C15294495117745780657AE02EDBE5D@ftrdmel1.rd.francetelecom.fr> Thread-Topic: ARE: I-D ACTION:draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt Thread-Index: AcVB80eYS2kUvd3iTW2jR3Dz9+WeLRRgvNLQ From: "LE ROUX Jean-Louis RD-CORE-LAN" <jeanlouis.leroux@francetelecom.com> To: "JP Vasseur" <jvasseur@cisco.com>, "Arthi Ayyangar" <arthi@juniper.net>, <Raymond_Zhang@infonet.com> Cc: <ccamp@ops.ietf.org> Hi JP, Arthi Raymond A few comments related to your inter-domain pd draft -Section: 1) common assumptions, Why not adding and addressing the case where a path is made of a set of = abstract nodes? For instance, in the inter-AS case, an ERO may list a set of ASes. This = may require specific procedures when computing the path and selecting = the downstream ASBR... -Section 4: 1)=20 Even if this is not really in the scope I think that the auto-discovery = mechanism would require some more words. This is partially addressed = latter in the draft(4.2.1), but you should mention here that such = auto-discovery may rely on IGP information and/or BGP information and = also on local policies (particularly in inter-AS/SP case). In the inter-AS/SP case there may be several candidate downstream ASBRs, = pertaining to distinct ASes/SPs, and the downstream AS selection may = require appying policies related to specific "TE peering contract" = negociated with neigbors Ases/SPs... -Section 4 2) b) ii) What do you exactly mean by "the boundary LSR is = not a FA-LSP/LSP segment candidate"? Does it mean in this case, that a contigous LSP is created?, this would = not be consistent with b)=20 -Top of Section 4.1.1: You should remove the first sentence after the = three path examples, "At least the set", as this sounds an ambiguous repetition of what is = just stated above -Section 6.2: "If the Inter-AS TE related Information is flooded in the IGP, each ASBR = is capable of computing the path..." Actually there may be some topologies where an ASBR cannot compute a = link protection path. See the following example AS1--> <---AS2---- ASBR1----ASBR2 | | | | ASBR3----ASBR4 ASBR1 is aware of links ASBR1->ASBR2, ASBR1->ASBR3 and ASBR3->ASBR4 but = not the link ASBR2->ASBR4 as there is no flooding between AS1 and AS2. Hence ASBR1 cannot compute a path to protect link ASBR1-ASBR2. Actually this case is similar to node protection issue, the path should = be computed according to the pd computation methods defined in the doc. Also it would be useful to mention that pd inter-domain backup path may = requires XRO incusion in RSVP messages... -Section 7.2 "Note that stiching or nesting rely on local optimization...hence the = reoptimization is not global and... Procedures describes in loose hop MUST be used... " It would be good to mention that global PCE-based reoptimization (with = potential change of border LSRs) also apply to stiching/nesting. In both cases, contiguous or stiching/nesting, PCE or any offline = computation is required to achieve end-to-end reoptimization. Hope it helps, Best Regards, JL > -----Message d'origine----- > De : owner-ccamp@ops.ietf.org=20 > [mailto:owner-ccamp@ops.ietf.org] De la part de=20 > Internet-Drafts@ietf.org > Envoy=E9 : vendredi 15 avril 2005 21:34 > =C0 : i-d-announce@ietf.org > Cc : ccamp@ops.ietf.org > Objet : I-D ACTION:draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt >=20 > A New Internet-Draft is available from the on-line=20 > Internet-Drafts directories. > This draft is a work item of the Common Control and=20 > Measurement Plane Working Group of the IETF. >=20 > Title : A Per-domain path computation method for=20 > computing Inter-domain Traffic=20 > Engineering (TE)=20 > Label Switched Path (LSP) > Author(s) : J. Vasseur, et al. > Filename :=20 > draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt > Pages : 0 > Date : 2005-4-15 > =09 > This document specifies a per-domain path computation method=20 > for computing inter-domain Traffic Engineering (TE)=20 > Multiprotocol Label Switching (MPLS) and Generalized MPLS=20 > (GMPLS) Label Switched (LSP) paths. In this document a domain=20 > is referred to as a collection of network elements within a=20 > common sphere of address management or path computational=20 > responsibility such as IGP areas and Autonomous Systems. >=20 > A URL for this Internet-Draft is: > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-inter-dom > ain-pd-path-comp-00.txt >=20 > To remove yourself from the I-D Announcement list, send a=20 > message to i-d-announce-request@ietf.org with the word=20 > unsubscribe in the body of the message. =20 > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > to change your subscription settings. >=20 >=20 > Internet-Drafts are also available by anonymous FTP. Login=20 > with the username "anonymous" and a password of your e-mail=20 > address. After logging in, type "cd internet-drafts" and then > "get draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt". >=20 > A list of Internet-Drafts directories can be found in=20 > http://www.ietf.org/shadow.html or=20 > ftp://ftp.ietf.org/ietf/1shadow-sites.txt >=20 >=20 > Internet-Drafts can also be obtained by e-mail. >=20 > Send a message to: > mailserv@ietf.org. > In the body type: > "FILE=20 > /internet-drafts/draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt". > =09 > NOTE: The mail server at ietf.org can return the document in > MIME-encoded form by using the "mpack" utility. To use this > feature, insert the command "ENCODING mime" before the "FILE" > command. To decode the response(s), you will need "munpack" or > a MIME-compliant mail reader. Different MIME-compliant=20 > mail readers > exhibit different behavior, especially when dealing with > "multipart" MIME messages (i.e. documents which have been split > up into multiple messages), so check your local documentation on > how to manipulate these messages. > =09 > =09 > Below is the data which will enable a MIME compliant mail reader > implementation to automatically retrieve the ASCII version of the > Internet-Draft. >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 28 Jul 2005 11:40:22 +0000 Message-ID: <42E8C3CC.6060208@pi.se> Date: Thu, 28 Jul 2005 13:38:52 +0200 From: Loa Andersson <loa@pi.se> Organization: Acreo AB User-Agent: Mozilla Thunderbird 1.0.5 (Windows/20050711) MIME-Version: 1.0 To: "Ali Sajassi (sajassi)" <sajassi@cisco.com> CC: ccamp@ops.ietf.org Subject: Re: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Ali Sajassi (sajassi) wrote: > Loa, > Ali, I don't parse the paragraph below. If you said "All the MPLS traffic can be carried..." I would understand it. But, then we have done that for the last 6 years. We were talking about extending the GMPLS control plane to control Ethernet traffic, and into a MRN (gosh I hate that name ;)) architecture. My take here is that we need to cooperate with IEEE to do this, since in my opinion, we need to discuss the forwarding plane. /Loa > I think you might have misinterpreted what I was saying. I was talking > about using MPLS encap first and then have MPLS traffic over its > designated VLAN(s). All the GMPLS traffic can be carried over a single > VLAN or you can have more than one for load balancing purposes. And thus > there shouldn't be any scalability issues. > > >>i don't know if we are ready to start the discussion on >>solution, but we seems to do that anyway ;) >> > > Apparently :-) > > >>there is a way for us to get a tpid from IEEE, we could >>specify the behavior of switches on that tpid >> >>my take that we shouldn't do that without involving ieee, but >>on the other it is normally quite easy to work with them >> > > > If interested one can come and discuss it during next 802.1 group (which > is the architecture group for IEEE bridges) meeting. > > >>but again we need to undestand requirements first >> > > > Completely agree. > > -Ali > > >>/Loa >> >>Ali Sajassi (sajassi) wrote: >> >>>Dimitri, >>> >>>By "existing Ethernet switches", I mean Ethernet switches based on >>>current approved IEEE 802 .1Q/.1D standards and not the >> >>hypothetical >> >>>switches with TBD forwarding characteristics. Anyway, if >> >>the objective >> >>>is to use RSVP-TE with "existing Ethernet switches", you >> >>can do that >> >>>easily by designating a VLAN or (set of VLANs) for RSVP-TE >> >>traffic and >> >>>support both EoMPLS and native bridged traffic on a single trunk. >>>There is some overhead with such encapsulation but the >> >>benefit of it >> >>>is that it works now. If we want to do it through some other means >>>(such as MAC address translation), then the benefit of such >> >>approach >> >>>is not clear to me; however, the disadvantages are very clear >>>:-) Anyway, I'll look forward to more discussions during >> >>and after the >> >>>meeting next week. >>> >>>Cheers, >>>Ali >>> >>> >> >>-------------------------------------------------------------- >>---------- >> >>> *From:* Dimitri.Papadimitriou@alcatel.be >>> [mailto:Dimitri.Papadimitriou@alcatel.be] >>> *Sent:* Tuesday, July 26, 2005 4:21 PM >>> *To:* Ali Sajassi (sajassi) >>> *Cc:* CHO, JAI HYUNG; Heiles Juergen; per@defero.se; >> >>ccamp@ops.ietf.org >> >>> *Subject:* RE: Frameformat in a l2cs gmpls rnvironment >> >>- Issues w/ >> >>> GMPLS Ethernet >>> >>> ali >>> >>> to be clear here you are hanging on using the term "existing >>> Ethernet switches" while the document that has been at >> >>the source of >> >>> the discussion does not state "GMPLS must be deployable >> >>on existing >> >>> Ethernet switches" this may be the case but there are >> >>limitations >> >>> (and there are perfectly well understood, don't worry) >> >>beside this >> >>> you should also tell us what do you exactly mean with the term >>> "existing" - but this is a side issue - >>> >>> this said, this does NOT mean that a GMPLS Ethernet >> >>switch (either >> >>> existing or even "extended") would need to support a >> >>new Ethernet >> >>> 802.3 frame forwarding paradigm, the difference is >> >>subtle but has to >> >>> be underlined >>> >>> note: there is a slot of 20 minutes during the next >> >>CCAMP WG meeting >> >>> where this document and all related issues will be >> >>heavily discussed >> >>> ... so i welcome you to join the festivities ;-) >>> >>> thanks, >>> >>> - dimitri. >>> >>> >>> >>> *"Ali Sajassi \(sajassi\)" <sajassi@cisco.com>* >>> Sent by: owner-ccamp@ops.ietf.org >>> 07/25/2005 16:13 MST >>> >>> To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles Juergen" >>> <juergen.heiles@siemens.com>, <per@defero.se> >>> cc: <ccamp@ops.ietf.org> >>> bcc: >>> Subject: RE: Frameformat in a l2cs gmpls rnvironment - Issues w/ >>> GMPLS Ethernet >>> >>> >>> Jaihyung, >>> >>> It seems to me that your primary objective is to use >> >>the existing >> >>> Ethernet switches with new control plane (GMPLS) to >> >>setup TE paths for >> >>> different applications. And you mentioned three >> >>approaches for doing >> >>> this in your email. However, all three of them have >> >>issues and are not >> >>> viable. Lets go over them ... >>> >>> > >>> > In below mail, I discussed about three proposed approaches. >>> > >>> > >>> > 1. using VLAN tag as it is for L2 label encoding. >>> >>> As you know this approach has scalability issue and you >> >>are limited to >> >>> 4K services in the network and if you try to use the >> >>upcoming IEEE >> >>> 802.1ad standard to make VLANs, interface specific, >> >>then you will need >> >>> new switches and not existing Ethernet switches. Even >> >>if you use these >> >>> newer switches, you will still be limited to 4K >> >>services per interface >> >>> and it would defeat the objective of using existing switches. >>> >>> > >>> > 2. defining new protocol ID (TPID) and borrow VLAN >> >>tag format, >> >>> > ?or extended VLAN tag to implement label swapping. >>> > >>> >>> This approach requires new data plane functionality to >> >>be implemented >> >>> (which defeats the objective of using existing switches). >>> >>> > 3. use lower 3 bytes of MAC address for L2 label encoding. >>> > >>> >>> This approach has several issues: >>> 1) how does MAC addresses get distributed among the >> >>bridges ? How does >> >>> this distribution works during a link or node failure ? >> >>How does it >> >>> recover from a link or node failure ? >>> >>> 2) Will you assign a separate provider MAC for each >> >>customer MAC. If so, >> >>> then how do you address MAC scalability. If not, then >> >>how do you take >> >>> care of 1-to-N mapping between provider MAC address and >> >>customer MAC >> >>> addresses and how do you handle multipoint connections. >>> >>> 3) how do handle customer FCS retention when doing this >> >>MAC translation >> >>> stuff ?? >>> >>> > >>> > 802.1Q bridge forwards Ethernet frames using two >> >>dataplane tables >> >>> > - MAC forwarding table and VLAN forwarding table. >>> > Bridge control protocols, such as GARP, GVRP, GMRP, >>> > manipulate one of the two dataplane entities. >>> >>> This is not correct. Bridges don't have different >> >>databases for MAC >> >>> forwarding and VLAN forwarding. They have filtering >> >>datebase(s) with >> >>> corresponding IDs (FIDs). VLAN IDs get mapped to FIDs >> >>based on operation >> >>> mode (IVL versus SVL). Then once a filtering database >> >>is identified, >> >>> then a MAC address lookup is performed for forwarding. >> >>Bridge control >> >>> protocols use a designated MAC addressed and based on these MAC >>> addressed, the bridge knows how to process the frame. >>> >>> > >>> > Similarly, option 1 and 3 are about which one of two >>> > dataplane entities >>> > GMPLS protocol should control on behalf of bridge >> >>control protocols. >> >>> > The two proposals do not intend to modify bridge behavior >>> > seriously, such as MAC learning, aging, filtering. >>> > Therefore, the approaches 1 and 3 are in the scope of CCAMP. >>> >>> If that's the intention, you may find out soon you get >> >>more than what >> >>> you asked for :-) >>> >>> > >>> > (% note however, I would regard implementing label swapping >>> > function is acceptable change considering the >> >>Internal Sublayer >> >>> > design of 802.1D bridge.) >>> > >>> >>> The only way of doing this is using VID translation >> >>table specified in >> >>> IEEE 802.1ad which requires new data-plane >> >>functionality and thus >> >>> existing bridges cannot be used !! >>> >>> Also if going the route of tag swapping, you may >> >>consider using IEEE >> >>> 802.1ah. As the editor of .1ah, I can go over the >> >>details when I get a >> >>> chance. >>> >>> > >>> > However, option 2 is about what TPID code we will >>> > choose to implement new switching function. >>> > The switch of this implementation should have new hardware >>> > design in addition to normal bride hardware. >>> > I think such proposal as option 2 is out of CCAMP scope. >>> >>> Yes. >>> >>> > >>> > Overall objective is improving scalability, traffic >> >>engineering (TE) >> >>> > characteristics of 802.1 bridge that it can be >> >>reliable, manageable >> >>> > enough to replace some core routers. >>> >>> Easier said than done :-) >>> >>> > The switching technique you mentioned above, such as >>> > Ethernet over MPLS as defined by PWE3 and L2VPN, >>> > are all actually router based technology, however this >>> > work is based on simple bridge architecture. >>> > Cost-effectiveness is the key differentiator. >>> > >>> >>> The devils are in the details and once you worked out >> >>the details, you >> >>> will see the issues. >>> >>> > >>> > In access or enterprise network, capability of >> >>providing end-to-end >> >>> > L2-LSP will enable service providers policing, >> >>measuring, charging >> >>> > application flows using Ethernet based network. This >> >>will eventually >> >>> > improve income structure and introduce new session >> >>based commercial >> >>> > service. I have been discussing this aspect in mail thread of >>> > title 'End-to-end L2-LSP'. Please read the mail >> >>thread and comment >> >>> > on the discussion. >>> > >>> > >>> > >You mentioned to use the MAC address or part of it >> >>as a label. >> >>> > >The TRILL working group is defining shortest path routing >>> > for Ethernet. >>> > >How would this fit together? >>> > >>> > >>> > TRILL employs new encapsulation layer outside of >> >>Ethernet frames. >> >>> > However in my proposal of option-3, I simply use the >> >>Ethernet header >> >>> > to encode GMPLS label. >>> > >>> > The overall format of rbridge is >>> > [Ethernet][r-tag][Ethernet][ data..] >>> > It is not LSP based technology but a connectionless >> >>routing bridge. >> >>> > TTL count is very important in rbridge because >> >>frames are routed >> >>> > hop-by-hop and it may loop. >>> > >>> > Whereas in option-3 above, Ethernet frames that have certain >>> > OUI prefix in MAC address are forwarded via a path that >>> > GMPLS protocol has configured on MAC forwarding table. >>> > >>> >>> How does the redundancy work in here. Are you using >> >>bridge protocol in >> >>> conjunction with GMPLS protocol. If you do, then there >> >>are bunch of >> >>> issues. If you don't, then what is it used and how does >> >>it interact with >> >>> bridging protocols ?? >>> >>> -Ali >>> >>> >>> > >>> > Thanks >>> > >>> > Jaihyung >>> > >>> > >>> > >>> > -----?? ???----- >>> > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com> >>> > ?? ??: 2005-07-25 ?? 6:44:34 >>> > ?? ??: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles >>> > Juergen" <juergen.heiles@siemens.com>, "per@defero.se" >>> > <per@defero.se> >>> > ??: "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> >>> > ??: RE: Frameformat in a l2cs gmpls rnvironment. >>> > >>> > >>> > >>> > >>> > >>> > >>> > Dear Jaihyung, >>> > >>> > you mention several options that could be used as a label. >>> > Most of them have strong impacts on the data plane and >>> > introduce new layer networks. Is this in the scope of CCAMP? >>> > What is the goal, to introduce a control plane for a existing >>> > L2 technology or introduce a new L2 technology? What would be >>> > the benefit of such a new L2 co switching technology compared >>> > to Ethernet over MPLS as defined by PWE3 and L2VPN? >>> > You mentioned to use the MAC address or part of it as a >>> > label. The TRILL working group is defining shortest path >>> > routing for Ethernet. How would this fit together? >>> > >>> > >>> > Regards >>> > >>> > Juergen >>> > >>> > >>> > >>> > >>> > >>> > > -----Original Message----- >>> > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] >>> > > Sent: Saturday, July 23, 2005 6:19 AM >>> > > To: Heiles Juergen; per@defero.se >>> > > Cc: ccamp@ops.ietf.org >>> > > Subject: RE: Frameformat in a l2cs gmpls rnvironment. >>> > > >>> > > >>> > > Dear Juergen, Par and all, >>> > > >>> > > The proposed framework of L2SC was not intended to suggest >>> > > any specific solution. It is still an open question. >>> > > As Adrian noted, it only aims to present requirements >>> > > as in scenarios that may drive people toward next >> >>step in many >> >>> > > viable solutions. >>> > > >>> > > In fact, there are other options we may also >> >>consider for L2 >> >>> > > label encoding. >>> > > IEEE 802.1 Ethernet bridges forward frames based >> >>on 48bits of >> >>> > > MAC address, >>> > > and additionally using VLAN tag. >>> > > >>> > > When the purpose of GMPLS control over Ethernet is not to >>> > > create new dataplane >>> > > but to utilize IEEE 802.1 bridge architecture, we may >>> > consider using >>> > > one of the two forwarding methods, MAC forwarding or VLAN >>> > forwarding. >>> > > (and perhaps any other combination of fields in >> >>MAC, but I'll >> >>> > > not discuss it) >>> > > >>> > > Use of VLAN ID for label encoding may automate >> >>VLAN configuration >> >>> > > using IP protocols. However, GMPLS protocol cannot use the >>> > > field exclusively >>> > > because public/private operators already use VLAN >> >>for various >> >>> > > purpose. >>> > > There is a potential conflict with existing use of VLAN and >>> > > GMPLS use of VLAN label. >>> > > Furthermore, scalability of VLAN ID has been >> >>frequently noted >> >>> > > as weakness >>> > > because the size of VLAN ID is at most 4096 (12bit). >>> > > >>> > > The scalability may be improved if the scope of >> >>VLAN label is >> >>> > > confined to >>> > > link-local, and some additional swapping function >> >>of VLAN ID >> >>> > > is introduced in >>> > > Internal Sublayer [802.1D] of GMPLS implemented switch. >>> > > However, this will only be effective when the configuration >>> > > of network >>> > > is mesh structure that multiple LSP paths exist. If the >>> > > configuration of network >>> > > is star or tree shape, as normal configuration of >> >>access network, >> >>> > > LSPs concentrate in root node and total number of >>> > > LSPs that the network can hold still be limited by >> >>available >> >>> > > label space >>> > > at a few root links. >>> > > >>> > > For these reasons, I do not think any form of VLAN >> >>ID label is >> >>> > > an appropriate choice for layer-2 label encoding. >>> > > >>> > > There are some other proposals assuming new >> >>assignment of Ethernet >> >>> > > Length/Type value (e.g. new TPID in VLAN tag) and >> >>re-definition of >> >>> > > information fields placed between 802.3 MAC header >> >>and IP packet. >> >>> > > In this case, only the format of VLAN tag or >> >>extended VLAN tag >> >>> > > is borrowed, however, inside the switching hardware, the >>> > > filter and relay, >>> > > etc. are totally different new dataplane switch. >>> > > >>> > > I do not see such approach is a GMPLS >> >>implementation for Ethernet >> >>> > > because the core switching technique is not 802.1 Ethernet >>> > > bridge at all. >>> > > >>> > > The other option we may consider is using MAC address filed >>> > as below. >>> > > >>> > > >>> > > +-------+-------+-------+-------+-------+-------+ >>> > > | 1byte | 2byte | 3byte | 4byte | 5byte | 6byte | >>> > > +-------+-------+-------+-------+-------+-------+ >>> > > +-----------------------+-----------------------+ >>> > > | OUI Prefix (=GMPLS) | DA-label (24bit) | >>> > > +-----------------------+-----------------------+ >>> > > | OUI Prefix (=GMPLS) | SA-label (24bit) | >>> > > +---------------+-------+-----------------------+ >>> > > | Length/Type | >>> > > +---------------+ >>> > > >>> > > >>> > > IEEE is designated by the ISO Council to act as the >>> > > registration authority >>> > > for the higher three-octet of OUI number in the MAC address >>> > > to be used by manufacturer. Ethernet manufacturer >> >>may generate >> >>> > > global unique MAC address using the OUI prefix and >> >>address block of >> >>> > > lower three-octet (24bit). Taking advantage of the >>> > addressing scheme, >>> > > GMPLS may use the lower three-octet exclusively if a unique >>> > > OUI number >>> > > is reserved for the protocol. With this labeling >> >>scheme, GMPLS will >> >>> > > control MAC forwarding entry, not VLAN table. >>> > > >>> > > All Ethernet frames controlled by GMPLS will have identical >>> > > OUI number >>> > > that they can easily be distinguished from other >> >>Ethernet frames. >> >>> > > In principle, the label lookup hardware is >> >>identical to MAC lookup >> >>> > > hardware in this labeling scheme. Therefore GMPLS >> >>implemented >> >>> > > switch may still function as normal Ethernet >> >>bridge to the frames >> >>> > > that OUI number is not GMPLS. This also facilitates GMPLS >>> > implemented >>> > > switches being deployed in operating Ethernet with minimum >>> > > service disruption. >>> > > >>> > > Note also that above proposed label encoding >> >>method is transparent >> >>> > > to the use of Ethernet Length/Type field. End-user >> >>device may use >> >>> > > the Length/Type field as defined in IEEE 802.3 protocol. >>> > > It also allows network operators configure VLAN >> >>for their own >> >>> > > purpose. >>> > > When IEEE 802.1p is used in conjunction with >> >>L2-LSP, the priority >> >>> > > field of VLAN tag can also be used for imposing >> >>consistent TE >> >>> > > policy in >>> > > legacy switches and GMPLS switches. >>> > > >>> > > Any way, my conclusion is, there are other options >> >>we may consider, >> >>> > > and this issue is still open to discuss. >>> > > >>> > > Thanks, >>> > > >>> > > Sincerely, >>> > > >>> > > Jaihyung >>> > > >>> > > >>> > > >>> > > >>> > > >>> > > -----?? ???----- >>> > > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com> >>> > > ?? ??: 2005-07-22 ?? 10:23:43 >>> > > ?? ??: "Loa Andersson" <loa@pi.se>, >> >>"richard.spencer@bt.com" >> >>> > > <richard.spencer@bt.com> >>> > > ??: "per@defero.se" <per@defero.se>, "ccamp@ops.ietf.org" >>> > > <ccamp@ops.ietf.org> >>> > > ??: RE: Frameformat in a l2cs gmpls rnvironment. >>> > > >>> > > >>> > > >>> > > >>> > > >>> > > >>> > > Loa, >>> > > >>> > > I interpret the ID as a proposal to use GMPLS for >> >>VLAN setup. >> >>> > > So GMPLS and VLAN to not compete. The VLAN is at the data >>> > > plane and GMPLS at the control plane. The question >> >>is how and >> >>> > > should different control plane techniques like GMPLS and >>> > > (GVRP and STB) work together? >>> > > >>> > > Regards >>> > > >>> > > Juergen >>> > > >>> > > >>> > > > -----Original Message----- >>> > > > From: owner-ccamp@ops.ietf.org >>> > > > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of >> >>Loa Andersson >> >>> > > > Sent: Friday, July 22, 2005 2:53 PM >>> > > > To: richard.spencer@bt.com >>> > > > Cc: per@defero.se; ccamp@ops.ietf.org >>> > > > Subject: Re: Frameformat in a l2cs gmpls rnvironment. >>> > > > >>> > > > Richard, >>> > > > >>> > > > I agree to most of this. Since we have doubts >> >>about the viability >> >>> > > > of taking GMPLS all the way to end-user or >> >>enterprise I think it >> >>> > > > would be good, from a wg perspective, if we >> >>agreed to solve the >> >>> > > > core network problems first. >>> > > > >>> > > > Do you have any comment on the requirement to >> >>run both VLANs and >> >>> > > > GMPLS on the same switch? >>> > > > >>> > > > /Loa >>> > > > >>> > > > richard.spencer@bt.com wrote: >>> > > > > Regardless of whether or not a switch is >> >>directly connected >> >>> > > > > to hosts, it must be able to forward packets using the >>> > > > connectionless >>> > > > > Ethernet data plane. This is due to the fundamental >>> > > > requirement that >>> > > > > for GMPLS switches to be able to exchange control >>> > > information with >>> > > > > each other, a data plane for control traffic >> >>must be present. >> >>> > > > > This is akin to using the IP data plane for MPLS >>> > > signalling in an >>> > > > > IP/MPLS network. An alternative would be to >> >>use a static >> >>> > > reserved >>> > > > > L2-LSP for control traffic in the same way >> >>that reserved >> >>> > > VPI/VCIs >>> > > > > are used for PNNI signalling in ATM. >>> > > > > >>> > > > > Regarding connecting hosts to GMPLS switches, I >>> > personally don't >>> > > > > think extending L2-LSPs into the >> >>enterprise/home network is >> >>> > > > > commercially viable. However, if you do want to use >>> > > GMPLS switches >>> > > > > in the home/enterprise network and for some reason >>> > don't want to >>> > > > > extend L2-LSPs down to the host then you will not be >>> > performing >>> > > > > normal Ethernet Mac address switching anyway. >> >>Instead you will >> >>> > > > > need some kind of policy on the switch that maps >>> > connectionless >>> > > > > Ethernet packets (e.g. based on MAC src/dest, 802.1p, >>> > VLAN) to a >>> > > > > L2-LSP. This is because multiple L2-LSPs to the same >>> > destination >>> > > > > (e.g. a gateway router) may exist for >> >>different services/flows >> >>> > > > > (e.g. video download, VoIP call, etc.). >>> > > > > >>> > > > > Regards, >>> > > > > Richard >>> > > > > >>> > > > > >>> > > > >>-----Original Message----- >>> > > > >>From: owner-ccamp@ops.ietf.org >>> > > [mailto:owner-ccamp@ops.ietf.org]On >>> > > > >>Behalf Of Par Mattsson >>> > > > >>Sent: 22 July 2005 11:42 >>> > > > >>To: Loa Andersson >>> > > > >>Cc: ccamp@ops.ietf.org >>> > > > >>Subject: Re: Frameformat in a l2cs gmpls rnvironment. >>> > > > >> >>> > > > >> >>> > > > >> >>> > > > >>>Per and Dimitri, >>> > > > >>> >>> > > > >>>I would like to come down stronger than that, >> >>for me it is >> >>> > > > >>>a very strong requirement that the same >> >>switch can handle >> >>> > > > >>>both VLANs and GMPLs trafic correctly. I >> >>can't dsee how that >> >>> > > > >>>could be done if using the VLAN tpid to indicate GMPLS >>> > > > >>>traffic. >>> > > > >> >>> > > > >>If you ever want that same switch to handle >> >>traffic for a >> >>> > > directly >>> > > > >>connected host (not to uncommen) you would >> >>want that to >> >>> > > use normal >>> > > > >>ethernet macaddress switching. So of course >> >>you do not want >> >>> > > > to have to >>> > > > >>choose between vlan and gmpls, you would want >> >>both at the >> >>> > > same time. >>> > > > >> >>> > > > >>/per >>> > > > >> >>> > > > >> >>> > > > >> >>> > > > >>>/Loa >>> > > > >>> >>> > > > >>>Par Mattsson wrote: >>> > > > >>> >>> > > > >>>>>hi par, one of the possibilities that has >> >>been considered >> >>> > > > >> >>> > > > >>to cope with >>> > > > >> >>> > > > >>>>>this requirement is to use a dedicated TPID for the >>> > > > >> >>> > > > >>Ethernet labeled >>> > > > >> >>> > > > >>>>>frames; this would allow differentiated >> >>processing with >> >>> > > > non-labeled >>> > > > >>>>>framesthanks. >>> > > > >>>> >>> > > > >>>> >>> > > > >>>>That seems to make more sence. If that frame is to be >>> > > sized like a >>> > > > >>>>802.1q >>> > > > >>>>frame. There is not that much space left to >> >>a label. Or is >> >>> > > > >> >>> > > > >>the demand to >>> > > > >> >>> > > > >>>>use jumboframes ? >>> > > > >>>>Has there been any discussion on labelstacking, and >>> > > > mainly where to >>> > > > >>>>place >>> > > > >>>>the information? >>> > > > >>>> >>> > > > >>>>Regards. >>> > > > >>>>Per >>> > > > >>>> >>> > > > >>>> >>> > > > >>>> >>> > > > >>> >>> > > > >>> >>> > > > >>>-- >>> > > > >>>Loa Andersson >>> > > > >>> >>> > > > >>>Principal Networking Architect >>> > > > >>>Acreo AB phone: >> >>+46 8 632 77 14 >> >>> > > > >>>Isafjordsgatan 22 mobile: >> >>+46 739 81 21 64 >> >>> > > > >>>Kista, Sweden email: >>> > > loa.andersson@acreo.se >>> > > > >>> loa@pi.se >>> > > > >>> >>> > > > >> >>> > > > >> >>> > > > >> >>> > > > > >>> > > > >>> > > > >>> > > > -- >>> > > > Loa Andersson >>> > > > >>> > > > Principal Networking Architect >>> > > > Acreo AB phone: +46 8 >> >>632 77 14 >> >>> > > > Isafjordsgatan 22 mobile: +46 >> >>739 81 21 64 >> >>> > > > Kista, Sweden email: >> >>loa.andersson@acreo.se >> >>> > > > loa@pi.se >>> > > > >>> > > >>> > > >>> > > >>> > > >>> > >>> > >>> > >>> >> >>-- >>Loa Andersson >> >>Principal Networking Architect >>Acreo AB phone: +46 8 632 77 14 >>Isafjordsgatan 22 mobile: +46 739 81 21 64 >>Kista, Sweden email: loa.andersson@acreo.se >> loa@pi.se >> > > -- Loa Andersson Principal Networking Architect Acreo AB phone: +46 8 632 77 14 Isafjordsgatan 22 mobile: +46 739 81 21 64 Kista, Sweden email: loa.andersson@acreo.se loa@pi.se Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 28 Jul 2005 10:42:06 +0000 Reply-To: <i-nishioka@cb.jp.nec.com> From: "Itaru Nishioka" <i-nishioka@cb.jp.nec.com> To: <ccamp@ops.ietf.org> Subject: RE: draft-otani-ccamp-interas-gmpls-te-03.txt Date: Thu, 28 Jul 2005 19:48:48 +0900 Organization: nec MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Thread-Index: AcWN3Lk8oJ25CktfRAW0r7o2p2kiVwFgdx6w Message-Id: <iss.25ee39ab.3d16.42e8b644.4d443.3d@mail.jp.nec.com> This draft gives a good start point for "Distributing Reachability and TE Information" in inter-domain framework (Section 4 of draft-ietf-ccamp-inter-domain-framework-04.txt). I also hope re-charter as soon as possible. Thanks, itaru. > -----Original Message----- > From: owner-ccamp@ops.ietf.org > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Tomohiro Otani > Sent: Thursday, July 21, 2005 7:09 PM > To: ccamp@ops.ietf.org > Subject: draft-otani-ccamp-interas-gmpls-te-03.txt > > Hi everyone, > > We updated the GMPLS inter-domain TE req. draft. Since last > meeting, we believe this document has been already stable and > well-recognized, however, per the last meeting discussion, we > have to wait for the charter being revised, in order to be a > WG document (This must be initiated by chairs). > > Until then, we continue to update the contents and indeed > welcome your feedback and comments through this list or > during the meeting in Paris. > > Regards, > > tomo Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 28 Jul 2005 09:30:13 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: I-D ACTION:draft-ietf-ccamp-inter-domain-rsvp-te-01.txt Date: Thu, 28 Jul 2005 11:29:25 +0200 Message-ID: <D109C8C97C15294495117745780657AE02EDBBC4@ftrdmel1.rd.francetelecom.fr> Thread-Topic: I-D ACTION:draft-ietf-ccamp-inter-domain-rsvp-te-01.txt Thread-Index: AcWMm75CDqtnEDxHRpSUDBa2yZwJWwGt1cgg From: "LE ROUX Jean-Louis RD-CORE-LAN" <jeanlouis.leroux@francetelecom.com> To: "Arthi Ayyangar" <arthi@juniper.net>, <jpv@cisco.com> Cc: <ccamp@ops.ietf.org> Hi Arthi, JP A few comments on this draft: -In section 5.2, I'm not sure that path p3 is a valid one: "RO-X1-ASBR1(loose)-ASBR4(strict)-ASBR7(strict)-ASBR9(loose)-R6. What does ERO sub-objects for ASBR4 and ASBR5 represent? Actually this cannot be an interface or router id, because the ERO would = be invalid (cannot be strict as ASBR4 and ASBR7 ar not directly = connected). The only way to proceed in this way is to include = ERO-sub-objects corresponding to FA-LSP or LSP-Segment TE-links = (numbered or unnumbered tunnel interface id).=20 One could assume that RO would obtain such ERO including FA-LSP or = LSP-segment as a result of a PCE based compuation... On receip of a Path message the ERO subobject would allow to directly = determine the underlying FA-LSP or LSP-Segment... By the way it would also be good to ilusrate the case of abstract nodes, = e.g. a path made of AS number ERO subobjects... -It would be good to add some text related to failure detection, at the = end of 6.1.3. In case of stitching or nesting, the PLR could rely on = BFD-MPLS so as to rapidliy detect FA-LSP/LSP-Segment tail-end failure. = This would help ensuring fast recovery... -In section 6.2 the FRR considerations described in 6.1.3 for stiching = and nesting, actually also apply to GMPLS Segment Recovery. -Section 7 would benefif from some words on global reoptimization (ie = with change of border routers) Even if this is more a path computation issue this may have some impact = on signaling... -And two minor edits Section 5.1.1: issue with B3 path, in the figure ASBR6 and ASBR7 are not = connnected.=20 Section 8: [INTER-AS-TE-REQ] instead of [INTER-AREA-TE-REQS] Hope this helps, Best Regards JL > -----Message d'origine----- > De : owner-ccamp@ops.ietf.org=20 > [mailto:owner-ccamp@ops.ietf.org] De la part de=20 > Internet-Drafts@ietf.org > Envoy=E9 : mardi 19 juillet 2005 21:50 > =C0 : i-d-announce@ietf.org > Cc : ccamp@ops.ietf.org > Objet : I-D ACTION:draft-ietf-ccamp-inter-domain-rsvp-te-01.txt=20 >=20 > A New Internet-Draft is available from the on-line=20 > Internet-Drafts directories. > This draft is a work item of the Common Control and=20 > Measurement Plane Working Group of the IETF. >=20 > Title : Inter domain GMPLS Traffic=20 > Engineering - RSVP-TE extensions > Author(s) : A. Ayyangar, J. Vasseur > Filename : draft-ietf-ccamp-inter-domain-rsvp-te-01.txt > Pages : 17 > Date : 2005-7-19 > =09 > This document describes extensions to Generalized=20 > Multi-Protocol Label Switching (GMPLS) Resource ReserVation=20 > Protocol - Traffic Engineering > (RSVP-TE) signaling required to support mechanisms for the=20 > establishment and maintenance of GMPLS Traffic Engineering=20 > (TE) Label Switched Paths (LSPs), both packet and non-packet,=20 > that traverse multiple domains. For the purpose of this=20 > document, a domain is considered to be any collection of=20 > network elements within a common realm of address space or=20 > path computation responsibility. Examples of such domains=20 > include Autonomous Systems, IGP areas and GMPLS overlay networks. >=20 > A URL for this Internet-Draft is: > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-inter-dom > ain-rsvp-te-01.txt >=20 > To remove yourself from the I-D Announcement list, send a=20 > message to i-d-announce-request@ietf.org with the word=20 > unsubscribe in the body of the message. =20 > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > to change your subscription settings. >=20 >=20 > Internet-Drafts are also available by anonymous FTP. Login=20 > with the username "anonymous" and a password of your e-mail=20 > address. After logging in, type "cd internet-drafts" and then > "get draft-ietf-ccamp-inter-domain-rsvp-te-01.txt". >=20 > A list of Internet-Drafts directories can be found in=20 > http://www.ietf.org/shadow.html or=20 > ftp://ftp.ietf.org/ietf/1shadow-sites.txt >=20 >=20 > Internet-Drafts can also be obtained by e-mail. >=20 > Send a message to: > mailserv@ietf.org. > In the body type: > "FILE=20 > /internet-drafts/draft-ietf-ccamp-inter-domain-rsvp-te-01.txt". > =09 > NOTE: The mail server at ietf.org can return the document in > MIME-encoded form by using the "mpack" utility. To use this > feature, insert the command "ENCODING mime" before the "FILE" > command. To decode the response(s), you will need "munpack" or > a MIME-compliant mail reader. Different MIME-compliant=20 > mail readers > exhibit different behavior, especially when dealing with > "multipart" MIME messages (i.e. documents which have been split > up into multiple messages), so check your local documentation on > how to manipulate these messages. > =09 > =09 > Below is the data which will enable a MIME compliant mail reader > implementation to automatically retrieve the ASCII version of the > Internet-Draft. >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 28 Jul 2005 09:00:49 +0000 Message-ID: <42E89E9E.4040608@chello.nl> Date: Thu, 28 Jul 2005 11:00:14 +0200 From: Huub van Helvoort <hhelvoort@chello.nl> User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax) MIME-Version: 1.0 To: ccamp <ccamp@ops.ietf.org> Subject: Test - please ignore Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sorry for this. Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 28 Jul 2005 08:41:58 +0000 Thread-Topic: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet thread-index: AcWTUFmFeqIRFVuSR82NxoJR0sjwhg== Reply-To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> From: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> To: "Ali Sajassi \(sajassi\)" <sajassi@cisco.com>, "Heiles Juergen" <juergen.heiles@siemens.com>, <per@defero.se> Cc: <ccamp@ops.ietf.org> Bcc: Subject: RE: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet Date: Thu, 28 Jul 2005 17:42:55 +0900 Comment: GQ19@|@ZEk=E?,18?x, BcN1b<z:P<.F@, 4c4g Message-ID: <852501c59350$598ac2e0$8310fe81@email1> MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit Content-Class: urn:content-classes:message Hi, Ali sorry for late in this response. a bit busy at preparing trip to Paris. see in-line [snip] >>> > >>> >> 3. use lower 3 bytes of MAC address for L2 label encoding. >>> > >>> >This approach has several issues: >>> >1) how does MAC addresses get distributed among the bridges ? >>> >How does this distribution works during a link or node failure ? >>> >How does it recover from a link or node failure ? >> >> Ali, >> The questions are really implementation dependent. >> Can we defer this discussion after Paris meeting? >> However, I see the solution will not be so different than >> what RSVP-TE has been done for other transport network. > >These questions are not implementation dependent. IEEE bridges have a >well established procedures for handling link and node failures based on >RSTP/MSTP and "independent" interaction between these procedures and >RSVP-TE is questionable at best. If you think this is not the case, then >I encourage you to specify how this interaction works. Honestly, it's not clear for me what is the point you trying to make. RSVP-TE works on well-established IP routing protocols, and MPLS/GMPLS define LSP restoration mechanism well. I believe you understand well how RSVP-TE distributes generic label between nodes in the event of failure. I'm a bit surprised a cisco guy saying such doubt on RSVP-TE. I still feel this is not necessary talk at this stage. Clear answer can be given after we settle down solution. >> >> >2) Will you assign a separate provider MAC for each customer MAC. >> >If so, then how do you address MAC scalability. >> >If not, then how do you take care of 1-to-N mapping between provider >> >MAC address and customer MAC addresses and how do you handle >> multipoint >> >connections. >> >> >> I think you are asking the issue of L2 label stacking or LSP >> aggregation, or perhaps some VPN application similar to 802.1ad. >> There are some opinions on whether label stacking is >> necessary in L2SC domain. This is actually the question >> Adrian keep asking us. > >Since you are using provider MAC address as a label, then by label >stacking, I presume you mean encapsulating customers MAC addresses >inside the provider MAC. That'll depend on service function configured in ingress node. 1. If ingress is PSC, (i.e. L3/L4 switching or MPLS), provider MAC header ?will encapsulate user IP packet or MPLS packet 2. If ingress provides L2 VPN service, provider MAC will encapsulate ?user MAC, thus it'll be MAC-in-MAC as you said. 3. If ingress provides end-to-end L2-LSP connection, it'll simply change ?user MAC address to GMPLS prefixed L2 label at ingress. this case only be valid in enterprise/access network. I told it before in reply to Juergen. please read past mails carefully. Scalability performance must be discussed respectively at the three cases. >If you want to go to that extend, then why not >just simply use IEEE 802.1ah that provide both stacking of MAC addresses >as well as it provides you with a 20-bit tag for service instance >identification. No, because of at least four reasons, 1. assuming new definition of dataplane is not in CCAMP scope, as you also agreed. 2. such machine will only be compatible with the same kind. ?backward compatibility to 802.1D/Q bridge is important goal to us, ?so that the GMPLS implemented switch may inter-mix with ?legacy bridges. 3. we are not intending to develop it just for metro-LAN service. as I told in above, several different service application, such as campus Internet service, MPLS-LSP backbone, etc. can also be considered. 4. we can achieve far more flexible, compatible, scalable solution than your proposal if we combine L2 label with MPLS shim labels. ?besides, if the format you propose is something like ?[MAC][20bit-tag][ IP ][ .. data ...] ?than is it in fact a MPLS label switching? ?if it is your proposal, why not just extend MPLS specs a bit, ?rather than creating new one? >> I personally think that we can achieve LSP aggregation with >> the aid of MPLS shim label, however I think this is too >> premature discussion. >> All I can say at the moment is, lower 3 octet of MAC address >> gives us 16M size of label space and this is enough to >> implement link-local, site-local or even some domain-local LSPs. >> Between the L2-LSP boundary, it needs further discussion. >> >It is one thing to have the address space, it is another thing to have >proper management and control protocols to manage that space. The big >question is what type of control and management protocols you intend to >use for that space and how it interacts with the existing IEEE control >and management protocols. Also, if it is one-to-one mapping, then we are >back to square one regarding MAC address scalability. control ? ==> of course, GMPLS. management? ==> not in my interest. it seems you are talking about address management mechanism. Precedence of such solution is far later. >> > >> >3) how do handle customer FCS retention when doing this MAC >> translation >> >stuff ?? >> >> I haven't seen any application that requires retention of >> original FCS image. >> 802.1D specification recommends recalculation of FCS when >> there is any change in MAC. I see no problem if FCS is >> recalculated at the label swapping switch. > >Please refer to all the discussions on the PWE3 mailing list regarding >FCS retention. There are certain applications that require it and that's >why there is a draft for it. ok, thanks, I found there's some people worrying about this, however, such worry is not in the scope of my idea. It is even not clear if there's requirement exist for PWE service using L2-LSP. [snip] >> >> If we define the scope of L2 label as site-local or >> domain-local, we don't need such label swapping operation. >> However, if we want link-local label, and find out that it is >> very useful >> (* actually I do think it'll be very useful *), I think we >> need to consult such modification with IEEE. >> Technically speaking, I don't think replacing the label value >> after normal MAC lookup will serious change hardware architecture. >> Also, it doesn't break conformance rule. >> Did any 802.1 spec. say that input MAC must be identical to >> output MAC? > >There is no MAC address swapping or translation in .1D, .1Q, or its >future standards. If there is MAC address translation, then you can run >into quite a few unwanted side affects. so, which conformance rule does it break in 802.1D/Q spec. if we translate MAC address ? In fact, some bridge vender sells MAC translating switch for security purpose because it can hide internal MAC information. It doesn't cause any side-effect. You need to be specific what side-effect it may cause. [snip] >> > >> >How does the redundancy work in here. Are you using bridge >> protocol in >> >conjunction with GMPLS protocol. If you do, then there are bunch of >> >issues. >> >> >> Yes I do. >> I intend GMPLS implemented switches being deployed in >> operating network >> as normal Ethernet switch that perform bridge protocols. >> When they detect (or snoop) RSVP-TE message, they establish L2-LSP >> and selectively forward those GMPLS prefixed frames via the L2-LSP. >> The message path that RSVP-TE is exchanged, thus L2-LSPs are >> established, does not necessarily be identical to spanning tree path >> because L2-LSPs can be established via shortest path. >> I welcome further discussion on this issue, but I hope to continue it >> after Paris meeting, if you may. > >If RSVP-TE traffic uses link-state based protocols and bridged traffic >uses distance-based protocols (e.g., RSTP) and these traffic operate >independently from each other and both types of traffic come through the >same port, then how does a standard IEEE bridge know how to discern >these packets and which packets to take action on ? - e.g., what IEEE >procedures do you recommend to be used ?? If you are talking about >different physical ports, then you are talking about different >independent networks with different set of links and topology for each >which is not the intent here. Normal MAC lookup is sufficient. MAC entries for normal frames are learned as bridge relay function. The frames are forwarded via active ports as usual. MAC entries for GMPLS frames are configured by GMPLS protocols. GMPLS frames may be forwarded via active ports as well as some standby ports that are not used by spanning tree. Of course in this case, GMPLS control messages need to be forwarded via standby ports. Thanks Jaihyung -----?? ???----- ?? ??: "Ali Sajassi (sajassi)" <sajassi@cisco.com> ?? ??: 2005-07-27 ?? 12:00:53 ?? ??: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles Juergen" <juergen.heiles@siemens.com>, "per@defero.se" <per@defero.se> ??: "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> ??: RE: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet Jaihyung, Please refer to my comments inline ... [snip] > > > >> 3. use lower 3 bytes of MAC address for L2 label encoding. > > > >This approach has several issues: > >1) how does MAC addresses get distributed among the bridges ? > >How does this distribution works during a link or node failure ? > >How does it recover from a link or node failure ? > > > > > Ali, > The questions are really implementation dependent. > Can we defer this discussion after Paris meeting? > However, I see the solution will not be so different than > what RSVP-TE has been done for other transport network. > These questions are not implementation dependent. IEEE bridges have a well established procedures for handling link and node failures based on RSTP/MSTP and "independent" interaction between these procedures and RSVP-TE is questionable at best. If you think this is not the case, then I encourage you to specify how this interaction works. > > >2) Will you assign a separate provider MAC for each customer MAC. > >If so, then how do you address MAC scalability. > >If not, then how do you take care of 1-to-N mapping between provider > >MAC address and customer MAC addresses and how do you handle > multipoint > >connections. > > > I think you are asking the issue of L2 label stacking or LSP > aggregation, or perhaps some VPN application similar to 802.1ad. > There are some opinions on whether label stacking is > necessary in L2SC domain. This is actually the question > Adrian keep asking us. Since you are using provider MAC address as a label, then by label stacking, I presume you mean encapsulating customers MAC addresses inside the provider MAC. If you want to go to that extend, then why not just simply use IEEE 802.1ah that provide both stacking of MAC addresses as well as it provides you with a 20-bit tag for service instance identification. > I personally think that we can achieve LSP aggregation with > the aid of MPLS shim label, however I think this is too > premature discussion. > All I can say at the moment is, lower 3 octet of MAC address > gives us 16M size of label space and this is enough to > implement link-local, site-local or even some domain-local LSPs. > Between the L2-LSP boundary, it needs further discussion. > It is one thing to have the address space, it is another thing to have proper management and control protocols to manage that space. The big question is what type of control and management protocols you intend to use for that space and how it interacts with the existing IEEE control and management protocols. Also, if it is one-to-one mapping, then we are back to square one regarding MAC address scalability. > > > > >3) how do handle customer FCS retention when doing this MAC > translation > >stuff ?? > > > > > I haven't seen any application that requires retention of > original FCS image. > 802.1D specification recommends recalculation of FCS when > there is any change in MAC. I see no problem if FCS is > recalculated at the label swapping switch. Please refer to all the discussions on the PWE3 mailing list regarding FCS retention. There are certain applications that require it and that's why there is a draft for it. > > > >> 802.1Q bridge forwards Ethernet frames using two dataplane tables > >> - MAC forwarding table and VLAN forwarding table. > >> Bridge control protocols, such as GARP, GVRP, GMRP, > manipulate one of > >> the two dataplane entities. > > > >This is not correct. Bridges don't have different databases for MAC > >forwarding and VLAN forwarding. They have filtering datebase(s) with > >corresponding IDs (FIDs). VLAN IDs get mapped to FIDs based on > >operation mode (IVL versus SVL). Then once a filtering database is > >identified, then a MAC address lookup is performed for forwarding. > >Bridge control protocols use a designated MAC addressed and based on > >these MAC addressed, the bridge knows how to process the frame. > > > oh, come on . > I explained it in conceptual level and you re-described it in > implementation level. The contextual meaning is same. > I believe you don't deny the existence of two types of > filtering databases, for MAC and VID, and the control of GARP > over these databases. > This is unnecessary argument, I think. > No, I didn't describe the implementation but instead I described it as specified in .1D and .1Q. However, I see what you were trying to say and I think it was just a poor choice of words. [snip] > > If we define the scope of L2 label as site-local or > domain-local, we don't need such label swapping operation. > However, if we want link-local label, and find out that it is > very useful > (* actually I do think it'll be very useful *), I think we > need to consult such modification with IEEE. > Technically speaking, I don't think replacing the label value > after normal MAC lookup will serious change hardware architecture. > Also, it doesn't break conformance rule. > Did any 802.1 spec. say that input MAC must be identical to > output MAC? There is no MAC address swapping or translation in .1D, .1Q, or its future standards. If there is MAC address translation, then you can run into quite a few unwanted side affects. [snip] > > > [snip] > > > >How does the redundancy work in here. Are you using bridge > protocol in > >conjunction with GMPLS protocol. If you do, then there are bunch of > >issues. > > > Yes I do. > I intend GMPLS implemented switches being deployed in > operating network > as normal Ethernet switch that perform bridge protocols. > When they detect (or snoop) RSVP-TE message, they establish L2-LSP > and selectively forward those GMPLS prefixed frames via the L2-LSP. > The message path that RSVP-TE is exchanged, thus L2-LSPs are > established, does not necessarily be identical to spanning tree path > because L2-LSPs can be established via shortest path. > I welcome further discussion on this issue, but I hope to continue it > after Paris meeting, if you may. If RSVP-TE traffic uses link-state based protocols and bridged traffic uses distance-based protocols (e.g., RSTP) and these traffic operate independently from each other and both types of traffic come through the same port, then how does a standard IEEE bridge know how to discern these packets and which packets to take action on ? - e.g., what IEEE procedures do you recommend to be used ?? If you are talking about different physical ports, then you are talking about different independent networks with different set of links and topology for each which is not the intent here. Cheers, Ali > > > Thanks > > Jaihyung > > > -----?? ???----- > ?? ??: "Ali Sajassi (sajassi)" <sajassi@cisco.com> > ?? ??: 2005-07-26 ?? 8:13:56 > ?? ??: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles > Juergen" <juergen.heiles@siemens.com>, "per@defero.se" > <per@defero.se> > ??: "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> > ??: RE: Frameformat in a l2cs gmpls rnvironment - Issues w/ > GMPLS Ethernet > > > > > > > > > Jaihyung, > > It seems to me that your primary objective is to use the existing > Ethernet switches with new control plane (GMPLS) to setup TE > paths for > different applications. And you mentioned three approaches for doing > this in your email. However, all three of them have issues > and are not > viable. Lets go over them ... > > > > > In below mail, I discussed about three proposed approaches. > > > > > > 1. using VLAN tag as it is for L2 label encoding. > > As you know this approach has scalability issue and you are > limited to > 4K services in the network and if you try to use the upcoming IEEE > 802.1ad standard to make VLANs, interface specific, then you > will need > new switches and not existing Ethernet switches. Even if you > use these > newer switches, you will still be limited to 4K services per > interface > and it would defeat the objective of using existing switches. > > > > > 2. defining new protocol ID (TPID) and borrow VLAN tag format, > > ?or extended VLAN tag to implement label swapping. > > > > This approach requires new data plane functionality to be implemented > (which defeats the objective of using existing switches). > > > 3. use lower 3 bytes of MAC address for L2 label encoding. > > > > This approach has several issues: > 1) how does MAC addresses get distributed among the bridges ? > How does > this distribution works during a link or node failure ? How does it > recover from a link or node failure ? > > 2) Will you assign a separate provider MAC for each customer > MAC. If so, > then how do you address MAC scalability. If not, then how do you take > care of 1-to-N mapping between provider MAC address and customer MAC > addresses and how do you handle multipoint connections. > > 3) how do handle customer FCS retention when doing this MAC > translation > stuff ?? > > > > > 802.1Q bridge forwards Ethernet frames using two dataplane tables > > - MAC forwarding table and VLAN forwarding table. > > Bridge control protocols, such as GARP, GVRP, GMRP, > > manipulate one of the two dataplane entities. > > This is not correct. Bridges don't have different databases for MAC > forwarding and VLAN forwarding. They have filtering datebase(s) with > corresponding IDs (FIDs). VLAN IDs get mapped to FIDs based > on operation > mode (IVL versus SVL). Then once a filtering database is identified, > then a MAC address lookup is performed for forwarding. Bridge control > protocols use a designated MAC addressed and based on these MAC > addressed, the bridge knows how to process the frame. > > > > > Similarly, option 1 and 3 are about which one of two > > dataplane entities > > GMPLS protocol should control on behalf of bridge control > protocols. > > The two proposals do not intend to modify bridge behavior > > seriously, such as MAC learning, aging, filtering. > > Therefore, the approaches 1 and 3 are in the scope of CCAMP. > > If that's the intention, you may find out soon you get more than what > you asked for :-) > > > > > (% note however, I would regard implementing label swapping > > function is acceptable change considering the Internal Sublayer > > design of 802.1D bridge.) > > > > The only way of doing this is using VID translation table > specified in > IEEE 802.1ad which requires new data-plane functionality and thus > existing bridges cannot be used !! > > Also if going the route of tag swapping, you may consider using IEEE > 802.1ah. As the editor of .1ah, I can go over the details > when I get a > chance. > > > > > However, option 2 is about what TPID code we will > > choose to implement new switching function. > > The switch of this implementation should have new hardware > > design in addition to normal bride hardware. > > I think such proposal as option 2 is out of CCAMP scope. > > Yes. > > > > > Overall objective is improving scalability, traffic > engineering (TE) > > characteristics of 802.1 bridge that it can be reliable, manageable > > enough to replace some core routers. > > Easier said than done :-) > > > The switching technique you mentioned above, such as > > Ethernet over MPLS as defined by PWE3 and L2VPN, > > are all actually router based technology, however this > > work is based on simple bridge architecture. > > Cost-effectiveness is the key differentiator. > > > > The devils are in the details and once you worked out the > details, you > will see the issues. > > > > > In access or enterprise network, capability of providing end-to-end > > L2-LSP will enable service providers policing, measuring, charging > > application flows using Ethernet based network. This will > eventually > > improve income structure and introduce new session based commercial > > service. I have been discussing this aspect in mail thread of > > title 'End-to-end L2-LSP'. Please read the mail thread and comment > > on the discussion. > > > > > > >You mentioned to use the MAC address or part of it as a label. > > >The TRILL working group is defining shortest path routing > > for Ethernet. > > >How would this fit together? > > > > > > TRILL employs new encapsulation layer outside of Ethernet frames. > > However in my proposal of option-3, I simply use the > Ethernet header > > to encode GMPLS label. > > > > The overall format of rbridge is > > [Ethernet][r-tag][Ethernet][ data..] > > It is not LSP based technology but a connectionless routing bridge. > > TTL count is very important in rbridge because frames are routed > > hop-by-hop and it may loop. > > > > Whereas in option-3 above, Ethernet frames that have certain > > OUI prefix in MAC address are forwarded via a path that > > GMPLS protocol has configured on MAC forwarding table. > > > > How does the redundancy work in here. Are you using bridge > protocol in > conjunction with GMPLS protocol. If you do, then there are bunch of > issues. If you don't, then what is it used and how does it > interact with > bridging protocols ?? > > -Ali > > > > > > Thanks > > > > Jaihyung > > > > > > > > -----?? ???----- > > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com> > > ?? ??: 2005-07-25 ?? 6:44:34 > > ?? ??: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles > > Juergen" <juergen.heiles@siemens.com>, "per@defero.se" > > <per@defero.se> > > ??: "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> > > ??: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > > > > > > > > > Dear Jaihyung, > > > > you mention several options that could be used as a label. > > Most of them have strong impacts on the data plane and > > introduce new layer networks. Is this in the scope of CCAMP? > > What is the goal, to introduce a control plane for a existing > > L2 technology or introduce a new L2 technology? What would be > > the benefit of such a new L2 co switching technology compared > > to Ethernet over MPLS as defined by PWE3 and L2VPN? > > You mentioned to use the MAC address or part of it as a > > label. The TRILL working group is defining shortest path > > routing for Ethernet. How would this fit together? > > > > > > Regards > > > > Juergen > > > > > > > > > > > > > -----Original Message----- > > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > > > Sent: Saturday, July 23, 2005 6:19 AM > > > To: Heiles Juergen; per@defero.se > > > Cc: ccamp@ops.ietf.org > > > Subject: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > > > > Dear Juergen, Par and all, > > > > > > The proposed framework of L2SC was not intended to suggest > > > any specific solution. It is still an open question. > > > As Adrian noted, it only aims to present requirements > > > as in scenarios that may drive people toward next step in many > > > viable solutions. > > > > > > In fact, there are other options we may also consider for L2 > > > label encoding. > > > IEEE 802.1 Ethernet bridges forward frames based on 48bits of > > > MAC address, > > > and additionally using VLAN tag. > > > > > > When the purpose of GMPLS control over Ethernet is not to > > > create new dataplane > > > but to utilize IEEE 802.1 bridge architecture, we may > > consider using > > > one of the two forwarding methods, MAC forwarding or VLAN > > forwarding. > > > (and perhaps any other combination of fields in MAC, but I'll > > > not discuss it) > > > > > > Use of VLAN ID for label encoding may automate VLAN configuration > > > using IP protocols. However, GMPLS protocol cannot use the > > > field exclusively > > > because public/private operators already use VLAN for various > > > purpose. > > > There is a potential conflict with existing use of VLAN and > > > GMPLS use of VLAN label. > > > Furthermore, scalability of VLAN ID has been frequently noted > > > as weakness > > > because the size of VLAN ID is at most 4096 (12bit). > > > > > > The scalability may be improved if the scope of VLAN label is > > > confined to > > > link-local, and some additional swapping function of VLAN ID > > > is introduced in > > > Internal Sublayer [802.1D] of GMPLS implemented switch. > > > However, this will only be effective when the configuration > > > of network > > > is mesh structure that multiple LSP paths exist. If the > > > configuration of network > > > is star or tree shape, as normal configuration of access network, > > > LSPs concentrate in root node and total number of > > > LSPs that the network can hold still be limited by available > > > label space > > > at a few root links. > > > > > > For these reasons, I do not think any form of VLAN ID label is > > > an appropriate choice for layer-2 label encoding. > > > > > > There are some other proposals assuming new assignment of > Ethernet > > > Length/Type value (e.g. new TPID in VLAN tag) and > re-definition of > > > information fields placed between 802.3 MAC header and IP packet. > > > In this case, only the format of VLAN tag or extended VLAN tag > > > is borrowed, however, inside the switching hardware, the > > > filter and relay, > > > etc. are totally different new dataplane switch. > > > > > > I do not see such approach is a GMPLS implementation for Ethernet > > > because the core switching technique is not 802.1 Ethernet > > > bridge at all. > > > > > > The other option we may consider is using MAC address filed > > as below. > > > > > > > > > +-------+-------+-------+-------+-------+-------+ > > > | 1byte | 2byte | 3byte | 4byte | 5byte | 6byte | > > > +-------+-------+-------+-------+-------+-------+ > > > +-----------------------+-----------------------+ > > > | OUI Prefix (=GMPLS) | DA-label (24bit) | > > > +-----------------------+-----------------------+ > > > | OUI Prefix (=GMPLS) | SA-label (24bit) | > > > +---------------+-------+-----------------------+ > > > | Length/Type | > > > +---------------+ > > > > > > > > > IEEE is designated by the ISO Council to act as the > > > registration authority > > > for the higher three-octet of OUI number in the MAC address > > > to be used by manufacturer. Ethernet manufacturer may generate > > > global unique MAC address using the OUI prefix and > address block of > > > lower three-octet (24bit). Taking advantage of the > > addressing scheme, > > > GMPLS may use the lower three-octet exclusively if a unique > > > OUI number > > > is reserved for the protocol. With this labeling scheme, > GMPLS will > > > control MAC forwarding entry, not VLAN table. > > > > > > All Ethernet frames controlled by GMPLS will have identical > > > OUI number > > > that they can easily be distinguished from other Ethernet frames. > > > In principle, the label lookup hardware is identical to > MAC lookup > > > hardware in this labeling scheme. Therefore GMPLS implemented > > > switch may still function as normal Ethernet bridge to the frames > > > that OUI number is not GMPLS. This also facilitates GMPLS > > implemented > > > switches being deployed in operating Ethernet with minimum > > > service disruption. > > > > > > Note also that above proposed label encoding method is > transparent > > > to the use of Ethernet Length/Type field. End-user device may use > > > the Length/Type field as defined in IEEE 802.3 protocol. > > > It also allows network operators configure VLAN for their own > > > purpose. > > > When IEEE 802.1p is used in conjunction with L2-LSP, the priority > > > field of VLAN tag can also be used for imposing consistent TE > > > policy in > > > legacy switches and GMPLS switches. > > > > > > Any way, my conclusion is, there are other options we may > consider, > > > and this issue is still open to discuss. > > > > > > Thanks, > > > > > > Sincerely, > > > > > > Jaihyung > > > > > > > > > > > > > > > > > > -----?? ???----- > > > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com> > > > ?? ??: 2005-07-22 ?? 10:23:43 > > > ?? ??: "Loa Andersson" <loa@pi.se>, "richard.spencer@bt.com" > > > <richard.spencer@bt.com> > > > ??: "per@defero.se" <per@defero.se>, "ccamp@ops.ietf.org" > > > <ccamp@ops.ietf.org> > > > ??: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > > > > > > > > > > > > > > > > Loa, > > > > > > I interpret the ID as a proposal to use GMPLS for VLAN setup. > > > So GMPLS and VLAN to not compete. The VLAN is at the data > > > plane and GMPLS at the control plane. The question is how and > > > should different control plane techniques like GMPLS and > > > (GVRP and STB) work together? > > > > > > Regards > > > > > > Juergen > > > > > > > > > > -----Original Message----- > > > > From: owner-ccamp@ops.ietf.org > > > > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson > > > > Sent: Friday, July 22, 2005 2:53 PM > > > > To: richard.spencer@bt.com > > > > Cc: per@defero.se; ccamp@ops.ietf.org > > > > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > > > > > > Richard, > > > > > > > > I agree to most of this. Since we have doubts about the > viability > > > > of taking GMPLS all the way to end-user or enterprise I > think it > > > > would be good, from a wg perspective, if we agreed to solve the > > > > core network problems first. > > > > > > > > Do you have any comment on the requirement to run both > VLANs and > > > > GMPLS on the same switch? > > > > > > > > /Loa > > > > > > > > richard.spencer@bt.com wrote: > > > > > Regardless of whether or not a switch is directly connected > > > > > to hosts, it must be able to forward packets using the > > > > connectionless > > > > > Ethernet data plane. This is due to the fundamental > > > > requirement that > > > > > for GMPLS switches to be able to exchange control > > > information with > > > > > each other, a data plane for control traffic must be > present. > > > > > This is akin to using the IP data plane for MPLS > > > signalling in an > > > > > IP/MPLS network. An alternative would be to use a static > > > reserved > > > > > L2-LSP for control traffic in the same way that reserved > > > VPI/VCIs > > > > > are used for PNNI signalling in ATM. > > > > > > > > > > Regarding connecting hosts to GMPLS switches, I > > personally don't > > > > > think extending L2-LSPs into the enterprise/home network is > > > > > commercially viable. However, if you do want to use > > > GMPLS switches > > > > > in the home/enterprise network and for some reason > > don't want to > > > > > extend L2-LSPs down to the host then you will not be > > performing > > > > > normal Ethernet Mac address switching anyway. > Instead you will > > > > > need some kind of policy on the switch that maps > > connectionless > > > > > Ethernet packets (e.g. based on MAC src/dest, 802.1p, > > VLAN) to a > > > > > L2-LSP. This is because multiple L2-LSPs to the same > > destination > > > > > (e.g. a gateway router) may exist for different > services/flows > > > > > (e.g. video download, VoIP call, etc.). > > > > > > > > > > Regards, > > > > > Richard > > > > > > > > > > > > > > >>-----Original Message----- > > > > >>From: owner-ccamp@ops.ietf.org > > > [mailto:owner-ccamp@ops.ietf.org]On > > > > >>Behalf Of Par Mattsson > > > > >>Sent: 22 July 2005 11:42 > > > > >>To: Loa Andersson > > > > >>Cc: ccamp@ops.ietf.org > > > > >>Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > > >> > > > > >> > > > > >> > > > > >>>Per and Dimitri, > > > > >>> > > > > >>>I would like to come down stronger than that, for me it is > > > > >>>a very strong requirement that the same switch can handle > > > > >>>both VLANs and GMPLs trafic correctly. I can't dsee how that > > > > >>>could be done if using the VLAN tpid to indicate GMPLS > > > > >>>traffic. > > > > >> > > > > >>If you ever want that same switch to handle traffic for a > > > directly > > > > >>connected host (not to uncommen) you would want that to > > > use normal > > > > >>ethernet macaddress switching. So of course you do not want > > > > to have to > > > > >>choose between vlan and gmpls, you would want both at the > > > same time. > > > > >> > > > > >>/per > > > > >> > > > > >> > > > > >> > > > > >>>/Loa > > > > >>> > > > > >>>Par Mattsson wrote: > > > > >>> > > > > >>>>>hi par, one of the possibilities that has been considered > > > > >> > > > > >>to cope with > > > > >> > > > > >>>>>this requirement is to use a dedicated TPID for the > > > > >> > > > > >>Ethernet labeled > > > > >> > > > > >>>>>frames; this would allow differentiated processing with > > > > non-labeled > > > > >>>>>framesthanks. > > > > >>>> > > > > >>>> > > > > >>>>That seems to make more sence. If that frame is to be > > > sized like a > > > > >>>>802.1q > > > > >>>>frame. There is not that much space left to a label. Or is > > > > >> > > > > >>the demand to > > > > >> > > > > >>>>use jumboframes ? > > > > >>>>Has there been any discussion on labelstacking, and > > > > mainly where to > > > > >>>>place > > > > >>>>the information? > > > > >>>> > > > > >>>>Regards. > > > > >>>>Per > > > > >>>> > > > > >>>> > > > > >>>> > > > > >>> > > > > >>> > > > > >>>-- > > > > >>>Loa Andersson > > > > >>> > > > > >>>Principal Networking Architect > > > > >>>Acreo AB phone: +46 8 632 77 14 > > > > >>>Isafjordsgatan 22 mobile: +46 739 81 21 64 > > > > >>>Kista, Sweden email: > > > loa.andersson@acreo.se > > > > >>> loa@pi.se > > > > >>> > > > > >> > > > > >> > > > > >> > > > > > > > > > > > > > > > > > -- > > > > Loa Andersson > > > > > > > > Principal Networking Architect > > > > Acreo AB phone: +46 8 632 77 14 > > > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > > > Kista, Sweden email: > loa.andersson@acreo.se > > > > loa@pi.se > > > > > > > > > > > > > > > > > > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 28 Jul 2005 08:02:32 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Date: Thu, 28 Jul 2005 10:00:25 +0200 Message-ID: <D109C8C97C15294495117745780657AE02EDBAAE@ftrdmel1.rd.francetelecom.fr> Thread-Topic: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Thread-Index: AcWJd4tCECNrgGhpTM+eZPnDSlXsgAJz1TFQ From: "LE ROUX Jean-Louis RD-CORE-LAN" <jeanlouis.leroux@francetelecom.com> To: <ccamp@ops.ietf.org>, "Arthi Ayyangar" <arthi@juniper.net>, <jpv@cisco.com> Hi Arthi, JP Thanks for this new revision that sounds really good I have two commments: -In section 4.1.2 you partially describe bidirectional LSP stitching = procedure. You mention that an Upstream Label MUST NOT be allocated by the = end-to-end LSP on the LSP segment, which is OK. But then how the LSP-Segment Egress will now that the end-to-end LSP is = bidirectional? What about defining a flag in the Attributes Flags TLV of = the LSP_ATTRIBUTE object so as to indicate that the LSP is = bidirectional? Also the selection of the LSP segment in case of bidirectional LSP = should be detailed (e.g. If the end-to-end LSP is bidirecitonal then=20 the LSP-segment MUST be bidirectional. Also shall we allow that two = unidrectional end-to-end LSP use the same bidirectional LSP segment (one = in each direction)? -At the end of section 4.2.5 you mention that LSP-Segment failure or = maintenance SHOULD be treated as a failure event for the end-to-end LSP. = I agree for LSP-Segment failure but not for LSP-Segment maintenance. = LSP-Segment maintenance should be treated as TE-link maintenance for the = end-to-end LSP, and procedures defined in GMPLS graceful TE-link = shutdown draft may be useful (Specific RSVP error code and TE-link = attribute)... Hope this helps, Best Regards JL > -----Message d'origine----- > De : owner-ccamp@ops.ietf.org=20 > [mailto:owner-ccamp@ops.ietf.org] De la part de=20 > Internet-Drafts@ietf.org > Envoy=E9 : vendredi 15 juillet 2005 21:50 > =C0 : i-d-announce@ietf.org > Cc : ccamp@ops.ietf.org > Objet : I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt=20 >=20 > A New Internet-Draft is available from the on-line=20 > Internet-Drafts directories. > This draft is a work item of the Common Control and=20 > Measurement Plane Working Group of the IETF. >=20 > Title : Label Switched Path Stitching with=20 > Generalized MPLS Traffic Engineering > Author(s) : A. Ayyangar, J. Vasseur > Filename : draft-ietf-ccamp-lsp-stitching-01.txt > Pages : 19 > Date : 2005-7-15 > =09 > In certain scenarios, there may be a need to combine together two > different Generalized Multi-Protocol Label Switching (GMPLS) Label > Switched Paths (LSPs) such that in the data plane, a single end-to- > end (e2e) LSP is achieved and all traffic from one LSP is switched > onto the other LSP. We will refer to this as "LSP=20 > stitching". This > document covers cases where: a) the node performing the stitching > does not require configuration of every LSP pair to be stitched > together b) the node performing the stitching is not the egress of > any of the LSPs c) LSP stitching not only results in an end-to-end > LSP in the data plane, but there is also a corresponding end-to-end > LSP (RSVP session) in the control plane. It might be possible to > configure a GMPLS node to switch the traffic from an LSP=20 > for which it > is the egress, to another LSP for which it is the ingress, without > requiring any signaling or routing extensions whatsoever,=20 > completely > transparent to other nodes. This will also result in LSP stitching > in the data plane. However, this document does not cover this > scenario of LSP stitching. >=20 > A URL for this Internet-Draft is: > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitc hing-01.txt >=20 > To remove yourself from the I-D Announcement list, send a=20 > message to i-d-announce-request@ietf.org with the word=20 > unsubscribe in the body of the message. =20 > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > to change your subscription settings. >=20 >=20 > Internet-Drafts are also available by anonymous FTP. Login=20 > with the username "anonymous" and a password of your e-mail=20 > address. After logging in, type "cd internet-drafts" and then > "get draft-ietf-ccamp-lsp-stitching-01.txt". >=20 > A list of Internet-Drafts directories can be found in=20 > http://www.ietf.org/shadow.html or=20 > ftp://ftp.ietf.org/ietf/1shadow-sites.txt >=20 >=20 > Internet-Drafts can also be obtained by e-mail. >=20 > Send a message to: > mailserv@ietf.org. > In the body type: > "FILE /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt". > =09 > NOTE: The mail server at ietf.org can return the document in > MIME-encoded form by using the "mpack" utility. To use this > feature, insert the command "ENCODING mime" before the "FILE" > command. To decode the response(s), you will need "munpack" or > a MIME-compliant mail reader. Different MIME-compliant=20 > mail readers > exhibit different behavior, especially when dealing with > "multipart" MIME messages (i.e. documents which have been split > up into multiple messages), so check your local documentation on > how to manipulate these messages. > =09 > =09 > Below is the data which will enable a MIME compliant mail reader > implementation to automatically retrieve the ASCII version of the > Internet-Draft. >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 22:39:41 +0000 Message-Id: <200507272237.j6RMbMf27528@boreas.isi.edu> To: ietf-announce@ietf.org Subject: RFC 4139 on Requirements for Generalized MPLS (GMPLS) Signaling Usage and Extensions for Automatically Switched Optical Network (ASON) Cc: rfc-editor@rfc-editor.org, ccamp@ops.ietf.org From: rfc-editor@rfc-editor.org Mime-Version: 1.0 Content-Type: Multipart/Mixed; Boundary=NextPart Date: Wed, 27 Jul 2005 15:37:22 -0700 --NextPart A new Request for Comments is now available in online RFC libraries. RFC 4139 Title: Requirements for Generalized MPLS (GMPLS) Signaling Usage and Extensions for Automatically Switched Optical Network (ASON) Author(s): D. Papadimitriou, J. Drake, J. Ash, A. Farrel, L. Ong Status: Informational Date: July 2005 Mailbox: dimitri.papadimitriou@alcatel.be, John.E.Drake2@boeing.com, gash@att.com, adrian@olddog.co.uk, lyong@ciena.com Pages: 16 Characters: 36660 Updates/Obsoletes/SeeAlso: None I-D Tag: draft-ietf-ccamp-gmpls-ason-reqts-07.txt URL: ftp://ftp.rfc-editor.org/in-notes/rfc4139.txt The Generalized Multi-Protocol Label Switching (GMPLS) suite of protocols has been defined to control different switching technologies and different applications. These include support for requesting Time Division Multiplexing (TDM) connections, including Synchronous Optical Network (SONET)/Synchronous Digital Hierarchy (SDH) and Optical Transport Networks (OTNs). This document concentrates on the signaling aspects of the GMPLS suite of protocols. It identifies the features to be covered by the GMPLS signaling protocol to support the capabilities of an Automatically Switched Optical Network (ASON). This document provides a problem statement and additional requirements for the GMPLS signaling protocol to support the ASON functionality. This document is a product of the Common Control and Measurement Plane Working Group of the IETF. This memo provides information for the Internet community. It does not specify an Internet standard of any kind. Distribution of this memo is unlimited. This announcement is sent to the IETF list and the RFC-DIST list. Requests to be added to or deleted from the IETF distribution list should be sent to IETF-REQUEST@IETF.ORG. Requests to be added to or deleted from the RFC-DIST distribution list should be sent to RFC-DIST-REQUEST@RFC-EDITOR.ORG. Details on obtaining RFCs via FTP or EMAIL may be obtained by sending an EMAIL message to rfc-info@RFC-EDITOR.ORG with the message body help: ways_to_get_rfcs. For example: To: rfc-info@RFC-EDITOR.ORG Subject: getting rfcs help: ways_to_get_rfcs Requests for special distribution should be addressed to either the author of the RFC in question, or to RFC-Manager@RFC-EDITOR.ORG. Unless specifically noted otherwise on the RFC itself, all RFCs are for unlimited distribution. Submissions for Requests for Comments should be sent to RFC-EDITOR@RFC-EDITOR.ORG. Please consult RFC 2223, Instructions to RFC Authors, for further information. Joyce K. Reynolds and Sandy Ginoza USC/Information Sciences Institute ... Below is the data which will enable a MIME compliant Mail Reader implementation to automatically retrieve the ASCII version of the RFCs. --NextPart Content-Type: Multipart/Alternative; Boundary="OtherAccess" --OtherAccess Content-Type: Message/External-body; access-type="mail-server"; server="RFC-INFO@RFC-EDITOR.ORG" Content-Type: text/plain Content-ID: <050727153247.RFC@RFC-EDITOR.ORG> RETRIEVE: rfc DOC-ID: rfc4139 --OtherAccess Content-Type: Message/External-body; name="rfc4139.txt"; site="ftp.isi.edu"; access-type="anon-ftp"; directory="in-notes" Content-Type: text/plain Content-ID: <050727153247.RFC@RFC-EDITOR.ORG> --OtherAccess-- --NextPart-- Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 20:56:43 +0000 Message-ID: <00f001c592ed$e72f0510$0601a8c0@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <ccamp@ops.ietf.org> Subject: Incoming correspondence from the OIF Date: Wed, 27 Jul 2005 21:57:45 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi, We have received the following email and attachment from the OIF. I will put these on the web at http://www.olddog.co.uk/incoming.htm Cheers, Adrian ======== Dear Mr. Farrel and Mr. Kampella, On behalf of Jim Jones, OIF Technical Committee Chair, attached please find a liaison letter regarding findings from the OIF Interoperability Demonstration at SUPERCOMM. Warm Regards, Stephanie Stephanie Lazicki / Association Coordinator Optical Internetworking Forum 39355 California Street / Suite 307 / Fremont, CA 94538 USA Phone: +1.510.608.5928 main / +1.510.608.5916 direct Fax: +1.510.608.5917 Email: slazicki@oiforum.com http://www.oiforum.com Managed by Association Management Solutions (AMS); Forum management, Meeting and Event Planning http://www.amsl.com ======== Text of attachment... To: Mr. Adrian Farrel, WG Co-Chair for IETF CCAMP Mr. Kireeti Kompella, WG Co-Chair for IETF CCAMP Copy to Alex Zinin and Bill Fenner, IETF Routing Area Directors From: Jim Jones, OIF Technical Committee Chair July 25, 2005 Dear Adrian and Kireeti, The OIF recently concluded a worldwide interoperability demonstration at Supercomm in June 2005 that successfully focused on dynamic setup of Ethernet Private Line service over an ASON enabled SONET/SDH network. A number of issues were encountered during testing that were felt to be of possible interest to the IETF CCAMP WG, and OIF invites the review and comment of IETF CCAMP on these issues. They include the following: 1. Use of the NCC and RCC fields for STS-3c/VC-4 connections During OIF testing it was noted that some ambiguity exists in the specification of encoding of NCC, RCC and NVC for certain types of connections: NCC and RCC for an STS-3c/VC-4 connection can be set to 0 or to 1 depending on which example of RFC 3946 is followed. Clarification is requested from IETF CCAMP as to which setting is considered correct, or if both settings should be accepted (this procedure was used during testing at Supercomm). 2. Setting of NVC for VCAT connections It was also noted that the setting of NVC may be somewhat ambiguous for the case where diverse connections are used within a single VCAT group. Each individual RSVP session controls a single connection, but the connection is part of a larger VCAT group and carries VCAT encoding of the H4 byte. Clarification is requested from IETF CCAMP and ITU-T Q.14/15 as to the correct setting of NVC for this case (0 or 1?). It should be noted that this case may occur with a VCAT group with only a single initial member, and that the NVC may provide an indication that VCAT encoding of the H4 byte is in use for the connection. 3. Length of the Interface Switching Capability TLV Although the Interface Switching Capability TLV defined by CCAMP for SONET/SDH connections was not used for the testing, it was noted that the text describing the length of the Interface Switching Capability TLV defined in draft-ietf-ccamp-ospf-gmpls-extensions-12.txt may be slightly ambiguous due to the use of padding bytes. RFC 3630 states that "The TLV is padded to four-octet alignment; padding is not included in the length field (so a three octet value would have a length of three, but the total size of the TLV would be eight octets)." Reading of the encoding in draft-ietf-ccamp-ospf-gmpls-extensions-12.txt specifies that the length of the TLV for TDM is 41 bytes plus 3 bytes of padding, and should be given in the length field as 41 bytes rather than 44. OIF requests verification of this interpretation from the experts in IETF CCAMP group. 4. Use of ADMIN_STATUS in an initial PATH message Some implementations sent an ADMIN_STATUS object with no flags set in the initial PATH message, i.e., when no status change was being requested. Although this did not serve any particular function, it was believed that this could be accepted as RFC3473, sect. 7.2 (page 18) states: "The absence of the object is equivalent to receiving an object containing values all set to zero (0)." It was our interpretation based on this text that a node should accept an ADMIN_STATUS object with no flags set in the same way as if the object was missing. Comment on this interpretation is welcome. 5. Handling of multiple received ResvConf Request objects When a connection desires a confirmation that the service (i.e. connection) requested is in place, a RESV_CONF_REQ object is included in the RESV message. As this object is received by the remote end of the reservation, it will send a RESV_CONF message back to the requester. However, it is unclear whether it is necessary to send a RESV_CONF message when the RSVP connection state is refreshed by subsequent RESV. This becomes potentially burdensome, especially when the reservation is being rapidly refreshed. Therefore we ask: should the remote end send a RESV_CONF message for subsequent RESV messages that still include the RESV_CONF_REQ object? Or is it required that the requestor of the reservation remove the RESV_CONF_REQ object to prevent the generation of further RESV_CONF messages? Comment on this issue from IETF CCAMP is requested. 6. Symmetry of Refresh Reduction usage During interop testing, we ran into a conflict caused by varying interpretations of RFC2961, regarding the use of SRefresh messages and the Refresh Reduction capabilities of the two ends of a given link. One interpretation of RFC2961 indicates that setting the Refresh Reduction Capability flag in the RSVP header indicates that that interface shall be capable of receiving messages related to Refresh Reduction - including the SRefresh message. This would be true even if the other end of the link for that interface were NOT indicating Refresh Reduction Capability, since the RFC makes no statement about symmetry in this matter. Another interpretation is that both ends of an interface must indicate Refresh Reduction Capability before either end can use such messages, i.e, use of Refresh Reduction on a link is symmetric. Comment from CCAMP WG on the correct interpretation is requested. 7. Sending of ACKs bundled with the RSVP HELLO During interop testing, it was observed that Message Acks were piggybacked onto RSVP Hello messages, when the receiving end was not using the Hello protocol. In this situation, the incoming Hello's were discarded and the Acks were lost. We believe that Message Acks should only be piggybacked onto mandatory messages, and not on Hello messages because of this problem. Comment on this interpretation is requested. 8. TSPEC format to be used for Ethernet connections For the interoperability demonstration, the INTSERV SENDER_TSPEC format was used. However, it appears to us that the INTSERV SENDER_TSPEC defined in RFC 2210 is not sufficient to support the Ethernet Private Line described in MEF and ITU-T documents as the MEF and ITU-T specifications provide a bandwidth profile description for Ethernet Private Line connections that includes six parameters described below, and the INTSERV format does not have fields that correspond with all six parameters. MEF6 [2] specifies that the Ethernet Bandwidth Profile is specified in terms of the following parameters: CIR, CBS, EIR, EBS, CM, CF. MEF10 [3] and G.8011 Am1 [1] define the parameters, summarized below: CIR (Committed Information Rate): Maximum rate required under normal conditions, service level guaranteed. CBS (Committed Burst Size): Limits the number of bytes for a burst while conforming to CIR. EIR (Excess Information Rate): Maximum rate required under burst conditions, no guarantee of performance. EBS (Excess Burst Size): Limits the number of bytes for a burst while conforming to EIR. CF (Coupling Flag) and CM (Colour Mode): Give a choice for different modes of operation. RFC2210 defines the following SENDER_TSPEC format for Ethernet Traffic Parameters as follows: 31 24 23 16 15 8 7 0 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1 | 0 (a) | reserved | 7 (b) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 2 | 1 (c) |0| reserved | 6 (d) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3 | 127 (e) | 0 (f) | 5 (g) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 4 | Token Bucket Rate [r] (32-bit IEEE floating point number) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 5 | Token Bucket Size [b] (32-bit IEEE floating point number) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 6 | Peak Data Rate [p] (32-bit IEEE floating point number) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 7 | Minimum Policed Unit [m] (32-bit integer) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 8 | Maximum Packet Size [M] (32-bit integer) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ (a) - Message format version number (0) (b) - Overall length (7 words not including header) (c) - Service header, service number 1 (default/global information) (d) - Length of service 1 data, 6 words not including header (e) - Parameter ID, parameter 127 (Token_Bucket_TSpec) (f) - Parameter 127 flags (none set) (g) - Parameter 127 length, 5 words not including header Advice of the IETF CCAMP WG or other experts within IETF is solicited regarding what TSPEC format if any is suitable to be used for setup of Ethernet connections. OIF greatly appreciates the input of the IETF CCAMP WG and other IETF experts in order to reach conclusion on these issues. Best regards, Jim Jones, OIF Technical Committee Chair cc: Trey Malpass, Lyndon Ong References: [1] ITU-T Recommendation G.8011/Y.1307 (2004), Am. 1 (2005), "Ethernet Services Framework" [2] Metro Ethernet Forum, MEF 6 - Ethernet Services Definition, Phase 1, Jun3 2004 [3] Metro Ethernet Forum, MEF.10 - Ethernet Services Attributes Phase 1, Nov. 2004 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 20:37:45 +0000 Message-ID: <00ca01c592eb$409abde0$0601a8c0@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <Greg.Jones@itu.int> Cc: <statements@ietf.org>, "Lam, Hing-Kam \(Kam\)" <hklam@lucent.com>, "Malcolm Betts" <betts01@nortel.com>, "Stephen Trowbridge" <sjtrowbridge@lucent.com>, <ccamp@ops.ietf.org>, "Scott Bradner" <sob@harvard.edu> Subject: CCAMP Response to your liaison on GMPLS/ASON Lexicography Date: Wed, 27 Jul 2005 21:39:05 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit To: ITU-T SG15 Questions 12 and 14 From: IETF CCAMP Working Group Subject: Response to your liaison on GMPLS/ASON Lexicography For: Action Deadline: September 30th 2005 The CCAMP Working Group would like to thank SG15 and especially Questions 12 and 14 for the time and effort they put into providing input and feedback on the GMPLS/ASON lexicography Internet-Draft that is being developed by CCAMP. Much of the material supplied as direct comments or as mark-ups to the Internet-Draft has been incorporated into the latest revision which is available at http://www.ietf.org/internet-drafts/draft-ietf-ccamp-gmpls-ason-lexicography-03.txt and is available for open discussion on the CCAMP mailing list. In your liaison you state some assumptions which we would like to comment on as follows. - The purpose of the document is to enable those familiar with ASON to understand GMPLS terminology in an ASON context Correct. - GMPLS terms are described informally in the document in a way that is generally consistent with GMPLS (existing RFCs and work in progress) While the definition of GMPLS terms would be helpful for understanding GMPLS terms in a GMPLS context, that is not the purpose of this document. - Based on these descriptions, interpretations of these GMPLS terms in ASON terminology are provided It is intended that the interpretation of the GMPLS terms is not based on these descriptions, but based on the full meaning of the GMPLS terms. The descriptions of the GMPLS terms are provided in this Internet-Draft for context and a brief summary. - CCAMP has primarily responsibility for the descriptions of GMPLS terms CCAMP is primarily responsible for the specification of GMPLS protocols and applicability, not just its terms. CCAMP is responsible for all of the content of this Internet-Draft, but will depend heavily on SG15 for assistance in providing suitable text for ASON interpretations as set out below. - Q12/15 has responsibility to provide input to the GMPLS-ASON interpretations (based on the descriptions in this draft) We welcome this assumption of responsibility by Q12/15 to provide input to the Internet-Draft. While the Internet-Draft remains under the overall editorial control of the CCAMP working group, we hope to be able to use the text supplied by Q12/15 with only editorial changes. Additionally, we welcome the pledge to continue work on this Internet-Draft through correspondence on the Q12/15 mailing list, and thank Q12/15 for assigning Ben Mack-Crane as the SG15 representative on this matter. Ben has proved very helpful in clarifying and developing comments received from SG15. We have had several email exchanges with Ben about a few of the specific comments made in your review of the 02 revision of the Internet-Draft. Although he has taken these comments to the Q12/15 mailing list no response has been received, perhaps due to pressure of time. We are taking the opportunity to restate several questions here for your convenience and would appreciate your answers as part of the response to this liaison. 1. Definition of a Resource (section 3.2) You have supplied us with the simple text that says... "In ASON applications: A resource refers only to a link connection (e.g. VC-11, VC-4 etc.)." We would like to clarify this because we don't understand how the original text is wrong. It used to say... "ITU-T [should be ASON] terms for resource: - Connection point (cp) in the context of link discovery and resource management (allocation, binding into cross-connects, etc.); - Link connection or trail termination in the context of routing, path computation and signaling." Recall that we are talking about the ASON terms for what GMPLS calls a resource. We think the following... In ASON a link connection is an association of output cp on one side of the link and input cp on the other side of the link. In GMPLS by resource we mean in most of the cases the local end of the resource (which is cp in ASON terminology). This is true in the context of link discovery. More importantly, this is also true in the context of a signaling application. Specifically, when a GMPLS signaling sub-system requests allocation of a resource, the GMPLS Resource Manager allocates only the local end of the resource. This happens on both sides of the link, that is, we allocate input and output resources. This contrasts with ASON, where there is a much more versatile LRM, and the connection manager (signaling application) allocates link connections only on the output side, leaving LRM to coordinate the allocation with the neighbor. However, in GMPLS TE advertisements we account only for outgoing resources, that is, we don't advertise incoming resources on the assumption that they match the outgoing resources for the discovered TE links. Bottom line: we believe that we should retain the previous definitions, and have not made this change in the new revision. We would like to discuss this point further with Q12/15. 2. Link ends and Data links (Section 3.5 - was section 3.4) 2a.Although we understand that ASON does not require the concept of a link end, GMPLS does. Therefore we should provide some form of language to help people do the mapping. The proposed new text from Q12/15 removes all reference to link ends from the ASON section, and we feel this is a mistake and propose the following text: "In ASON, a GMPLS unidirectional data link end is a collection of connection points from the same client layer that are supported by a single trail termination (access point)." This text depends on the fact that a link end is described as a set of resources that transfer traffic to a neighbor. 2b.The new text from Q12/15 gives us... "A GMPLS data link is an ASON link supported by a single server trail." We are not sure that we understand this correctly. It appears to say that a unidirectional data link could be supported by multiple trail terminations (access points). This seems to suggest that a client would have to make a choice about where to send data, which seems to imply it really has to choose between data links! We are confused! We think some of this problem may come from the *need* in GMPLS to identify data link ends. There may be some value in stressing that a GMPLS link end is supported by exactly one trail termination. When several trail terminations are involved, we can still talk about single TE link (as a bundle), but each component link will be a distinct data link. We think that in ASON the SNPP (TE link in GMPLS language) could be supported by multiple terminations. 3. Link interfaces (section 3.6 - was 3.5) The new text from Q12/15 gives us... "A GMPLS interface is "all the stuff between a physical interface and a matrix in an ASON network element." More precisely, it is a trail termination function and adaptation function for which adapted client layer connection points are bound to a matrix." We think "physical interface" may be ambiguous in the multi- layer context. We think the "stuff" should be between a link and a matrix, since we are talking about links in distinct layers. In addition to your feedback on the specific points above, we would welcome your continued comments and review of the Internet-Draft. Regards, Adrian Farrel and Kireeti Kompella IETF CCAMP Working Group Co-chairs Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 20:02:48 +0000 Message-ID: <009901c592e6$45005430$0601a8c0@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: "Ong, Lyndon" <Lyong@Ciena.com> Cc: "ccamp" <ccamp@ops.ietf.org> Subject: Re: VCn-xV LSPs terminology Date: Wed, 27 Jul 2005 20:51:14 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi, Yes, please check everything you can find in the lexicography draft, and comment accordingly. Thanks, Adrian ----- Original Message ----- From: "Ong, Lyndon" <Lyong@Ciena.com> To: "Diego Caviglia" <Diego.Caviglia@marconi.com>; "Wataru Imajuku <imajuku.wataru" <imajuku.wataru@lab.ntt.co.jp> Cc: "Dieter Beller <D.Beller" <D.Beller@alcatel.de>; "ccamp" <ccamp@ops.ietf.org> Sent: Wednesday, July 27, 2005 4:41 PM Subject: RE: VCn-xV LSPs terminology Hi Folks, It would be worthwhile to check this against the lexicography draft. Cheers, Lyndon -----Original Message----- From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Diego Caviglia Sent: Wednesday, July 27, 2005 1:25 AM To: Wataru Imajuku <imajuku.wataru Cc: Dieter Beller <D.Beller; ccamp Subject: Re: VCn-xV LSPs terminology Hi Wataru, Dieter, yes I agree with both of you. Regards Diego Wataru Imajuku <imajuku.wataru@lab.ntt.co.jp> on 27/07/2005 10.00.02 To: Dieter Beller <D.Beller@alcatel.de>, Diego Caviglia <Diego.Caviglia@marconi.com> cc: ccamp@ops.ietf.org Subject: Re: VCn-xV LSPs terminology Hi, Diego I understand your question is whether we shoud call LSP on the basis of signaling session or not. If so, IMHO is Yes. The VC4-7v with single RSVP session is single LSP. On the other hand, VC4-3v + VC4-4v with two RSVP session is two LSPs. However, both of them are single LSP in the client layer as Dieter said, . That is my understanding. But is that right, all ? Best Regards Wataru >Hi Diego, > >Diego Caviglia wrote: > >>Hi all, >> I'd like to have a terminology clarification about virtual >>concatenated LSP. >> >>When, say a VC4-7v, VCAT circuit is set-up with a single signalling session >>RFC3946 (Path, Resv and may be Resv Confirm) is correct to say that >>the VCAT group is made by a single LSP or that is made by 7 LSPs? >> >>IMHO in that case there is a single LSP. >> > >in my understanding, you are describing a multi-layer scenario here >because the VCAT client layer sits on top of the VC4 server layer (see >ITU-T Rec. G.806). > >If you are in the VCAT client layer, there is IMHO a single LSP that is >decomposed into 7 HO VC4 LSPs in the VC4 server layer in your example. > >Now, if multiple (not necessarily all) VC4 LSPs are routed along the >same path you could set up a server layer call containing multiple >VC4 connections using a single RSVP signaling session (as it was done >in the OIF interop demo this year). > >I hope this clarifies the issue. > > >Cheers, >Dieter > >-- > >++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > V >+---------------+ Dieter Beller >| A L C A T E L | Alcatel Optics Group >+---------------+ Optical Networks Division > Dept. US/BA3 > > Alcatel SEL AG > Lorenzstrasse 10 > D-70435 Stuttgart, Germany > > Phone: +49 711 821-43125 > Fax: +49 711 821-47045 > > E-mail: D.Beller@alcatel.de > >++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > --------------------------------- Wataru Imajuku Senior Research Engineer @NTT Network Innovation Labs. TEL +81-46-859-4315 FAX +81-46-859-5541 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 17:52:58 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: quoted-printable Subject: RE: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet Date: Wed, 27 Jul 2005 10:51:38 -0700 Message-ID: <75B2A84D9323BC4CA3977CF378CE75EB59FBFA@xmb-sjc-21e.amer.cisco.com> Thread-Topic: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet Thread-Index: AcWSmNMqdEdyNji/QsiAHoAalcOz4wAOQUog From: "Ali Sajassi \(sajassi\)" <sajassi@cisco.com> To: "Loa Andersson" <loa@pi.se> Cc: <Dimitri.Papadimitriou@alcatel.be>, "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles Juergen" <juergen.heiles@siemens.com>, <per@defero.se>, <ccamp@ops.ietf.org> Loa, =20 > yes - we can make a solution based on designated vlans=20 > "work", but if we do that we will have serious scaling issues I think you might have misinterpreted what I was saying. I was talking about using MPLS encap first and then have MPLS traffic over its designated VLAN(s). All the GMPLS traffic can be carried over a single VLAN or you can have more than one for load balancing purposes. And thus there shouldn't be any scalability issues. =20 >=20 > i don't know if we are ready to start the discussion on=20 > solution, but we seems to do that anyway ;) >=20 Apparently :-) > there is a way for us to get a tpid from IEEE, we could=20 > specify the behavior of switches on that tpid >=20 > my take that we shouldn't do that without involving ieee, but=20 > on the other it is normally quite easy to work with them >=20 If interested one can come and discuss it during next 802.1 group (which is the architecture group for IEEE bridges) meeting. > but again we need to undestand requirements first >=20 Completely agree.=20 -Ali > /Loa >=20 > Ali Sajassi (sajassi) wrote: > > Dimitri, > > =20 > > By "existing Ethernet switches", I mean Ethernet switches based on=20 > > current approved IEEE 802 .1Q/.1D standards and not the=20 > hypothetical=20 > > switches with TBD forwarding characteristics. Anyway, if=20 > the objective=20 > > is to use RSVP-TE with "existing Ethernet switches", you=20 > can do that=20 > > easily by designating a VLAN or (set of VLANs) for RSVP-TE=20 > traffic and=20 > > support both EoMPLS and native bridged traffic on a single trunk.=20 > > There is some overhead with such encapsulation but the=20 > benefit of it=20 > > is that it works now. If we want to do it through some other means=20 > > (such as MAC address translation), then the benefit of such=20 > approach=20 > > is not clear to me; however, the disadvantages are very clear > > :-) Anyway, I'll look forward to more discussions during=20 > and after the=20 > > meeting next week. > > =20 > > Cheers, > > Ali > >=20 > > =20 > -------------------------------------------------------------- > ---------- > > *From:* Dimitri.Papadimitriou@alcatel.be > > [mailto:Dimitri.Papadimitriou@alcatel.be] > > *Sent:* Tuesday, July 26, 2005 4:21 PM > > *To:* Ali Sajassi (sajassi) > > *Cc:* CHO, JAI HYUNG; Heiles Juergen; per@defero.se;=20 > ccamp@ops.ietf.org > > *Subject:* RE: Frameformat in a l2cs gmpls rnvironment=20 > - Issues w/ > > GMPLS Ethernet > >=20 > > ali > >=20 > > to be clear here you are hanging on using the term "existing > > Ethernet switches" while the document that has been at=20 > the source of > > the discussion does not state "GMPLS must be deployable=20 > on existing > > Ethernet switches" this may be the case but there are=20 > limitations > > (and there are perfectly well understood, don't worry)=20 > beside this > > you should also tell us what do you exactly mean with the term > > "existing" - but this is a side issue - > >=20 > > this said, this does NOT mean that a GMPLS Ethernet=20 > switch (either > > existing or even "extended") would need to support a=20 > new Ethernet > > 802.3 frame forwarding paradigm, the difference is=20 > subtle but has to > > be underlined > >=20 > > note: there is a slot of 20 minutes during the next=20 > CCAMP WG meeting > > where this document and all related issues will be=20 > heavily discussed > > ... so i welcome you to join the festivities ;-) > >=20 > > thanks, > >=20 > > - dimitri. > >=20 > > =20 > >=20 > > *"Ali Sajassi \(sajassi\)" <sajassi@cisco.com>* > > Sent by: owner-ccamp@ops.ietf.org > > 07/25/2005 16:13 MST > >=20 > > To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles Juergen" > > <juergen.heiles@siemens.com>, <per@defero.se> > > cc: <ccamp@ops.ietf.org> > > bcc: > > Subject: RE: Frameformat in a l2cs gmpls rnvironment - Issues w/ > > GMPLS Ethernet > >=20 > >=20 > > Jaihyung, > >=20 > > It seems to me that your primary objective is to use=20 > the existing > > Ethernet switches with new control plane (GMPLS) to=20 > setup TE paths for > > different applications. And you mentioned three=20 > approaches for doing > > this in your email. However, all three of them have=20 > issues and are not > > viable. Lets go over them ... > >=20 > > > > > > In below mail, I discussed about three proposed approaches. > > > > > > > > > 1. using VLAN tag as it is for L2 label encoding. > >=20 > > As you know this approach has scalability issue and you=20 > are limited to > > 4K services in the network and if you try to use the=20 > upcoming IEEE > > 802.1ad standard to make VLANs, interface specific,=20 > then you will need > > new switches and not existing Ethernet switches. Even=20 > if you use these > > newer switches, you will still be limited to 4K=20 > services per interface > > and it would defeat the objective of using existing switches. > >=20 > > > > > > 2. defining new protocol ID (TPID) and borrow VLAN=20 > tag format, > > > ?or extended VLAN tag to implement label swapping. > > > > >=20 > > This approach requires new data plane functionality to=20 > be implemented > > (which defeats the objective of using existing switches). > >=20 > > > 3. use lower 3 bytes of MAC address for L2 label encoding. > > > > >=20 > > This approach has several issues: > > 1) how does MAC addresses get distributed among the=20 > bridges ? How does > > this distribution works during a link or node failure ?=20 > How does it > > recover from a link or node failure ? > >=20 > > 2) Will you assign a separate provider MAC for each=20 > customer MAC. If so, > > then how do you address MAC scalability. If not, then=20 > how do you take > > care of 1-to-N mapping between provider MAC address and=20 > customer MAC > > addresses and how do you handle multipoint connections. > >=20 > > 3) how do handle customer FCS retention when doing this=20 > MAC translation > > stuff ?? > >=20 > > > > > > 802.1Q bridge forwards Ethernet frames using two=20 > dataplane tables > > > - MAC forwarding table and VLAN forwarding table. > > > Bridge control protocols, such as GARP, GVRP, GMRP, > > > manipulate one of the two dataplane entities. > >=20 > > This is not correct. Bridges don't have different=20 > databases for MAC > > forwarding and VLAN forwarding. They have filtering=20 > datebase(s) with > > corresponding IDs (FIDs). VLAN IDs get mapped to FIDs=20 > based on operation > > mode (IVL versus SVL). Then once a filtering database=20 > is identified, > > then a MAC address lookup is performed for forwarding.=20 > Bridge control > > protocols use a designated MAC addressed and based on these MAC > > addressed, the bridge knows how to process the frame. > >=20 > > > > > > Similarly, option 1 and 3 are about which one of two > > > dataplane entities > > > GMPLS protocol should control on behalf of bridge=20 > control protocols. > > > The two proposals do not intend to modify bridge behavior > > > seriously, such as MAC learning, aging, filtering. > > > Therefore, the approaches 1 and 3 are in the scope of CCAMP. > >=20 > > If that's the intention, you may find out soon you get=20 > more than what > > you asked for :-) > >=20 > > > > > > (% note however, I would regard implementing label swapping > > > function is acceptable change considering the=20 > Internal Sublayer > > > design of 802.1D bridge.) > > > > >=20 > > The only way of doing this is using VID translation=20 > table specified in > > IEEE 802.1ad which requires new data-plane=20 > functionality and thus > > existing bridges cannot be used !! > >=20 > > Also if going the route of tag swapping, you may=20 > consider using IEEE > > 802.1ah. As the editor of .1ah, I can go over the=20 > details when I get a > > chance. > >=20 > > > > > > However, option 2 is about what TPID code we will > > > choose to implement new switching function. > > > The switch of this implementation should have new hardware > > > design in addition to normal bride hardware. > > > I think such proposal as option 2 is out of CCAMP scope. > >=20 > > Yes. > >=20 > > > > > > Overall objective is improving scalability, traffic=20 > engineering (TE) > > > characteristics of 802.1 bridge that it can be=20 > reliable, manageable > > > enough to replace some core routers. > >=20 > > Easier said than done :-) > >=20 > > > The switching technique you mentioned above, such as > > > Ethernet over MPLS as defined by PWE3 and L2VPN, > > > are all actually router based technology, however this > > > work is based on simple bridge architecture. > > > Cost-effectiveness is the key differentiator. > > > > >=20 > > The devils are in the details and once you worked out=20 > the details, you > > will see the issues. > >=20 > > > > > > In access or enterprise network, capability of=20 > providing end-to-end > > > L2-LSP will enable service providers policing,=20 > measuring, charging > > > application flows using Ethernet based network. This=20 > will eventually > > > improve income structure and introduce new session=20 > based commercial > > > service. I have been discussing this aspect in mail thread of > > > title 'End-to-end L2-LSP'. Please read the mail=20 > thread and comment > > > on the discussion. > > > > > > > > > >You mentioned to use the MAC address or part of it=20 > as a label. > > > >The TRILL working group is defining shortest path routing > > > for Ethernet. > > > >How would this fit together? > > > > > > > > > TRILL employs new encapsulation layer outside of=20 > Ethernet frames. > > > However in my proposal of option-3, I simply use the=20 > Ethernet header > > > to encode GMPLS label. > > > > > > The overall format of rbridge is > > > [Ethernet][r-tag][Ethernet][ data..] > > > It is not LSP based technology but a connectionless=20 > routing bridge. > > > TTL count is very important in rbridge because=20 > frames are routed > > > hop-by-hop and it may loop. > > > > > > Whereas in option-3 above, Ethernet frames that have certain > > > OUI prefix in MAC address are forwarded via a path that > > > GMPLS protocol has configured on MAC forwarding table. > > > > >=20 > > How does the redundancy work in here. Are you using=20 > bridge protocol in > > conjunction with GMPLS protocol. If you do, then there=20 > are bunch of > > issues. If you don't, then what is it used and how does=20 > it interact with > > bridging protocols ?? > >=20 > > -Ali > >=20 > >=20 > > > > > > Thanks > > > > > > Jaihyung > > > > > > > > > > > > -----?? ???----- > > > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com> > > > ?? ??: 2005-07-25 ?? 6:44:34 > > > ?? ??: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles > > > Juergen" <juergen.heiles@siemens.com>, "per@defero.se" > > > <per@defero.se> > > > ??: "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> > > > ??: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > > > > > > > > > > > > > > > > Dear Jaihyung, > > > > > > you mention several options that could be used as a label. > > > Most of them have strong impacts on the data plane and > > > introduce new layer networks. Is this in the scope of CCAMP? > > > What is the goal, to introduce a control plane for a existing > > > L2 technology or introduce a new L2 technology? What would be > > > the benefit of such a new L2 co switching technology compared > > > to Ethernet over MPLS as defined by PWE3 and L2VPN? > > > You mentioned to use the MAC address or part of it as a > > > label. The TRILL working group is defining shortest path > > > routing for Ethernet. How would this fit together? > > > > > > > > > Regards > > > > > > Juergen > > > > > > > > > > > > > > > > > > > -----Original Message----- > > > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > > > > Sent: Saturday, July 23, 2005 6:19 AM > > > > To: Heiles Juergen; per@defero.se > > > > Cc: ccamp@ops.ietf.org > > > > Subject: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > > > > > > > Dear Juergen, Par and all, > > > > > > > > The proposed framework of L2SC was not intended to suggest > > > > any specific solution. It is still an open question. > > > > As Adrian noted, it only aims to present requirements > > > > as in scenarios that may drive people toward next=20 > step in many > > > > viable solutions. > > > > > > > > In fact, there are other options we may also=20 > consider for L2 > > > > label encoding. > > > > IEEE 802.1 Ethernet bridges forward frames based=20 > on 48bits of > > > > MAC address, > > > > and additionally using VLAN tag. > > > > > > > > When the purpose of GMPLS control over Ethernet is not to > > > > create new dataplane > > > > but to utilize IEEE 802.1 bridge architecture, we may > > > consider using > > > > one of the two forwarding methods, MAC forwarding or VLAN > > > forwarding. > > > > (and perhaps any other combination of fields in=20 > MAC, but I'll > > > > not discuss it) > > > > > > > > Use of VLAN ID for label encoding may automate=20 > VLAN configuration > > > > using IP protocols. However, GMPLS protocol cannot use the > > > > field exclusively > > > > because public/private operators already use VLAN=20 > for various > > > > purpose. > > > > There is a potential conflict with existing use of VLAN and > > > > GMPLS use of VLAN label. > > > > Furthermore, scalability of VLAN ID has been=20 > frequently noted > > > > as weakness > > > > because the size of VLAN ID is at most 4096 (12bit). > > > > > > > > The scalability may be improved if the scope of=20 > VLAN label is > > > > confined to > > > > link-local, and some additional swapping function=20 > of VLAN ID > > > > is introduced in > > > > Internal Sublayer [802.1D] of GMPLS implemented switch. > > > > However, this will only be effective when the configuration > > > > of network > > > > is mesh structure that multiple LSP paths exist. If the > > > > configuration of network > > > > is star or tree shape, as normal configuration of=20 > access network, > > > > LSPs concentrate in root node and total number of > > > > LSPs that the network can hold still be limited by=20 > available > > > > label space > > > > at a few root links. > > > > > > > > For these reasons, I do not think any form of VLAN=20 > ID label is > > > > an appropriate choice for layer-2 label encoding. > > > > > > > > There are some other proposals assuming new=20 > assignment of Ethernet > > > > Length/Type value (e.g. new TPID in VLAN tag) and=20 > re-definition of > > > > information fields placed between 802.3 MAC header=20 > and IP packet. > > > > In this case, only the format of VLAN tag or=20 > extended VLAN tag > > > > is borrowed, however, inside the switching hardware, the > > > > filter and relay, > > > > etc. are totally different new dataplane switch. > > > > > > > > I do not see such approach is a GMPLS=20 > implementation for Ethernet > > > > because the core switching technique is not 802.1 Ethernet > > > > bridge at all. > > > > > > > > The other option we may consider is using MAC address filed > > > as below. > > > > > > > > > > > > +-------+-------+-------+-------+-------+-------+ > > > > | 1byte | 2byte | 3byte | 4byte | 5byte | 6byte | > > > > +-------+-------+-------+-------+-------+-------+ > > > > +-----------------------+-----------------------+ > > > > | OUI Prefix (=3DGMPLS) | DA-label (24bit) | > > > > +-----------------------+-----------------------+ > > > > | OUI Prefix (=3DGMPLS) | SA-label (24bit) | > > > > +---------------+-------+-----------------------+ > > > > | Length/Type | > > > > +---------------+ > > > > > > > > > > > > IEEE is designated by the ISO Council to act as the > > > > registration authority > > > > for the higher three-octet of OUI number in the MAC address > > > > to be used by manufacturer. Ethernet manufacturer=20 > may generate > > > > global unique MAC address using the OUI prefix and=20 > address block of > > > > lower three-octet (24bit). Taking advantage of the > > > addressing scheme, > > > > GMPLS may use the lower three-octet exclusively if a unique > > > > OUI number > > > > is reserved for the protocol. With this labeling=20 > scheme, GMPLS will > > > > control MAC forwarding entry, not VLAN table. > > > > > > > > All Ethernet frames controlled by GMPLS will have identical > > > > OUI number > > > > that they can easily be distinguished from other=20 > Ethernet frames. > > > > In principle, the label lookup hardware is=20 > identical to MAC lookup > > > > hardware in this labeling scheme. Therefore GMPLS=20 > implemented > > > > switch may still function as normal Ethernet=20 > bridge to the frames > > > > that OUI number is not GMPLS. This also facilitates GMPLS > > > implemented > > > > switches being deployed in operating Ethernet with minimum > > > > service disruption. > > > > > > > > Note also that above proposed label encoding=20 > method is transparent > > > > to the use of Ethernet Length/Type field. End-user=20 > device may use > > > > the Length/Type field as defined in IEEE 802.3 protocol. > > > > It also allows network operators configure VLAN=20 > for their own > > > > purpose. > > > > When IEEE 802.1p is used in conjunction with=20 > L2-LSP, the priority > > > > field of VLAN tag can also be used for imposing=20 > consistent TE > > > > policy in > > > > legacy switches and GMPLS switches. > > > > > > > > Any way, my conclusion is, there are other options=20 > we may consider, > > > > and this issue is still open to discuss. > > > > > > > > Thanks, > > > > > > > > Sincerely, > > > > > > > > Jaihyung > > > > > > > > > > > > > > > > > > > > > > > > -----?? ???----- > > > > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com> > > > > ?? ??: 2005-07-22 ?? 10:23:43 > > > > ?? ??: "Loa Andersson" <loa@pi.se>,=20 > "richard.spencer@bt.com" > > > > <richard.spencer@bt.com> > > > > ??: "per@defero.se" <per@defero.se>, "ccamp@ops.ietf.org" > > > > <ccamp@ops.ietf.org> > > > > ??: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > > > > > > > > > > > > > > > > > > > > > > > Loa, > > > > > > > > I interpret the ID as a proposal to use GMPLS for=20 > VLAN setup. > > > > So GMPLS and VLAN to not compete. The VLAN is at the data > > > > plane and GMPLS at the control plane. The question=20 > is how and > > > > should different control plane techniques like GMPLS and > > > > (GVRP and STB) work together? > > > > > > > > Regards > > > > > > > > Juergen > > > > > > > > > > > > > -----Original Message----- > > > > > From: owner-ccamp@ops.ietf.org > > > > > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of=20 > Loa Andersson > > > > > Sent: Friday, July 22, 2005 2:53 PM > > > > > To: richard.spencer@bt.com > > > > > Cc: per@defero.se; ccamp@ops.ietf.org > > > > > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > > > > > > > > Richard, > > > > > > > > > > I agree to most of this. Since we have doubts=20 > about the viability > > > > > of taking GMPLS all the way to end-user or=20 > enterprise I think it > > > > > would be good, from a wg perspective, if we=20 > agreed to solve the > > > > > core network problems first. > > > > > > > > > > Do you have any comment on the requirement to=20 > run both VLANs and > > > > > GMPLS on the same switch? > > > > > > > > > > /Loa > > > > > > > > > > richard.spencer@bt.com wrote: > > > > > > Regardless of whether or not a switch is=20 > directly connected > > > > > > to hosts, it must be able to forward packets using the > > > > > connectionless > > > > > > Ethernet data plane. This is due to the fundamental > > > > > requirement that > > > > > > for GMPLS switches to be able to exchange control > > > > information with > > > > > > each other, a data plane for control traffic=20 > must be present. > > > > > > This is akin to using the IP data plane for MPLS > > > > signalling in an > > > > > > IP/MPLS network. An alternative would be to=20 > use a static > > > > reserved > > > > > > L2-LSP for control traffic in the same way=20 > that reserved > > > > VPI/VCIs > > > > > > are used for PNNI signalling in ATM. > > > > > > > > > > > > Regarding connecting hosts to GMPLS switches, I > > > personally don't > > > > > > think extending L2-LSPs into the=20 > enterprise/home network is > > > > > > commercially viable. However, if you do want to use > > > > GMPLS switches > > > > > > in the home/enterprise network and for some reason > > > don't want to > > > > > > extend L2-LSPs down to the host then you will not be > > > performing > > > > > > normal Ethernet Mac address switching anyway.=20 > Instead you will > > > > > > need some kind of policy on the switch that maps > > > connectionless > > > > > > Ethernet packets (e.g. based on MAC src/dest, 802.1p, > > > VLAN) to a > > > > > > L2-LSP. This is because multiple L2-LSPs to the same > > > destination > > > > > > (e.g. a gateway router) may exist for=20 > different services/flows > > > > > > (e.g. video download, VoIP call, etc.). > > > > > > > > > > > > Regards, > > > > > > Richard > > > > > > > > > > > > > > > > > >>-----Original Message----- > > > > > >>From: owner-ccamp@ops.ietf.org > > > > [mailto:owner-ccamp@ops.ietf.org]On > > > > > >>Behalf Of Par Mattsson > > > > > >>Sent: 22 July 2005 11:42 > > > > > >>To: Loa Andersson > > > > > >>Cc: ccamp@ops.ietf.org > > > > > >>Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > > > >> > > > > > >> > > > > > >> > > > > > >>>Per and Dimitri, > > > > > >>> > > > > > >>>I would like to come down stronger than that,=20 > for me it is > > > > > >>>a very strong requirement that the same=20 > switch can handle > > > > > >>>both VLANs and GMPLs trafic correctly. I=20 > can't dsee how that > > > > > >>>could be done if using the VLAN tpid to indicate GMPLS > > > > > >>>traffic. > > > > > >> > > > > > >>If you ever want that same switch to handle=20 > traffic for a > > > > directly > > > > > >>connected host (not to uncommen) you would=20 > want that to > > > > use normal > > > > > >>ethernet macaddress switching. So of course=20 > you do not want > > > > > to have to > > > > > >>choose between vlan and gmpls, you would want=20 > both at the > > > > same time. > > > > > >> > > > > > >>/per > > > > > >> > > > > > >> > > > > > >> > > > > > >>>/Loa > > > > > >>> > > > > > >>>Par Mattsson wrote: > > > > > >>> > > > > > >>>>>hi par, one of the possibilities that has=20 > been considered > > > > > >> > > > > > >>to cope with > > > > > >> > > > > > >>>>>this requirement is to use a dedicated TPID for the > > > > > >> > > > > > >>Ethernet labeled > > > > > >> > > > > > >>>>>frames; this would allow differentiated=20 > processing with > > > > > non-labeled > > > > > >>>>>framesthanks. > > > > > >>>> > > > > > >>>> > > > > > >>>>That seems to make more sence. If that frame is to be > > > > sized like a > > > > > >>>>802.1q > > > > > >>>>frame. There is not that much space left to=20 > a label. Or is > > > > > >> > > > > > >>the demand to > > > > > >> > > > > > >>>>use jumboframes ? > > > > > >>>>Has there been any discussion on labelstacking, and > > > > > mainly where to > > > > > >>>>place > > > > > >>>>the information? > > > > > >>>> > > > > > >>>>Regards. > > > > > >>>>Per > > > > > >>>> > > > > > >>>> > > > > > >>>> > > > > > >>> > > > > > >>> > > > > > >>>-- > > > > > >>>Loa Andersson > > > > > >>> > > > > > >>>Principal Networking Architect > > > > > >>>Acreo AB phone: =20 > +46 8 632 77 14 > > > > > >>>Isafjordsgatan 22 mobile:=20 > +46 739 81 21 64 > > > > > >>>Kista, Sweden email: > > > > loa.andersson@acreo.se > > > > > >>> loa@pi.se > > > > > >>> > > > > > >> > > > > > >> > > > > > >> > > > > > > > > > > > > > > > > > > > > > -- > > > > > Loa Andersson > > > > > > > > > > Principal Networking Architect > > > > > Acreo AB phone: +46 8=20 > 632 77 14 > > > > > Isafjordsgatan 22 mobile: +46=20 > 739 81 21 64 > > > > > Kista, Sweden email: =20 > loa.andersson@acreo.se > > > > > loa@pi.se > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > >=20 >=20 >=20 > -- > Loa Andersson >=20 > Principal Networking Architect > Acreo AB phone: +46 8 632 77 14 > Isafjordsgatan 22 mobile: +46 739 81 21 64 > Kista, Sweden email: loa.andersson@acreo.se > loa@pi.se >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 15:42:30 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: RE: VCn-xV LSPs terminology Date: Wed, 27 Jul 2005 11:41:14 -0400 Message-ID: <0901D1988E815341A0103206A834DA0755559D@mdmxm02.ciena.com> Thread-Topic: VCn-xV LSPs terminology Thread-Index: AcWShORY0rNJvxaSS92zwFwRUIfh1wAPG+Eg From: "Ong, Lyndon" <Lyong@Ciena.com> To: "Diego Caviglia" <Diego.Caviglia@marconi.com>, "Wataru Imajuku <imajuku.wataru" <imajuku.wataru@lab.ntt.co.jp> Cc: "Dieter Beller <D.Beller" <D.Beller@alcatel.de>, "ccamp" <ccamp@ops.ietf.org> Hi Folks, It would be worthwhile to check this against the lexicography draft. Cheers, Lyndon=20 -----Original Message----- From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Diego Caviglia Sent: Wednesday, July 27, 2005 1:25 AM To: Wataru Imajuku <imajuku.wataru Cc: Dieter Beller <D.Beller; ccamp Subject: Re: VCn-xV LSPs terminology Hi Wataru, Dieter, yes I agree with both of you. Regards Diego Wataru Imajuku <imajuku.wataru@lab.ntt.co.jp> on 27/07/2005 10.00.02 To: Dieter Beller <D.Beller@alcatel.de>, Diego Caviglia <Diego.Caviglia@marconi.com> cc: ccamp@ops.ietf.org Subject: Re: VCn-xV LSPs terminology Hi, Diego I understand your question is whether we shoud call LSP on the basis of signaling session or not. If so, IMHO is Yes. The VC4-7v with single RSVP session is single LSP. On the other hand, VC4-3v + VC4-4v with two RSVP session is two LSPs. However, both of them are single LSP in the client layer as Dieter said, . That is my understanding. But is that right, all ? Best Regards Wataru >Hi Diego, > >Diego Caviglia wrote: > >>Hi all, >> I'd like to have a terminology clarification about virtual=20 >>concatenated LSP. >> >>When, say a VC4-7v, VCAT circuit is set-up with a single signalling session >>RFC3946 (Path, Resv and may be Resv Confirm) is correct to say that=20 >>the VCAT group is made by a single LSP or that is made by 7 LSPs? >> >>IMHO in that case there is a single LSP. >> > >in my understanding, you are describing a multi-layer scenario here=20 >because the VCAT client layer sits on top of the VC4 server layer (see=20 >ITU-T Rec. G.806). > >If you are in the VCAT client layer, there is IMHO a single LSP that is >decomposed into 7 HO VC4 LSPs in the VC4 server layer in your example. > >Now, if multiple (not necessarily all) VC4 LSPs are routed along the=20 >same path you could set up a server layer call containing multiple >VC4 connections using a single RSVP signaling session (as it was done=20 >in the OIF interop demo this year). > >I hope this clarifies the issue. > > >Cheers, >Dieter > >-- > >++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > V >+---------------+ Dieter Beller >| A L C A T E L | Alcatel Optics Group >+---------------+ Optical Networks Division > Dept. US/BA3 > > Alcatel SEL AG > Lorenzstrasse 10 > D-70435 Stuttgart, Germany > > Phone: +49 711 821-43125 > Fax: +49 711 821-47045 > > E-mail: D.Beller@alcatel.de > >++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > --------------------------------- Wataru Imajuku Senior Research Engineer @NTT Network Innovation Labs. TEL +81-46-859-4315 FAX +81-46-859-5541 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 14:01:46 +0000 Message-Id: <5.1.1.9.2.20050727223753.056bde80@mailsv4.y.ecl.ntt.co.jp> Date: Wed, 27 Jul 2005 23:01:13 +0900 To: ccamp@ops.ietf.org From: Wataru Imajuku <imajuku.wataru@lab.ntt.co.jp> Subject: Re: LCAS and GMPLS Mime-Version: 1.0 Content-Type: text/plain; charset="ISO-2022-JP"; format=flowed Content-Transfer-Encoding: 7bit Hi, all I forward a nice comment from Mr. Huub van Helvoort. His comments indicates we need liason from ITU-T SG-15 regarding to this issue. Best Regards Wataru >Date: Wed, 27 Jul 2005 15:35:08 +0200 >From: Huub van Helvoort <hhelvoort@chello.nl> >User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) >Gecko/20040804 Netscape/7.2 (ax) >X-Accept-Language: en-us, en >To: "Wataru Imajuku <imajuku.wataru" <imajuku.wataru@lab.ntt.co.jp> >Subject: [Fwd: Re: LCAS and GMPLS] > >Hello Wataru, > >Thank you for your reply. >Would you be so kind to send the attached response also to >the ccamp list? In my name. > >Kindest regards, Huub. >PS My messages are not on the list: > https://psg.com/lists/ccamp/ccamp.2005/maillist.html > >-------- Original Message -------- >Subject: Re: LCAS and GMPLS >Date: Wed, 27 Jul 2005 13:49:19 +0200 >From: Huub van Helvoort <hhelvoort@chello.nl> >To: ccamp@ops.ietf.org >CC: Trevor Wilson <wilsont@nortel.com>, Wataru Imajuku ><imajuku.wataru@lab.ntt.co.jp>, Michael.Dueser@t-systems.com >References: <B03A52266C74544CA7C8CDB507AC5314018DA310@zharhxm1.corp.nortel.com> > >Hello again, > >I agree with what Trevor wrote: > >>VCAT (and LCAS) functionality is required ONLY at the terminating TDM >>NEs of a Virtually Concatenated signal. As long as the 'intermediate >>NEs', through which the individual VC-ns are transported, have the >>capacity to carry those VC-ns then the SONET/SDH network will provide >>the necessary connection. For any given VCAT connection, the >>intermediate NEs do not need to be VCAT/LCAS capable. > >and: > >>For LCAS to operate it needs at least one member in both directions; but >>essentially the LCAS operation is uni-directional in nature. If >>bandwidth adjustment is required in both directions it is necessary to >>initiate these separately. > >One point for the problem statement could be that in order to >guarantee a hitless removal of one of the members in a VCG >(decrease of bandwidth) there is a mandatory sequence: >First remove member at ingress node, wait for confirm and >only then remove path and member at egress node. > >However, a solution was presented at the last ITU-T SG15/q11 >meeting that removes this requirement. > >Cheers, Huub. --------------------------------- Wataru Imajuku Senior Research Engineer @NTT Network Innovation Labs. TEL +81-46-859-4315 FAX +81-46-859-5541 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 13:18:44 +0000 Message-Id: <5.1.1.9.2.20050727221547.0576bae0@mailsv4.y.ecl.ntt.co.jp> Date: Wed, 27 Jul 2005 22:17:21 +0900 To: Huub van Helvoort <hhelvoort@chello.nl>, Trevor Wilson <wilsont@nortel.com> From: Wataru Imajuku <imajuku.wataru@lab.ntt.co.jp> Subject: Re: VCn-xV LSPs terminology Cc: ccamp <ccamp@ops.ietf.org> Mime-Version: 1.0 Content-Type: text/plain; charset="ISO-2022-JP"; format=flowed Content-Transfer-Encoding: 7bit Hi, Huub Your comment is very valuable. I will try to reflect your comment on -01.txt version. Thank you Best Regards, Wataru >Hello, > >In addition to what Trevor Wilson wrote: > >>Firtsly, lets get the nomenclature correct. A virtually concatenated >>signal is denoted as VC-n-Xv (please note the use of hyphens and >>upper/lower case letters). > >Note that "n" can have the value 4, 3, 2, 12, 11 and for SONET >the nomenclature would be VTn-Xv (n=1.5,6)and STS-n-Xv (n=1,3c). > >Looking at draft-imajuku-ccamp-gmpls-vcat-lagr-req-00 >I find in the second paragraph of the introduction: > "The LCAS achieves hitless bandwidth modification via > signaling in the H4 byte of the SDH/SONET link" >This restricts this document to VC-n-Xv (n=3,4) and STS-n-Xv. > >Changing the text to: > "The LCAS achieves hitless bandwidth modification via > signaling in the VCAT overhead byte of the SDH/SONET link" >will make it more generic (even applicable to OTN). > >H4 bite is also mentioned at the top of page 5. > >Cheers, Huub (van Helvoort) > >-- >================================================================ > http://members.chello.nl/hhelvoort/ >================================================================ >Always remember that you are unique...just like everyone else... --------------------------------- Wataru Imajuku Senior Research Engineer @NTT Network Innovation Labs. TEL +81-46-859-4315 FAX +81-46-859-5541 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 11:42:28 +0000 Message-Id: <B946BD9AB30355458B1B9629C6A601BE0356F661@E8PBE.blf01.telekom.de> From: Michael.Dueser@t-systems.com To: wilsont@nortel.com, imajuku.wataru@lab.ntt.co.jp, ccamp@ops.ietf.org Subject: AW: LCAS and GMPLS Date: Wed, 27 Jul 2005 13:41:41 +0200 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Hi Trevor, you're right! Thanks for the clarification, Michael -----Urspr=FCngliche Nachricht----- Von: Trevor Wilson [mailto:wilsont@nortel.com]=20 Gesendet: Mittwoch, 27. Juli 2005 13:31 An: Wataru Imajuku; Dueser, Michael; ccamp@ops.ietf.org Betreff: RE: LCAS and GMPLS Hi, VCAT (and LCAS) functionality is required ONLY at the terminating TDM = NEs of a Virtually Concatenated signal. As long as the 'intermediate = NEs', through which the individual VC-ns are transported, have the = capacity to carry those VC-ns then the SONET/SDH network will provide = the necessary connection. For any given VCAT connection, the = intermediate NEs do not need to be VCAT/LCAS capable. Therefore I do = not understand why, in the 'problem statement of = 'draft-imajuku-ccamp-gmpls-vcat-lagr-req-00.txt', each TDM node should = advertise its VCAT/LCAS capabilities. When the connection is being = provisioned, it is up to the SDH/SONET NMS to determine whether or not = the network (and it's NEs) are capable of a VCAT/LCAS operation. In the = TDM network, label switching is not performed, except perhaps at the = edge of the network where the client signal is encapsulated into the = transport technology. Therefore I believe that the routes for the = individual VC-ns of a virtually concatenated group (VCG) are of no = consequence to the client signal, and the VCG can be looked upon as a = single LSP between the TDM ingress node and the TDM egress node. Michael: Below you state that LCAS is Bi-directional. For LCAS to operate it needs at least one member in both directions; = but essentially the LCAS operation is uni-directional in nature. If = bandwidth adjustment is required in both directions it is necessary to = initiate these separately. Regards, Trevor Trevor Wilson Nortel Networks Belfast Lab Tel: +44 2890 363701 ESN 751 3701 >This is the Way. This is Nortel. > -----Original Message----- From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Wataru Imajuku Sent: 27 July 2005 05:28 To: Michael.Dueser@t-systems.com; ccamp@ops.ietf.org Subject: Re: LCAS and GMPLS Hi, Mr. Dueser and all Thank you for your comment. As you pointed out, the GFP/LCAS/VCAT functionality is temination capability. But, it can easily migrate with non-GFP/LCAS/VCAT TDM network, The one of ID's requirements includes the extension for the routing protocol in order to advertise such termination capability in each = node. I believe this routing extension helps smooth migration amongst GFP/LCAS/VCAT capable and non capable networks. In the -00.txt version, the ID suggests the extension to Interface Switching Capability Descriptor (ISCD) as a temporal suggestion. Because, I understand that the meaning of interface switching capability=20 descriptor includes both transit capability and termination capability of LSPs. However, the explicit addition of termination capability to the ISCD may=20 be one of choice, considering the more generic node architeture as discussed in Dimitri's = MRN-extensions draft. I hope the discussion in Paris meeting also for above issue . Best Regards, Wataru Hi, >some comments re I-D draft-imajuku-ccamp-gmpls-vcat-lagr-req-00: > >- LCAS was not intended for grouping of VC-X-yvs, but for=20 >synchronization >of the VC-X which is added or dropped. And it's bi-directional. >- Only the terminating nodes need to have the full GFP+LCAS+VCAT functionality. >- You need to be careful when using the verbs may, shoul and must - = are >they conform to IETF convention? > >I believe in general this is an important I-D also wrt to the ITU-T >liaison and for implementation! I believe we can also discuss details=20 >during the IETF next week. > >Michael Dueser --------------------------------- Wataru Imajuku Senior Research Engineer @NTT Network Innovation Labs. TEL +81-46-859-4315 FAX +81-46-859-5541=20 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 11:31:35 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: RE: LCAS and GMPLS Date: Wed, 27 Jul 2005 12:30:45 +0100 Message-ID: <B03A52266C74544CA7C8CDB507AC5314018DA310@zharhxm1.corp.nortel.com> Thread-Topic: LCAS and GMPLS Thread-Index: AcWSlot9u10xs/HcRcma9seXRr/9NQAArRng From: "Trevor Wilson" <wilsont@nortel.com> To: "Wataru Imajuku" <imajuku.wataru@lab.ntt.co.jp>, <Michael.Dueser@t-systems.com>, <ccamp@ops.ietf.org> Hi, VCAT (and LCAS) functionality is required ONLY at the terminating TDM NEs of a Virtually Concatenated signal. As long as the 'intermediate NEs', through which the individual VC-ns are transported, have the capacity to carry those VC-ns then the SONET/SDH network will provide the necessary connection. For any given VCAT connection, the intermediate NEs do not need to be VCAT/LCAS capable. Therefore I do not understand why, in the 'problem statement of 'draft-imajuku-ccamp-gmpls-vcat-lagr-req-00.txt', each TDM node should advertise its VCAT/LCAS capabilities. When the connection is being provisioned, it is up to the SDH/SONET NMS to determine whether or not the network (and it's NEs) are capable of a VCAT/LCAS operation. In the TDM network, label switching is not performed, except perhaps at the edge of the network where the client signal is encapsulated into the transport technology. Therefore I believe that the routes for the individual VC-ns of a virtually concatenated group (VCG) are of no consequence to the client signal, and the VCG can be looked upon as a single LSP between the TDM ingress node and the TDM egress node. Michael: Below you state that LCAS is Bi-directional. For LCAS to operate it needs at least one member in both directions; but essentially the LCAS operation is uni-directional in nature. If bandwidth adjustment is required in both directions it is necessary to initiate these separately. Regards, Trevor Trevor Wilson Nortel Networks Belfast Lab Tel: +44 2890 363701 ESN 751 3701 >This is the Way. This is Nortel.=20 > -----Original Message----- From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Wataru Imajuku Sent: 27 July 2005 05:28 To: Michael.Dueser@t-systems.com; ccamp@ops.ietf.org Subject: Re: LCAS and GMPLS Hi, Mr. Dueser and all Thank you for your comment. As you pointed out, the GFP/LCAS/VCAT functionality is temination capability. But, it can easily migrate with non-GFP/LCAS/VCAT TDM network, The one of ID's requirements includes the extension for the routing protocol in order to advertise such termination capability in each node. I believe this routing extension helps smooth migration amongst GFP/LCAS/VCAT capable and non capable networks. In the -00.txt version, the ID suggests the extension to Interface Switching Capability Descriptor (ISCD) as a temporal suggestion. Because, I understand that the meaning of interface switching capability=20 descriptor includes both transit capability and termination capability of LSPs. However, the explicit addition of termination capability to the ISCD may=20 be one of choice, considering the more generic node architeture as discussed in Dimitri's=20 MRN-extensions draft. I hope the discussion in Paris meeting also for above issue . Best Regards, Wataru Hi, >some comments re I-D draft-imajuku-ccamp-gmpls-vcat-lagr-req-00: > >- LCAS was not intended for grouping of VC-X-yvs, but for=20 >synchronization >of the VC-X which is added or dropped. And it's bi-directional. >- Only the terminating nodes need to have the full GFP+LCAS+VCAT functionality. >- You need to be careful when using the verbs may, shoul and must - are >they conform to IETF convention? > >I believe in general this is an important I-D also wrt to the ITU-T >liaison and for implementation! I believe we can also discuss details=20 >during the IETF next week. > >Michael Dueser --------------------------------- Wataru Imajuku Senior Research Engineer @NTT Network Innovation Labs. TEL +81-46-859-4315 FAX +81-46-859-5541=20 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 10:50:56 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: RE: VCn-xV LSPs terminology Date: Wed, 27 Jul 2005 11:50:04 +0100 Message-ID: <B03A52266C74544CA7C8CDB507AC5314018DA250@zharhxm1.corp.nortel.com> Thread-Topic: VCn-xV LSPs terminology Thread-Index: AcWSlrke3R0gMyHfSzW45vcOAYICPQAAJMiw From: "Trevor Wilson" <wilsont@nortel.com> To: "Diego Caviglia" <Diego.Caviglia@marconi.com>, "Wataru Imajuku <imajuku.wataru" <imajuku.wataru@lab.ntt.co.jp>, "Alan Bailey \(Alan Bailey\)" <IMCEAEX-_O=NORTEL_OU=EuropeM01_cn=Recipients_cn=Exchange_cn=3220803@AmericasM01.nt.com> Cc: "Dieter Beller <D.Beller" <D.Beller@alcatel.de>, "ccamp" <ccamp@ops.ietf.org>, "Ludovico Prattico" <prattico@nortel.com>, "Pete Anslow" <pja@nortel.com>, "Malcolm Betts" <betts01@nortel.com>, "Graham Copley" <copley@nortel.com>, "Astrid Lozano" <astrid@nortel.com>, "David Martin" <dwmartin@nortel.com>, "Michael Mayer" <mgm@nortel.com> Hi, Firtsly, lets get the nomenclature correct. A virtually concatenated signal is denoted as VC-n-Xv (please note the use of hyphens and upper/lower case letters). Also I will use VCG =3D Virtually = Concatenated Group in my reply below. Wataru: You wrote "The VC4-7v with single RSVP session is single LSP. On the other hand, VC4-3v + VC4-4v with two RSVP session is two LSPs " I assume that when you say 'VC-4-3v + VC-4-4v' you mean that the VCG has a total of seven members transported over two routes, one with three members and one with four members. This is and always will be a VC-4-7v. It is not a summation of two VCGs. You cannot think of each route as a separate VCG. The VCAT function at the terminations of the connection will regard all individual VC-4s as members of the VC-4-7v, regardless of which route they were transported on. I hope this helps. Trevor Wilson Nortel Networks Belfast Lab Tel: +44 2890 363701 ESN 751 3701 >This is the Way. This is Nortel.=20 > -----Original Message----- From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Diego Caviglia Sent: 27 July 2005 09:25 To: Wataru Imajuku <imajuku.wataru Cc: Dieter Beller <D.Beller; ccamp Subject: Re: VCn-xV LSPs terminology Hi Wataru, Dieter, yes I agree with both of you. Regards Diego Wataru Imajuku <imajuku.wataru@lab.ntt.co.jp> on 27/07/2005 10.00.02 To: Dieter Beller <D.Beller@alcatel.de>, Diego Caviglia <Diego.Caviglia@marconi.com> cc: ccamp@ops.ietf.org Subject: Re: VCn-xV LSPs terminology Hi, Diego I understand your question is whether we shoud call LSP on the basis of signaling session or not. If so, IMHO is Yes. The VC4-7v with single RSVP session is single LSP. On the other hand, VC4-3v + VC4-4v with two RSVP session is two LSPs. However, both of them are single LSP in the client layer as Dieter said, . That is my understanding. But is that right, all ? Best Regards Wataru >Hi Diego, > >Diego Caviglia wrote: > >>Hi all, >> I'd like to have a terminology clarification about virtual=20 >>concatenated LSP. >> >>When, say a VC4-7v, VCAT circuit is set-up with a single signalling session >>RFC3946 (Path, Resv and may be Resv Confirm) is correct to say that=20 >>the VCAT group is made by a single LSP or that is made by 7 LSPs? >> >>IMHO in that case there is a single LSP. >> > >in my understanding, you are describing a multi-layer scenario here=20 >because the VCAT client layer sits on top of the VC4 server layer (see=20 >ITU-T Rec. G.806). > >If you are in the VCAT client layer, there is IMHO a single LSP that is >decomposed into 7 HO VC4 LSPs in the VC4 server layer in your example. > >Now, if multiple (not necessarily all) VC4 LSPs are routed along the=20 >same path you could set up a server layer call containing multiple VC4=20 >connections using a single RSVP signaling session (as it was done in=20 >the OIF interop demo this year). > >I hope this clarifies the issue. > > >Cheers, >Dieter > >-- > >++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > V >+---------------+ Dieter Beller >| A L C A T E L | Alcatel Optics Group >+---------------+ Optical Networks Division > Dept. US/BA3 > > Alcatel SEL AG > Lorenzstrasse 10 > D-70435 Stuttgart, Germany > > Phone: +49 711 821-43125 > Fax: +49 711 821-47045 > > E-mail: D.Beller@alcatel.de > >++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > --------------------------------- Wataru Imajuku Senior Research Engineer @NTT Network Innovation Labs. TEL +81-46-859-4315 FAX +81-46-859-5541 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 10:49:38 +0000 Message-ID: <42E765A3.6080408@pi.se> Date: Wed, 27 Jul 2005 12:44:51 +0200 From: Loa Andersson <loa@pi.se> Organization: Acreo AB User-Agent: Mozilla Thunderbird 1.0.5 (Windows/20050711) MIME-Version: 1.0 To: "Ali Sajassi (sajassi)" <sajassi@cisco.com> CC: Dimitri.Papadimitriou@alcatel.be, "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, Heiles Juergen <juergen.heiles@siemens.com>, per@defero.se, ccamp@ops.ietf.org Subject: Re: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Ali, yes - we can make a solution based on designated vlans "work", but if we do that we will have serious scaling issues i don't know if we are ready to start the discussion on solution, but we seems to do that anyway ;) there is a way for us to get a tpid from IEEE, we could specify the behavior of switches on that tpid my take that we shouldn't do that without involving ieee, but on the other it is normally quite easy to work with them but again we need to undestand requirements first /Loa Ali Sajassi (sajassi) wrote: > Dimitri, > > By "existing Ethernet switches", I mean Ethernet switches based on > current approved IEEE 802 .1Q/.1D standards and not the hypothetical > switches with TBD forwarding characteristics. Anyway, if the objective > is to use RSVP-TE with "existing Ethernet switches", you can do that > easily by designating a VLAN or (set of VLANs) for RSVP-TE traffic > and support both EoMPLS and native bridged traffic on a > single trunk. There is some overhead with such encapsulation but the > benefit of it is that it works now. If we want to do it through some > other means (such as MAC address translation), then the benefit of such > approach is not clear to me; however, the disadvantages are very clear > :-) Anyway, I'll look forward to more discussions during and after the > meeting next week. > > Cheers, > Ali > > ------------------------------------------------------------------------ > *From:* Dimitri.Papadimitriou@alcatel.be > [mailto:Dimitri.Papadimitriou@alcatel.be] > *Sent:* Tuesday, July 26, 2005 4:21 PM > *To:* Ali Sajassi (sajassi) > *Cc:* CHO, JAI HYUNG; Heiles Juergen; per@defero.se; ccamp@ops.ietf.org > *Subject:* RE: Frameformat in a l2cs gmpls rnvironment - Issues w/ > GMPLS Ethernet > > ali > > to be clear here you are hanging on using the term "existing > Ethernet switches" while the document that has been at the source of > the discussion does not state "GMPLS must be deployable on existing > Ethernet switches" this may be the case but there are limitations > (and there are perfectly well understood, don't worry) beside this > you should also tell us what do you exactly mean with the term > "existing" - but this is a side issue - > > this said, this does NOT mean that a GMPLS Ethernet switch (either > existing or even "extended") would need to support a new Ethernet > 802.3 frame forwarding paradigm, the difference is subtle but has to > be underlined > > note: there is a slot of 20 minutes during the next CCAMP WG meeting > where this document and all related issues will be heavily discussed > ... so i welcome you to join the festivities ;-) > > thanks, > > - dimitri. > > > > *"Ali Sajassi \(sajassi\)" <sajassi@cisco.com>* > Sent by: owner-ccamp@ops.ietf.org > 07/25/2005 16:13 MST > > To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles Juergen" > <juergen.heiles@siemens.com>, <per@defero.se> > cc: <ccamp@ops.ietf.org> > bcc: > Subject: RE: Frameformat in a l2cs gmpls rnvironment - Issues w/ > GMPLS Ethernet > > > Jaihyung, > > It seems to me that your primary objective is to use the existing > Ethernet switches with new control plane (GMPLS) to setup TE paths for > different applications. And you mentioned three approaches for doing > this in your email. However, all three of them have issues and are not > viable. Lets go over them ... > > > > > In below mail, I discussed about three proposed approaches. > > > > > > 1. using VLAN tag as it is for L2 label encoding. > > As you know this approach has scalability issue and you are limited to > 4K services in the network and if you try to use the upcoming IEEE > 802.1ad standard to make VLANs, interface specific, then you will need > new switches and not existing Ethernet switches. Even if you use these > newer switches, you will still be limited to 4K services per interface > and it would defeat the objective of using existing switches. > > > > > 2. defining new protocol ID (TPID) and borrow VLAN tag format, > > ?or extended VLAN tag to implement label swapping. > > > > This approach requires new data plane functionality to be implemented > (which defeats the objective of using existing switches). > > > 3. use lower 3 bytes of MAC address for L2 label encoding. > > > > This approach has several issues: > 1) how does MAC addresses get distributed among the bridges ? How does > this distribution works during a link or node failure ? How does it > recover from a link or node failure ? > > 2) Will you assign a separate provider MAC for each customer MAC. If so, > then how do you address MAC scalability. If not, then how do you take > care of 1-to-N mapping between provider MAC address and customer MAC > addresses and how do you handle multipoint connections. > > 3) how do handle customer FCS retention when doing this MAC translation > stuff ?? > > > > > 802.1Q bridge forwards Ethernet frames using two dataplane tables > > - MAC forwarding table and VLAN forwarding table. > > Bridge control protocols, such as GARP, GVRP, GMRP, > > manipulate one of the two dataplane entities. > > This is not correct. Bridges don't have different databases for MAC > forwarding and VLAN forwarding. They have filtering datebase(s) with > corresponding IDs (FIDs). VLAN IDs get mapped to FIDs based on operation > mode (IVL versus SVL). Then once a filtering database is identified, > then a MAC address lookup is performed for forwarding. Bridge control > protocols use a designated MAC addressed and based on these MAC > addressed, the bridge knows how to process the frame. > > > > > Similarly, option 1 and 3 are about which one of two > > dataplane entities > > GMPLS protocol should control on behalf of bridge control protocols. > > The two proposals do not intend to modify bridge behavior > > seriously, such as MAC learning, aging, filtering. > > Therefore, the approaches 1 and 3 are in the scope of CCAMP. > > If that's the intention, you may find out soon you get more than what > you asked for :-) > > > > > (% note however, I would regard implementing label swapping > > function is acceptable change considering the Internal Sublayer > > design of 802.1D bridge.) > > > > The only way of doing this is using VID translation table specified in > IEEE 802.1ad which requires new data-plane functionality and thus > existing bridges cannot be used !! > > Also if going the route of tag swapping, you may consider using IEEE > 802.1ah. As the editor of .1ah, I can go over the details when I get a > chance. > > > > > However, option 2 is about what TPID code we will > > choose to implement new switching function. > > The switch of this implementation should have new hardware > > design in addition to normal bride hardware. > > I think such proposal as option 2 is out of CCAMP scope. > > Yes. > > > > > Overall objective is improving scalability, traffic engineering (TE) > > characteristics of 802.1 bridge that it can be reliable, manageable > > enough to replace some core routers. > > Easier said than done :-) > > > The switching technique you mentioned above, such as > > Ethernet over MPLS as defined by PWE3 and L2VPN, > > are all actually router based technology, however this > > work is based on simple bridge architecture. > > Cost-effectiveness is the key differentiator. > > > > The devils are in the details and once you worked out the details, you > will see the issues. > > > > > In access or enterprise network, capability of providing end-to-end > > L2-LSP will enable service providers policing, measuring, charging > > application flows using Ethernet based network. This will eventually > > improve income structure and introduce new session based commercial > > service. I have been discussing this aspect in mail thread of > > title 'End-to-end L2-LSP'. Please read the mail thread and comment > > on the discussion. > > > > > > >You mentioned to use the MAC address or part of it as a label. > > >The TRILL working group is defining shortest path routing > > for Ethernet. > > >How would this fit together? > > > > > > TRILL employs new encapsulation layer outside of Ethernet frames. > > However in my proposal of option-3, I simply use the Ethernet header > > to encode GMPLS label. > > > > The overall format of rbridge is > > [Ethernet][r-tag][Ethernet][ data..] > > It is not LSP based technology but a connectionless routing bridge. > > TTL count is very important in rbridge because frames are routed > > hop-by-hop and it may loop. > > > > Whereas in option-3 above, Ethernet frames that have certain > > OUI prefix in MAC address are forwarded via a path that > > GMPLS protocol has configured on MAC forwarding table. > > > > How does the redundancy work in here. Are you using bridge protocol in > conjunction with GMPLS protocol. If you do, then there are bunch of > issues. If you don't, then what is it used and how does it interact with > bridging protocols ?? > > -Ali > > > > > > Thanks > > > > Jaihyung > > > > > > > > -----?? ???----- > > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com> > > ?? ??: 2005-07-25 ?? 6:44:34 > > ?? ??: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles > > Juergen" <juergen.heiles@siemens.com>, "per@defero.se" > > <per@defero.se> > > ??: "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> > > ??: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > > > > > > > > > Dear Jaihyung, > > > > you mention several options that could be used as a label. > > Most of them have strong impacts on the data plane and > > introduce new layer networks. Is this in the scope of CCAMP? > > What is the goal, to introduce a control plane for a existing > > L2 technology or introduce a new L2 technology? What would be > > the benefit of such a new L2 co switching technology compared > > to Ethernet over MPLS as defined by PWE3 and L2VPN? > > You mentioned to use the MAC address or part of it as a > > label. The TRILL working group is defining shortest path > > routing for Ethernet. How would this fit together? > > > > > > Regards > > > > Juergen > > > > > > > > > > > > > -----Original Message----- > > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > > > Sent: Saturday, July 23, 2005 6:19 AM > > > To: Heiles Juergen; per@defero.se > > > Cc: ccamp@ops.ietf.org > > > Subject: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > > > > Dear Juergen, Par and all, > > > > > > The proposed framework of L2SC was not intended to suggest > > > any specific solution. It is still an open question. > > > As Adrian noted, it only aims to present requirements > > > as in scenarios that may drive people toward next step in many > > > viable solutions. > > > > > > In fact, there are other options we may also consider for L2 > > > label encoding. > > > IEEE 802.1 Ethernet bridges forward frames based on 48bits of > > > MAC address, > > > and additionally using VLAN tag. > > > > > > When the purpose of GMPLS control over Ethernet is not to > > > create new dataplane > > > but to utilize IEEE 802.1 bridge architecture, we may > > consider using > > > one of the two forwarding methods, MAC forwarding or VLAN > > forwarding. > > > (and perhaps any other combination of fields in MAC, but I'll > > > not discuss it) > > > > > > Use of VLAN ID for label encoding may automate VLAN configuration > > > using IP protocols. However, GMPLS protocol cannot use the > > > field exclusively > > > because public/private operators already use VLAN for various > > > purpose. > > > There is a potential conflict with existing use of VLAN and > > > GMPLS use of VLAN label. > > > Furthermore, scalability of VLAN ID has been frequently noted > > > as weakness > > > because the size of VLAN ID is at most 4096 (12bit). > > > > > > The scalability may be improved if the scope of VLAN label is > > > confined to > > > link-local, and some additional swapping function of VLAN ID > > > is introduced in > > > Internal Sublayer [802.1D] of GMPLS implemented switch. > > > However, this will only be effective when the configuration > > > of network > > > is mesh structure that multiple LSP paths exist. If the > > > configuration of network > > > is star or tree shape, as normal configuration of access network, > > > LSPs concentrate in root node and total number of > > > LSPs that the network can hold still be limited by available > > > label space > > > at a few root links. > > > > > > For these reasons, I do not think any form of VLAN ID label is > > > an appropriate choice for layer-2 label encoding. > > > > > > There are some other proposals assuming new assignment of Ethernet > > > Length/Type value (e.g. new TPID in VLAN tag) and re-definition of > > > information fields placed between 802.3 MAC header and IP packet. > > > In this case, only the format of VLAN tag or extended VLAN tag > > > is borrowed, however, inside the switching hardware, the > > > filter and relay, > > > etc. are totally different new dataplane switch. > > > > > > I do not see such approach is a GMPLS implementation for Ethernet > > > because the core switching technique is not 802.1 Ethernet > > > bridge at all. > > > > > > The other option we may consider is using MAC address filed > > as below. > > > > > > > > > +-------+-------+-------+-------+-------+-------+ > > > | 1byte | 2byte | 3byte | 4byte | 5byte | 6byte | > > > +-------+-------+-------+-------+-------+-------+ > > > +-----------------------+-----------------------+ > > > | OUI Prefix (=GMPLS) | DA-label (24bit) | > > > +-----------------------+-----------------------+ > > > | OUI Prefix (=GMPLS) | SA-label (24bit) | > > > +---------------+-------+-----------------------+ > > > | Length/Type | > > > +---------------+ > > > > > > > > > IEEE is designated by the ISO Council to act as the > > > registration authority > > > for the higher three-octet of OUI number in the MAC address > > > to be used by manufacturer. Ethernet manufacturer may generate > > > global unique MAC address using the OUI prefix and address block of > > > lower three-octet (24bit). Taking advantage of the > > addressing scheme, > > > GMPLS may use the lower three-octet exclusively if a unique > > > OUI number > > > is reserved for the protocol. With this labeling scheme, GMPLS will > > > control MAC forwarding entry, not VLAN table. > > > > > > All Ethernet frames controlled by GMPLS will have identical > > > OUI number > > > that they can easily be distinguished from other Ethernet frames. > > > In principle, the label lookup hardware is identical to MAC lookup > > > hardware in this labeling scheme. Therefore GMPLS implemented > > > switch may still function as normal Ethernet bridge to the frames > > > that OUI number is not GMPLS. This also facilitates GMPLS > > implemented > > > switches being deployed in operating Ethernet with minimum > > > service disruption. > > > > > > Note also that above proposed label encoding method is transparent > > > to the use of Ethernet Length/Type field. End-user device may use > > > the Length/Type field as defined in IEEE 802.3 protocol. > > > It also allows network operators configure VLAN for their own > > > purpose. > > > When IEEE 802.1p is used in conjunction with L2-LSP, the priority > > > field of VLAN tag can also be used for imposing consistent TE > > > policy in > > > legacy switches and GMPLS switches. > > > > > > Any way, my conclusion is, there are other options we may consider, > > > and this issue is still open to discuss. > > > > > > Thanks, > > > > > > Sincerely, > > > > > > Jaihyung > > > > > > > > > > > > > > > > > > -----?? ???----- > > > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com> > > > ?? ??: 2005-07-22 ?? 10:23:43 > > > ?? ??: "Loa Andersson" <loa@pi.se>, "richard.spencer@bt.com" > > > <richard.spencer@bt.com> > > > ??: "per@defero.se" <per@defero.se>, "ccamp@ops.ietf.org" > > > <ccamp@ops.ietf.org> > > > ??: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > > > > > > > > > > > > > > > > Loa, > > > > > > I interpret the ID as a proposal to use GMPLS for VLAN setup. > > > So GMPLS and VLAN to not compete. The VLAN is at the data > > > plane and GMPLS at the control plane. The question is how and > > > should different control plane techniques like GMPLS and > > > (GVRP and STB) work together? > > > > > > Regards > > > > > > Juergen > > > > > > > > > > -----Original Message----- > > > > From: owner-ccamp@ops.ietf.org > > > > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson > > > > Sent: Friday, July 22, 2005 2:53 PM > > > > To: richard.spencer@bt.com > > > > Cc: per@defero.se; ccamp@ops.ietf.org > > > > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > > > > > > Richard, > > > > > > > > I agree to most of this. Since we have doubts about the viability > > > > of taking GMPLS all the way to end-user or enterprise I think it > > > > would be good, from a wg perspective, if we agreed to solve the > > > > core network problems first. > > > > > > > > Do you have any comment on the requirement to run both VLANs and > > > > GMPLS on the same switch? > > > > > > > > /Loa > > > > > > > > richard.spencer@bt.com wrote: > > > > > Regardless of whether or not a switch is directly connected > > > > > to hosts, it must be able to forward packets using the > > > > connectionless > > > > > Ethernet data plane. This is due to the fundamental > > > > requirement that > > > > > for GMPLS switches to be able to exchange control > > > information with > > > > > each other, a data plane for control traffic must be present. > > > > > This is akin to using the IP data plane for MPLS > > > signalling in an > > > > > IP/MPLS network. An alternative would be to use a static > > > reserved > > > > > L2-LSP for control traffic in the same way that reserved > > > VPI/VCIs > > > > > are used for PNNI signalling in ATM. > > > > > > > > > > Regarding connecting hosts to GMPLS switches, I > > personally don't > > > > > think extending L2-LSPs into the enterprise/home network is > > > > > commercially viable. However, if you do want to use > > > GMPLS switches > > > > > in the home/enterprise network and for some reason > > don't want to > > > > > extend L2-LSPs down to the host then you will not be > > performing > > > > > normal Ethernet Mac address switching anyway. Instead you will > > > > > need some kind of policy on the switch that maps > > connectionless > > > > > Ethernet packets (e.g. based on MAC src/dest, 802.1p, > > VLAN) to a > > > > > L2-LSP. This is because multiple L2-LSPs to the same > > destination > > > > > (e.g. a gateway router) may exist for different services/flows > > > > > (e.g. video download, VoIP call, etc.). > > > > > > > > > > Regards, > > > > > Richard > > > > > > > > > > > > > > >>-----Original Message----- > > > > >>From: owner-ccamp@ops.ietf.org > > > [mailto:owner-ccamp@ops.ietf.org]On > > > > >>Behalf Of Par Mattsson > > > > >>Sent: 22 July 2005 11:42 > > > > >>To: Loa Andersson > > > > >>Cc: ccamp@ops.ietf.org > > > > >>Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > > >> > > > > >> > > > > >> > > > > >>>Per and Dimitri, > > > > >>> > > > > >>>I would like to come down stronger than that, for me it is > > > > >>>a very strong requirement that the same switch can handle > > > > >>>both VLANs and GMPLs trafic correctly. I can't dsee how that > > > > >>>could be done if using the VLAN tpid to indicate GMPLS > > > > >>>traffic. > > > > >> > > > > >>If you ever want that same switch to handle traffic for a > > > directly > > > > >>connected host (not to uncommen) you would want that to > > > use normal > > > > >>ethernet macaddress switching. So of course you do not want > > > > to have to > > > > >>choose between vlan and gmpls, you would want both at the > > > same time. > > > > >> > > > > >>/per > > > > >> > > > > >> > > > > >> > > > > >>>/Loa > > > > >>> > > > > >>>Par Mattsson wrote: > > > > >>> > > > > >>>>>hi par, one of the possibilities that has been considered > > > > >> > > > > >>to cope with > > > > >> > > > > >>>>>this requirement is to use a dedicated TPID for the > > > > >> > > > > >>Ethernet labeled > > > > >> > > > > >>>>>frames; this would allow differentiated processing with > > > > non-labeled > > > > >>>>>framesthanks. > > > > >>>> > > > > >>>> > > > > >>>>That seems to make more sence. If that frame is to be > > > sized like a > > > > >>>>802.1q > > > > >>>>frame. There is not that much space left to a label. Or is > > > > >> > > > > >>the demand to > > > > >> > > > > >>>>use jumboframes ? > > > > >>>>Has there been any discussion on labelstacking, and > > > > mainly where to > > > > >>>>place > > > > >>>>the information? > > > > >>>> > > > > >>>>Regards. > > > > >>>>Per > > > > >>>> > > > > >>>> > > > > >>>> > > > > >>> > > > > >>> > > > > >>>-- > > > > >>>Loa Andersson > > > > >>> > > > > >>>Principal Networking Architect > > > > >>>Acreo AB phone: +46 8 632 77 14 > > > > >>>Isafjordsgatan 22 mobile: +46 739 81 21 64 > > > > >>>Kista, Sweden email: > > > loa.andersson@acreo.se > > > > >>> loa@pi.se > > > > >>> > > > > >> > > > > >> > > > > >> > > > > > > > > > > > > > > > > > -- > > > > Loa Andersson > > > > > > > > Principal Networking Architect > > > > Acreo AB phone: +46 8 632 77 14 > > > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > > > Kista, Sweden email: loa.andersson@acreo.se > > > > loa@pi.se > > > > > > > > > > > > > > > > > > > > > > > -- Loa Andersson Principal Networking Architect Acreo AB phone: +46 8 632 77 14 Isafjordsgatan 22 mobile: +46 739 81 21 64 Kista, Sweden email: loa.andersson@acreo.se loa@pi.se Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 08:25:23 +0000 Sensitivity: Subject: Re: VCn-xV LSPs terminology To: "Wataru Imajuku <imajuku.wataru" <imajuku.wataru@lab.ntt.co.jp> Cc: "Dieter Beller <D.Beller" <D.Beller@alcatel.de>, "ccamp" <ccamp@ops.ietf.org> Message-ID: <OF0001B3D1.ED3B97CC-ONC125704B.002CAE14-C125704B.002E3DDE@uk.marconicomms.com> From: "Diego Caviglia" <Diego.Caviglia@marconi.com> Date: Wed, 27 Jul 2005 10:24:45 +0200 MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii Hi Wataru, Dieter, yes I agree with both of you. Regards Diego Wataru Imajuku <imajuku.wataru@lab.ntt.co.jp> on 27/07/2005 10.00.02 To: Dieter Beller <D.Beller@alcatel.de>, Diego Caviglia <Diego.Caviglia@marconi.com> cc: ccamp@ops.ietf.org Subject: Re: VCn-xV LSPs terminology Hi, Diego I understand your question is whether we shoud call LSP on the basis of signaling session or not. If so, IMHO is Yes. The VC4-7v with single RSVP session is single LSP. On the other hand, VC4-3v + VC4-4v with two RSVP session is two LSPs. However, both of them are single LSP in the client layer as Dieter said, . That is my understanding. But is that right, all ? Best Regards Wataru >Hi Diego, > >Diego Caviglia wrote: > >>Hi all, >> I'd like to have a terminology clarification about virtual >>concatenated LSP. >> >>When, say a VC4-7v, VCAT circuit is set-up with a single signalling session >>RFC3946 (Path, Resv and may be Resv Confirm) is correct to say that the >>VCAT group is made by a single LSP or that is made by 7 LSPs? >> >>IMHO in that case there is a single LSP. >> > >in my understanding, you are describing a multi-layer scenario here >because the VCAT client layer sits on top of the VC4 server layer (see >ITU-T Rec. G.806). > >If you are in the VCAT client layer, there is IMHO a single LSP that >is decomposed into 7 HO VC4 LSPs in the VC4 server layer in your >example. > >Now, if multiple (not necessarily all) VC4 LSPs are routed along the >same path you could set up a server layer call containing multiple >VC4 connections using a single RSVP signaling session (as it was >done in the OIF interop demo this year). > >I hope this clarifies the issue. > > >Cheers, >Dieter > >-- > >++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > V >+---------------+ Dieter Beller >| A L C A T E L | Alcatel Optics Group >+---------------+ Optical Networks Division > Dept. US/BA3 > > Alcatel SEL AG > Lorenzstrasse 10 > D-70435 Stuttgart, Germany > > Phone: +49 711 821-43125 > Fax: +49 711 821-47045 > > E-mail: D.Beller@alcatel.de > >++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > --------------------------------- Wataru Imajuku Senior Research Engineer @NTT Network Innovation Labs. TEL +81-46-859-4315 FAX +81-46-859-5541 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 08:00:30 +0000 Message-Id: <5.1.1.9.2.20050727164908.0568bd28@mailsv4.y.ecl.ntt.co.jp> Date: Wed, 27 Jul 2005 17:00:02 +0900 To: Dieter Beller <D.Beller@alcatel.de>, Diego Caviglia <Diego.Caviglia@marconi.com> From: Wataru Imajuku <imajuku.wataru@lab.ntt.co.jp> Subject: Re: VCn-xV LSPs terminology Cc: ccamp@ops.ietf.org Mime-Version: 1.0 Content-Type: text/plain; charset="ISO-2022-JP"; format=flowed Content-Transfer-Encoding: 7bit Hi, Diego I understand your question is whether we shoud call LSP on the basis of signaling session or not. If so, IMHO is Yes. The VC4-7v with single RSVP session is single LSP. On the other hand, VC4-3v + VC4-4v with two RSVP session is two LSPs. However, both of them are single LSP in the client layer as Dieter said, . That is my understanding. But is that right, all ? Best Regards Wataru >Hi Diego, > >Diego Caviglia wrote: > >>Hi all, >> I'd like to have a terminology clarification about virtual >>concatenated LSP. >> >>When, say a VC4-7v, VCAT circuit is set-up with a single signalling session >>RFC3946 (Path, Resv and may be Resv Confirm) is correct to say that the >>VCAT group is made by a single LSP or that is made by 7 LSPs? >> >>IMHO in that case there is a single LSP. >> > >in my understanding, you are describing a multi-layer scenario here >because the VCAT client layer sits on top of the VC4 server layer (see >ITU-T Rec. G.806). > >If you are in the VCAT client layer, there is IMHO a single LSP that >is decomposed into 7 HO VC4 LSPs in the VC4 server layer in your >example. > >Now, if multiple (not necessarily all) VC4 LSPs are routed along the >same path you could set up a server layer call containing multiple >VC4 connections using a single RSVP signaling session (as it was >done in the OIF interop demo this year). > >I hope this clarifies the issue. > > >Cheers, >Dieter > >-- > >++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > V >+---------------+ Dieter Beller >| A L C A T E L | Alcatel Optics Group >+---------------+ Optical Networks Division > Dept. US/BA3 > > Alcatel SEL AG > Lorenzstrasse 10 > D-70435 Stuttgart, Germany > > Phone: +49 711 821-43125 > Fax: +49 711 821-47045 > > E-mail: D.Beller@alcatel.de > >++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > --------------------------------- Wataru Imajuku Senior Research Engineer @NTT Network Innovation Labs. TEL +81-46-859-4315 FAX +81-46-859-5541 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 07:41:37 +0000 Message-ID: <42E73A68.8050803@alcatel.de> Date: Wed, 27 Jul 2005 09:40:24 +0200 From: Dieter Beller <D.Beller@alcatel.de> Organization: Alcatel SEL AG, OND Germany User-Agent: Mozilla Thunderbird 1.0.6 (Windows/20050716) MIME-Version: 1.0 To: Diego Caviglia <Diego.Caviglia@marconi.com> CC: ccamp@ops.ietf.org Subject: Re: VCn-xV LSPs terminology Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Hi Diego, Diego Caviglia wrote: >Hi all, > I'd like to have a terminology clarification about virtual >concatenated LSP. > >When, say a VC4-7v, VCAT circuit is set-up with a single signalling session >RFC3946 (Path, Resv and may be Resv Confirm) is correct to say that the >VCAT group is made by a single LSP or that is made by 7 LSPs? > >IMHO in that case there is a single LSP. > > in my understanding, you are describing a multi-layer scenario here because the VCAT client layer sits on top of the VC4 server layer (see ITU-T Rec. G.806). If you are in the VCAT client layer, there is IMHO a single LSP that is decomposed into 7 HO VC4 LSPs in the VC4 server layer in your example. Now, if multiple (not necessarily all) VC4 LSPs are routed along the same path you could set up a server layer call containing multiple VC4 connections using a single RSVP signaling session (as it was done in the OIF interop demo this year). I hope this clarifies the issue. Cheers, Dieter -- ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ V +---------------+ Dieter Beller | A L C A T E L | Alcatel Optics Group +---------------+ Optical Networks Division Dept. US/BA3 Alcatel SEL AG Lorenzstrasse 10 D-70435 Stuttgart, Germany Phone: +49 711 821-43125 Fax: +49 711 821-47045 E-mail: D.Beller@alcatel.de ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 06:57:50 +0000 Sensitivity: Subject: VCn-xV LSPs terminology To: ccamp@ops.ietf.org Message-ID: <OF9C119FEC.04F62719-ONC125704B.0024F9CD-C125704B.002635FD@uk.marconicomms.com> From: "Diego Caviglia" <Diego.Caviglia@marconi.com> Date: Wed, 27 Jul 2005 08:57:02 +0200 MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii Hi all, I'd like to have a terminology clarification about virtual concatenated LSP. When, say a VC4-7v, VCAT circuit is set-up with a single signalling session RFC3946 (Path, Resv and may be Resv Confirm) is correct to say that the VCAT group is made by a single LSP or that is made by 7 LSPs? IMHO in that case there is a single LSP. Regards Diego Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 06:26:21 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01C59273.E41176C9" Subject: RE: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet Date: Tue, 26 Jul 2005 23:24:47 -0700 Message-ID: <75B2A84D9323BC4CA3977CF378CE75EB59FB1A@xmb-sjc-21e.amer.cisco.com> Thread-Topic: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet Thread-Index: AcWSOK/zbxRzx3LXQQm+Mbn3ehy40AAOI+2w From: "Ali Sajassi \(sajassi\)" <sajassi@cisco.com> To: <Dimitri.Papadimitriou@alcatel.be> Cc: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles Juergen" <juergen.heiles@siemens.com>, <per@defero.se>, <ccamp@ops.ietf.org> This is a multi-part message in MIME format. ------_=_NextPart_001_01C59273.E41176C9 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: quoted-printable Dimitri, =20 By "existing Ethernet switches", I mean Ethernet switches based on current approved IEEE 802 .1Q/.1D standards and not the hypothetical switches with TBD forwarding characteristics. Anyway, if the objective is to use RSVP-TE with "existing Ethernet switches", you can do that easily by designating a VLAN or (set of VLANs) for RSVP-TE traffic and support both EoMPLS and native bridged traffic on a single trunk. There is some overhead with such encapsulation but the benefit of it is that it works now. If we want to do it through some other means (such as MAC address translation), then the benefit of such approach is not clear to me; however, the disadvantages are very clear :-) Anyway, I'll look forward to more discussions during and after the meeting next week. =20 Cheers, Ali ________________________________ From: Dimitri.Papadimitriou@alcatel.be [mailto:Dimitri.Papadimitriou@alcatel.be]=20 Sent: Tuesday, July 26, 2005 4:21 PM To: Ali Sajassi (sajassi) Cc: CHO, JAI HYUNG; Heiles Juergen; per@defero.se; ccamp@ops.ietf.org Subject: RE: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet =09 =09 ali to be clear here you are hanging on using the term "existing Ethernet switches" while the document that has been at the source of the discussion does not state "GMPLS must be deployable on existing Ethernet switches" this may be the case but there are limitations (and there are perfectly well understood, don't worry) beside this you should also tell us what do you exactly mean with the term "existing" - but this is a side issue - this said, this does NOT mean that a GMPLS Ethernet switch (either existing or even "extended") would need to support a new Ethernet 802.3 frame forwarding paradigm, the difference is subtle but has to be underlined note: there is a slot of 20 minutes during the next CCAMP WG meeting where this document and all related issues will be heavily discussed ... so i welcome you to join the festivities ;-) thanks, - dimitri. =20 =09 "Ali Sajassi \(sajassi\)" <sajassi@cisco.com> Sent by: owner-ccamp@ops.ietf.org 07/25/2005 16:13 MST =09 To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles Juergen" <juergen.heiles@siemens.com>, <per@defero.se> cc: <ccamp@ops.ietf.org> bcc:=20 Subject: RE: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet =09 =09 =09 Jaihyung, =09 It seems to me that your primary objective is to use the existing Ethernet switches with new control plane (GMPLS) to setup TE paths for different applications. And you mentioned three approaches for doing this in your email. However, all three of them have issues and are not viable. Lets go over them ... =09 > > In below mail, I discussed about three proposed approaches. > > > 1. using VLAN tag as it is for L2 label encoding. =09 As you know this approach has scalability issue and you are limited to 4K services in the network and if you try to use the upcoming IEEE 802.1ad standard to make VLANs, interface specific, then you will need new switches and not existing Ethernet switches. Even if you use these newer switches, you will still be limited to 4K services per interface and it would defeat the objective of using existing switches. =09 > > 2. defining new protocol ID (TPID) and borrow VLAN tag format, > ?or extended VLAN tag to implement label swapping. > =09 This approach requires new data plane functionality to be implemented (which defeats the objective of using existing switches). =09 > 3. use lower 3 bytes of MAC address for L2 label encoding. > =09 This approach has several issues: 1) how does MAC addresses get distributed among the bridges ? How does this distribution works during a link or node failure ? How does it recover from a link or node failure ? =09 2) Will you assign a separate provider MAC for each customer MAC. If so, then how do you address MAC scalability. If not, then how do you take care of 1-to-N mapping between provider MAC address and customer MAC addresses and how do you handle multipoint connections. =09 3) how do handle customer FCS retention when doing this MAC translation stuff ?? =09 > > 802.1Q bridge forwards Ethernet frames using two dataplane tables > - MAC forwarding table and VLAN forwarding table. > Bridge control protocols, such as GARP, GVRP, GMRP, > manipulate one of the two dataplane entities. =09 This is not correct. Bridges don't have different databases for MAC forwarding and VLAN forwarding. They have filtering datebase(s) with corresponding IDs (FIDs). VLAN IDs get mapped to FIDs based on operation mode (IVL versus SVL). Then once a filtering database is identified, then a MAC address lookup is performed for forwarding. Bridge control protocols use a designated MAC addressed and based on these MAC addressed, the bridge knows how to process the frame. =09 > > Similarly, option 1 and 3 are about which one of two > dataplane entities > GMPLS protocol should control on behalf of bridge control protocols. > The two proposals do not intend to modify bridge behavior > seriously, such as MAC learning, aging, filtering. > Therefore, the approaches 1 and 3 are in the scope of CCAMP. =09 If that's the intention, you may find out soon you get more than what you asked for :-) =09 > > (% note however, I would regard implementing label swapping > function is acceptable change considering the Internal Sublayer > design of 802.1D bridge.) > =09 The only way of doing this is using VID translation table specified in IEEE 802.1ad which requires new data-plane functionality and thus existing bridges cannot be used !! =09 Also if going the route of tag swapping, you may consider using IEEE 802.1ah. As the editor of .1ah, I can go over the details when I get a chance. =09 > > However, option 2 is about what TPID code we will > choose to implement new switching function. > The switch of this implementation should have new hardware > design in addition to normal bride hardware. > I think such proposal as option 2 is out of CCAMP scope. =09 Yes. =09 > > Overall objective is improving scalability, traffic engineering (TE) > characteristics of 802.1 bridge that it can be reliable, manageable > enough to replace some core routers. =09 Easier said than done :-) =09 > The switching technique you mentioned above, such as > Ethernet over MPLS as defined by PWE3 and L2VPN, > are all actually router based technology, however this > work is based on simple bridge architecture. > Cost-effectiveness is the key differentiator. > =09 The devils are in the details and once you worked out the details, you will see the issues. =09 > > In access or enterprise network, capability of providing end-to-end > L2-LSP will enable service providers policing, measuring, charging > application flows using Ethernet based network. This will eventually > improve income structure and introduce new session based commercial > service. I have been discussing this aspect in mail thread of > title 'End-to-end L2-LSP'. Please read the mail thread and comment > on the discussion. > > > >You mentioned to use the MAC address or part of it as a label. > >The TRILL working group is defining shortest path routing > for Ethernet. > >How would this fit together? > > > TRILL employs new encapsulation layer outside of Ethernet frames. > However in my proposal of option-3, I simply use the Ethernet header > to encode GMPLS label. > > The overall format of rbridge is > [Ethernet][r-tag][Ethernet][ data..] > It is not LSP based technology but a connectionless routing bridge. > TTL count is very important in rbridge because frames are routed > hop-by-hop and it may loop. > > Whereas in option-3 above, Ethernet frames that have certain > OUI prefix in MAC address are forwarded via a path that > GMPLS protocol has configured on MAC forwarding table. > =09 How does the redundancy work in here. Are you using bridge protocol in conjunction with GMPLS protocol. If you do, then there are bunch of issues. If you don't, then what is it used and how does it interact with bridging protocols ?? =09 -Ali =09 =09 > > Thanks > > Jaihyung > > > > -----?? ???----- > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com> > ?? ??: 2005-07-25 ?? 6:44:34 > ?? ??: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles > Juergen" <juergen.heiles@siemens.com>, "per@defero.se" > <per@defero.se> > ??: "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> > ??: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > > Dear Jaihyung, > > you mention several options that could be used as a label. > Most of them have strong impacts on the data plane and > introduce new layer networks. Is this in the scope of CCAMP? > What is the goal, to introduce a control plane for a existing > L2 technology or introduce a new L2 technology? What would be > the benefit of such a new L2 co switching technology compared > to Ethernet over MPLS as defined by PWE3 and L2VPN? > You mentioned to use the MAC address or part of it as a > label. The TRILL working group is defining shortest path > routing for Ethernet. How would this fit together? > > > Regards > > Juergen > > > > > > > -----Original Message----- > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > > Sent: Saturday, July 23, 2005 6:19 AM > > To: Heiles Juergen; per@defero.se > > Cc: ccamp@ops.ietf.org > > Subject: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > Dear Juergen, Par and all, > > > > The proposed framework of L2SC was not intended to suggest > > any specific solution. It is still an open question. > > As Adrian noted, it only aims to present requirements > > as in scenarios that may drive people toward next step in many > > viable solutions. > > > > In fact, there are other options we may also consider for L2 > > label encoding. > > IEEE 802.1 Ethernet bridges forward frames based on 48bits of > > MAC address, > > and additionally using VLAN tag. > > > > When the purpose of GMPLS control over Ethernet is not to > > create new dataplane > > but to utilize IEEE 802.1 bridge architecture, we may > consider using > > one of the two forwarding methods, MAC forwarding or VLAN > forwarding. > > (and perhaps any other combination of fields in MAC, but I'll > > not discuss it) > > > > Use of VLAN ID for label encoding may automate VLAN configuration > > using IP protocols. However, GMPLS protocol cannot use the > > field exclusively > > because public/private operators already use VLAN for various > > purpose. > > There is a potential conflict with existing use of VLAN and > > GMPLS use of VLAN label. > > Furthermore, scalability of VLAN ID has been frequently noted > > as weakness > > because the size of VLAN ID is at most 4096 (12bit). > > > > The scalability may be improved if the scope of VLAN label is > > confined to > > link-local, and some additional swapping function of VLAN ID > > is introduced in > > Internal Sublayer [802.1D] of GMPLS implemented switch. > > However, this will only be effective when the configuration > > of network > > is mesh structure that multiple LSP paths exist. If the > > configuration of network > > is star or tree shape, as normal configuration of access network, > > LSPs concentrate in root node and total number of > > LSPs that the network can hold still be limited by available > > label space > > at a few root links. > > > > For these reasons, I do not think any form of VLAN ID label is > > an appropriate choice for layer-2 label encoding. > > > > There are some other proposals assuming new assignment of Ethernet > > Length/Type value (e.g. new TPID in VLAN tag) and re-definition of > > information fields placed between 802.3 MAC header and IP packet. > > In this case, only the format of VLAN tag or extended VLAN tag > > is borrowed, however, inside the switching hardware, the > > filter and relay, > > etc. are totally different new dataplane switch. > > > > I do not see such approach is a GMPLS implementation for Ethernet > > because the core switching technique is not 802.1 Ethernet > > bridge at all. > > > > The other option we may consider is using MAC address filed > as below. > > > > > > +-------+-------+-------+-------+-------+-------+ > > | 1byte | 2byte | 3byte | 4byte | 5byte | 6byte | > > +-------+-------+-------+-------+-------+-------+ > > +-----------------------+-----------------------+ > > | OUI Prefix (=3DGMPLS) | DA-label (24bit) | > > +-----------------------+-----------------------+ > > | OUI Prefix (=3DGMPLS) | SA-label (24bit) | > > +---------------+-------+-----------------------+ > > | Length/Type | > > +---------------+ > > > > > > IEEE is designated by the ISO Council to act as the > > registration authority > > for the higher three-octet of OUI number in the MAC address > > to be used by manufacturer. Ethernet manufacturer may generate > > global unique MAC address using the OUI prefix and address block of > > lower three-octet (24bit). Taking advantage of the > addressing scheme, > > GMPLS may use the lower three-octet exclusively if a unique > > OUI number > > is reserved for the protocol. With this labeling scheme, GMPLS will > > control MAC forwarding entry, not VLAN table. > > > > All Ethernet frames controlled by GMPLS will have identical > > OUI number > > that they can easily be distinguished from other Ethernet frames. > > In principle, the label lookup hardware is identical to MAC lookup > > hardware in this labeling scheme. Therefore GMPLS implemented > > switch may still function as normal Ethernet bridge to the frames > > that OUI number is not GMPLS. This also facilitates GMPLS > implemented > > switches being deployed in operating Ethernet with minimum > > service disruption. > > > > Note also that above proposed label encoding method is transparent > > to the use of Ethernet Length/Type field. End-user device may use > > the Length/Type field as defined in IEEE 802.3 protocol. > > It also allows network operators configure VLAN for their own > > purpose. > > When IEEE 802.1p is used in conjunction with L2-LSP, the priority > > field of VLAN tag can also be used for imposing consistent TE > > policy in > > legacy switches and GMPLS switches. > > > > Any way, my conclusion is, there are other options we may consider, > > and this issue is still open to discuss. > > > > Thanks, > > > > Sincerely, > > > > Jaihyung > > > > > > > > > > > > -----?? ???----- > > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com> > > ?? ??: 2005-07-22 ?? 10:23:43 > > ?? ??: "Loa Andersson" <loa@pi.se>, "richard.spencer@bt.com" > > <richard.spencer@bt.com> > > ??: "per@defero.se" <per@defero.se>, "ccamp@ops.ietf.org" > > <ccamp@ops.ietf.org> > > ??: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > > > > > > > > > Loa, > > > > I interpret the ID as a proposal to use GMPLS for VLAN setup. > > So GMPLS and VLAN to not compete. The VLAN is at the data > > plane and GMPLS at the control plane. The question is how and > > should different control plane techniques like GMPLS and > > (GVRP and STB) work together? > > > > Regards > > > > Juergen > > > > > > > -----Original Message----- > > > From: owner-ccamp@ops.ietf.org > > > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson > > > Sent: Friday, July 22, 2005 2:53 PM > > > To: richard.spencer@bt.com > > > Cc: per@defero.se; ccamp@ops.ietf.org > > > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > > > > Richard, > > > > > > I agree to most of this. Since we have doubts about the viability > > > of taking GMPLS all the way to end-user or enterprise I think it > > > would be good, from a wg perspective, if we agreed to solve the > > > core network problems first. > > > > > > Do you have any comment on the requirement to run both VLANs and > > > GMPLS on the same switch? > > > > > > /Loa > > > > > > richard.spencer@bt.com wrote: > > > > Regardless of whether or not a switch is directly connected > > > > to hosts, it must be able to forward packets using the > > > connectionless > > > > Ethernet data plane. This is due to the fundamental > > > requirement that > > > > for GMPLS switches to be able to exchange control > > information with > > > > each other, a data plane for control traffic must be present. > > > > This is akin to using the IP data plane for MPLS > > signalling in an > > > > IP/MPLS network. An alternative would be to use a static > > reserved > > > > L2-LSP for control traffic in the same way that reserved > > VPI/VCIs > > > > are used for PNNI signalling in ATM. > > > > > > > > Regarding connecting hosts to GMPLS switches, I > personally don't > > > > think extending L2-LSPs into the enterprise/home network is > > > > commercially viable. However, if you do want to use > > GMPLS switches > > > > in the home/enterprise network and for some reason > don't want to > > > > extend L2-LSPs down to the host then you will not be > performing > > > > normal Ethernet Mac address switching anyway. Instead you will > > > > need some kind of policy on the switch that maps > connectionless > > > > Ethernet packets (e.g. based on MAC src/dest, 802.1p,=20 > VLAN) to a > > > > L2-LSP. This is because multiple L2-LSPs to the same > destination > > > > (e.g. a gateway router) may exist for different services/flows > > > > (e.g. video download, VoIP call, etc.). > > > > > > > > Regards, > > > > Richard > > > > > > > > > > > >>-----Original Message----- > > > >>From: owner-ccamp@ops.ietf.org > > [mailto:owner-ccamp@ops.ietf.org]On > > > >>Behalf Of Par Mattsson > > > >>Sent: 22 July 2005 11:42 > > > >>To: Loa Andersson > > > >>Cc: ccamp@ops.ietf.org > > > >>Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > >> > > > >> > > > >> > > > >>>Per and Dimitri, > > > >>> > > > >>>I would like to come down stronger than that, for me it is > > > >>>a very strong requirement that the same switch can handle > > > >>>both VLANs and GMPLs trafic correctly. I can't dsee how that > > > >>>could be done if using the VLAN tpid to indicate GMPLS > > > >>>traffic. > > > >> > > > >>If you ever want that same switch to handle traffic for a > > directly > > > >>connected host (not to uncommen) you would want that to > > use normal > > > >>ethernet macaddress switching. So of course you do not want > > > to have to > > > >>choose between vlan and gmpls, you would want both at the > > same time. > > > >> > > > >>/per > > > >> > > > >> > > > >> > > > >>>/Loa > > > >>> > > > >>>Par Mattsson wrote: > > > >>> > > > >>>>>hi par, one of the possibilities that has been considered > > > >> > > > >>to cope with > > > >> > > > >>>>>this requirement is to use a dedicated TPID for the > > > >> > > > >>Ethernet labeled > > > >> > > > >>>>>frames; this would allow differentiated processing with > > > non-labeled > > > >>>>>framesthanks. > > > >>>> > > > >>>> > > > >>>>That seems to make more sence. If that frame is to be > > sized like a > > > >>>>802.1q > > > >>>>frame. There is not that much space left to a label. Or is > > > >> > > > >>the demand to > > > >> > > > >>>>use jumboframes ? > > > >>>>Has there been any discussion on labelstacking, and > > > mainly where to > > > >>>>place > > > >>>>the information? > > > >>>> > > > >>>>Regards. > > > >>>>Per > > > >>>> > > > >>>> > > > >>>> > > > >>> > > > >>> > > > >>>-- > > > >>>Loa Andersson > > > >>> > > > >>>Principal Networking Architect > > > >>>Acreo AB phone: +46 8 632 77 14 > > > >>>Isafjordsgatan 22 mobile: +46 739 81 21 64 > > > >>>Kista, Sweden email: > > loa.andersson@acreo.se > > > >>> loa@pi.se > > > >>> > > > >> > > > >> > > > >> > > > > > > > > > > > > > -- > > > Loa Andersson > > > > > > Principal Networking Architect > > > Acreo AB phone: +46 8 632 77 14 > > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > > Kista, Sweden email: loa.andersson@acreo.se > > > loa@pi.se > > > > > > > > > > > > > > ------_=_NextPart_001_01C59273.E41176C9 Content-Type: text/html; charset="US-ASCII" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META http-equiv=3DContent-Type content=3D"text/html; = charset=3Dus-ascii"> <META content=3D"MSHTML 6.00.2800.1505" name=3DGENERATOR></HEAD> <BODY> <DIV dir=3Dltr align=3Dleft><SPAN class=3D696540506-27072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>Dimitri,</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D696540506-27072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2></FONT></SPAN> </DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D696540506-27072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>By "existing Ethernet switches", I mean = Ethernet switches=20 based on current approved IEEE 802 .1Q/.1D standards and = not the=20 hypothetical switches with TBD forwarding characteristics. Anyway, = if the=20 objective is to use RSVP-TE with "existing Ethernet switches", you = can do=20 that easily by designating a VLAN or (set of VLANs) for = RSVP-TE=20 traffic and support both EoMPLS and native bridged traffic on a=20 single trunk. There is some overhead with such = encapsulation but=20 the benefit of it is that it works now. If we want to=20 do it through some other means (such as MAC address = translation), then=20 the benefit of such approach is not clear to me; however, the = disadvantages are=20 very clear :-) Anyway, I'll look forward to more discussions during and = after=20 the meeting next week.</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D696540506-27072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2></FONT></SPAN> </DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D696540506-27072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>Cheers,</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D696540506-27072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>Ali</FONT></SPAN></DIV><BR> <BLOCKQUOTE dir=3Dltr=20 style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #0000ff 2px = solid; MARGIN-RIGHT: 0px"> <DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr align=3Dleft> <HR tabIndex=3D-1> <FONT face=3DTahoma size=3D2><B>From:</B> = Dimitri.Papadimitriou@alcatel.be=20 [mailto:Dimitri.Papadimitriou@alcatel.be] <BR><B>Sent:</B> Tuesday, = July 26,=20 2005 4:21 PM<BR><B>To:</B> Ali Sajassi (sajassi)<BR><B>Cc:</B> CHO, = JAI HYUNG;=20 Heiles Juergen; per@defero.se; ccamp@ops.ietf.org<BR><B>Subject:</B> = RE:=20 Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS=20 Ethernet<BR></FONT><BR></DIV> <DIV></DIV> <P><FONT face=3DMonospace,Courier>ali</FONT></P> <P><FONT face=3DMonospace,Courier>to be clear here you are hanging on = using the=20 term "existing Ethernet switches" while the document that has been at = the=20 source of the discussion does not state "GMPLS must be deployable on = existing=20 Ethernet switches" this may be the case but there are limitations (and = there=20 are perfectly well understood, don't worry) beside this you should = also tell=20 us what do you exactly mean with the term "existing" - but this is a = side=20 issue -</FONT></P> <P><FONT face=3DMonospace,Courier>this said, this does NOT mean that a = GMPLS=20 Ethernet switch (either existing or even "extended") would need to = support a=20 new Ethernet 802.3 frame forwarding paradigm, the difference is subtle = but has=20 to be underlined</FONT></P> <P><FONT face=3DMonospace,Courier>note: there is a slot of 20 minutes = during the=20 next CCAMP WG meeting where this document and all related issues will = be=20 heavily discussed ... so i welcome you to join the festivities = ;-)</FONT></P> <P><FONT face=3DMonospace,Courier>thanks,</FONT></P> <P><FONT face=3DMonospace,Courier>- dimitri.</FONT></P> <P><FONT face=3DMonospace,Courier> </FONT><BR><BR><FONT = size=3D2><B>"Ali=20 Sajassi \(sajassi\)" <sajassi@cisco.com></B></FONT><BR><FONT = size=3D2>Sent=20 by: owner-ccamp@ops.ietf.org</FONT><BR><FONT size=3D2>07/25/2005 16:13 = MST</FONT><BR><BR><FONT size=3D2>To:</FONT> <FONT size=3D2>"CHO, JAI = HYUNG"=20 <jaihyung@etri.re.kr>, "Heiles Juergen"=20 <juergen.heiles@siemens.com>, = <per@defero.se></FONT><BR><FONT=20 size=3D2>cc:</FONT> <FONT = size=3D2><ccamp@ops.ietf.org></FONT><BR><FONT=20 size=3D2>bcc:</FONT> <BR><FONT size=3D2>Subject:</FONT> <FONT = size=3D2>RE:=20 Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS=20 Ethernet</FONT><BR><BR><BR></P> <P><FONT face=3DMonospace,Courier>Jaihyung,<BR></FONT><BR><FONT=20 face=3DMonospace,Courier>It seems to me that your primary objective is = to use=20 the existing<BR>Ethernet switches with new control plane (GMPLS) to = setup TE=20 paths for<BR>different applications. And you mentioned three = approaches for=20 doing<BR>this in your email. However, all three of them have issues = and are=20 not<BR>viable. Lets go over them ...<BR></FONT><BR><FONT=20 face=3DMonospace,Courier>><BR>> In below mail, I discussed about = three=20 proposed approaches.<BR>><BR>><BR>> 1. using VLAN tag as it = is for L2=20 label encoding.<BR></FONT><BR><FONT face=3DMonospace,Courier>As you = know this=20 approach has scalability issue and you are limited to<BR>4K services = in the=20 network and if you try to use the upcoming IEEE<BR>802.1ad standard to = make=20 VLANs, interface specific, then you will need<BR>new switches and not = existing=20 Ethernet switches. Even if you use these<BR>newer switches, you will = still be=20 limited to 4K services per interface<BR>and it would defeat the = objective of=20 using existing switches.<BR></FONT><BR><FONT=20 face=3DMonospace,Courier>><BR>> 2. defining new protocol ID = (TPID) and=20 borrow VLAN tag format,<BR>> ?or extended VLAN tag to implement = label=20 swapping.<BR>><BR></FONT><BR><FONT face=3DMonospace,Courier>This = approach=20 requires new data plane functionality to be implemented<BR>(which = defeats the=20 objective of using existing switches).<BR></FONT><BR><FONT=20 face=3DMonospace,Courier>> 3. use lower 3 bytes of MAC address for = L2 label=20 encoding.<BR>><BR></FONT><BR><FONT face=3DMonospace,Courier>This = approach has=20 several issues:<BR>1) how does MAC addresses get distributed among the = bridges=20 ? How does<BR>this distribution works during a link or node failure ? = How does=20 it<BR>recover from a link or node failure ?<BR></FONT><BR><FONT=20 face=3DMonospace,Courier>2) Will you assign a separate provider MAC = for each=20 customer MAC. If so,<BR>then how do you address MAC scalability. If = not, then=20 how do you take<BR>care of 1-to-N mapping between provider MAC address = and=20 customer MAC<BR>addresses and how do you handle multipoint=20 connections.<BR></FONT><BR><FONT face=3DMonospace,Courier>3) how do = handle=20 customer FCS retention when doing this MAC translation<BR>stuff=20 ??<BR></FONT><BR><FONT face=3DMonospace,Courier>><BR>> 802.1Q = bridge=20 forwards Ethernet frames using two dataplane tables<BR>> - MAC = forwarding=20 table and VLAN forwarding table.<BR>> Bridge control protocols, = such as=20 GARP, GVRP, GMRP,<BR>> manipulate one of the two dataplane=20 entities.<BR></FONT><BR><FONT face=3DMonospace,Courier>This is not = correct.=20 Bridges don't have different databases for MAC<BR>forwarding and VLAN=20 forwarding. They have filtering datebase(s) with<BR>corresponding IDs = (FIDs).=20 VLAN IDs get mapped to FIDs based on operation<BR>mode (IVL versus = SVL). Then=20 once a filtering database is identified,<BR>then a MAC address lookup = is=20 performed for forwarding. Bridge control<BR>protocols use a designated = MAC=20 addressed and based on these MAC<BR>addressed, the bridge knows how to = process=20 the frame.<BR></FONT><BR><FONT face=3DMonospace,Courier>><BR>> = Similarly,=20 option 1 and 3 are about which one of two<BR>> dataplane = entities<BR>>=20 GMPLS protocol should control on behalf of bridge control = protocols.<BR>>=20 The two proposals do not intend to modify bridge behavior<BR>> = seriously,=20 such as MAC learning, aging, filtering.<BR>> Therefore, the = approaches 1=20 and 3 are in the scope of CCAMP.<BR></FONT><BR><FONT = face=3DMonospace,Courier>If=20 that's the intention, you may find out soon you get more than = what<BR>you=20 asked for :-)<BR></FONT><BR><FONT = face=3DMonospace,Courier>><BR>> (% note=20 however, I would regard implementing label swapping<BR>> function = is=20 acceptable change considering the Internal Sublayer<BR>> design of = 802.1D=20 bridge.)<BR>><BR></FONT><BR><FONT face=3DMonospace,Courier>The only = way of=20 doing this is using VID translation table specified in<BR>IEEE 802.1ad = which=20 requires new data-plane functionality and thus<BR>existing bridges = cannot be=20 used !!<BR></FONT><BR><FONT face=3DMonospace,Courier>Also if going the = route of=20 tag swapping, you may consider using IEEE<BR>802.1ah. As the editor of = .1ah, I=20 can go over the details when I get a<BR>chance.<BR></FONT><BR><FONT=20 face=3DMonospace,Courier>><BR>> However, option 2 is about what = TPID code=20 we will<BR>> choose to implement new switching function.<BR>> = The switch=20 of this implementation should have new hardware<BR>> design in = addition to=20 normal bride hardware.<BR>> I think such proposal as option 2 is = out of=20 CCAMP scope.<BR></FONT><BR><FONT=20 face=3DMonospace,Courier>Yes.<BR></FONT><BR><FONT=20 face=3DMonospace,Courier>><BR>> Overall objective is improving=20 scalability, traffic engineering (TE)<BR>> characteristics of 802.1 = bridge=20 that it can be reliable, manageable<BR>> enough to replace some = core=20 routers.<BR></FONT><BR><FONT face=3DMonospace,Courier>Easier said than = done=20 :-)<BR></FONT><BR><FONT face=3DMonospace,Courier>> The switching = technique=20 you mentioned above, such as<BR>> Ethernet over MPLS as defined by = PWE3 and=20 L2VPN,<BR>> are all actually router based technology, however = this<BR>>=20 work is based on simple bridge architecture.<BR>> = Cost-effectiveness is the=20 key differentiator.<BR>><BR></FONT><BR><FONT = face=3DMonospace,Courier>The=20 devils are in the details and once you worked out the details, = you<BR>will see=20 the issues.<BR></FONT><BR><FONT face=3DMonospace,Courier>><BR>> = In access=20 or enterprise network, capability of providing end-to-end<BR>> = L2-LSP will=20 enable service providers policing, measuring, charging<BR>> = application=20 flows using Ethernet based network. This will eventually<BR>> = improve=20 income structure and introduce new session based commercial<BR>> = service. I=20 have been discussing this aspect in mail thread of<BR>> title = 'End-to-end=20 L2-LSP'. Please read the mail thread and comment<BR>> on the=20 discussion.<BR>><BR>><BR>> >You mentioned to use the MAC = address=20 or part of it as a label.<BR>> >The TRILL working group is = defining=20 shortest path routing<BR>> for Ethernet.<BR>> >How would this = fit=20 together?<BR>><BR>><BR>> TRILL employs new encapsulation = layer=20 outside of Ethernet frames.<BR>> However in my proposal of = option-3, I=20 simply use the Ethernet header<BR>> to encode GMPLS = label.<BR>><BR>>=20 The overall format of rbridge is<BR>> [Ethernet][r-tag][Ethernet][=20 data..]<BR>> It is not LSP based technology but a connectionless = routing=20 bridge.<BR>> TTL count is very important in rbridge because frames = are=20 routed<BR>> hop-by-hop and it may loop.<BR>><BR>> Whereas in = option-3=20 above, Ethernet frames that have certain<BR>> OUI prefix in MAC = address are=20 forwarded via a path that<BR>> GMPLS protocol has configured on MAC = forwarding table.<BR>><BR></FONT><BR><FONT = face=3DMonospace,Courier>How does=20 the redundancy work in here. Are you using bridge protocol = in<BR>conjunction=20 with GMPLS protocol. If you do, then there are bunch of<BR>issues. If = you=20 don't, then what is it used and how does it interact with<BR>bridging=20 protocols ??<BR></FONT><BR><FONT=20 face=3DMonospace,Courier>-Ali<BR></FONT><BR><BR><FONT=20 face=3DMonospace,Courier>><BR>> Thanks<BR>><BR>>=20 Jaihyung<BR>><BR>><BR>><BR>> -----?? ???-----<BR>> ?? = ??:=20 "Heiles Juergen" <juergen.heiles@siemens.com><BR>> ?? ??: = 2005-07-25=20 ?? 6:44:34<BR>> ?? ??: "CHO, JAI HYUNG" = <jaihyung@etri.re.kr>,=20 "Heiles<BR>> Juergen" <juergen.heiles@siemens.com>,=20 "per@defero.se"<BR>> <per@defero.se><BR>> ??: = "ccamp@ops.ietf.org"=20 <ccamp@ops.ietf.org><BR>> ??: RE: Frameformat in a l2cs gmpls = rnvironment.<BR>><BR>><BR>><BR>><BR>><BR>><BR>> = Dear=20 Jaihyung,<BR>><BR>> you mention several options that could be = used as a=20 label.<BR>> Most of them have strong impacts on the data plane = and<BR>>=20 introduce new layer networks. Is this in the scope of CCAMP?<BR>> = What is=20 the goal, to introduce a control plane for a existing<BR>> L2 = technology or=20 introduce a new L2 technology? What would be<BR>> the benefit of = such a new=20 L2 co switching technology compared<BR>> to Ethernet over MPLS as = defined=20 by PWE3 and L2VPN?<BR>> You mentioned to use the MAC address or = part of it=20 as a<BR>> label. The TRILL working group is defining shortest = path<BR>>=20 routing for Ethernet. How would this fit = together?<BR>><BR>><BR>>=20 Regards<BR>><BR>>=20 Juergen<BR>><BR>><BR>><BR>><BR>><BR>> > = -----Original=20 Message-----<BR>> > From: CHO, JAI HYUNG=20 [mailto:jaihyung@etri.re.kr]<BR>> > Sent: Saturday, July 23, = 2005 6:19=20 AM<BR>> > To: Heiles Juergen; per@defero.se<BR>> > Cc:=20 ccamp@ops.ietf.org<BR>> > Subject: RE: Frameformat in a l2cs = gmpls=20 rnvironment.<BR>> ><BR>> ><BR>> > Dear Juergen, Par = and=20 all,<BR>> ><BR>> > The proposed framework of L2SC was not = intended=20 to suggest<BR>> > any specific solution. It is still an open=20 question.<BR>> > As Adrian noted, it only aims to present=20 requirements<BR>> > as in scenarios that may drive people toward = next=20 step in many<BR>> > viable solutions.<BR>> ><BR>> > = In fact,=20 there are other options we may also consider for L2<BR>> > label = encoding.<BR>> > IEEE 802.1 Ethernet bridges forward frames = based on=20 48bits of<BR>> > MAC address,<BR>> > and additionally = using VLAN=20 tag.<BR>> ><BR>> > When the purpose of GMPLS control over = Ethernet=20 is not to<BR>> > create new dataplane<BR>> > but to = utilize IEEE=20 802.1 bridge architecture, we may<BR>> consider using<BR>> > = one of=20 the two forwarding methods, MAC forwarding or VLAN<BR>> = forwarding.<BR>>=20 > (and perhaps any other combination of fields in MAC, but = I'll<BR>>=20 > not discuss it)<BR>> ><BR>> > Use of VLAN ID for = label=20 encoding may automate VLAN configuration<BR>> > using IP = protocols.=20 However, GMPLS protocol cannot use the<BR>> > field = exclusively<BR>>=20 > because public/private operators already use VLAN for = various<BR>>=20 > purpose.<BR>> > There is a potential conflict with existing = use of=20 VLAN and<BR>> > GMPLS use of VLAN label.<BR>> > = Furthermore,=20 scalability of VLAN ID has been frequently noted<BR>> > as=20 weakness<BR>> > because the size of VLAN ID is at most 4096=20 (12bit).<BR>> ><BR>> > The scalability may be improved if = the=20 scope of VLAN label is<BR>> > confined to<BR>> > = link-local, and=20 some additional swapping function of VLAN ID<BR>> > is = introduced=20 in<BR>> > Internal Sublayer [802.1D] of GMPLS implemented=20 switch.<BR>> > However, this will only be effective when the=20 configuration<BR>> > of network<BR>> > is mesh structure = that=20 multiple LSP paths exist. If the<BR>> > configuration of = network<BR>>=20 > is star or tree shape, as normal configuration of access = network,<BR>>=20 > LSPs concentrate in root node and total number of<BR>> > = LSPs that=20 the network can hold still be limited by available<BR>> > label=20 space<BR>> > at a few root links.<BR>> ><BR>> > For = these=20 reasons, I do not think any form of VLAN ID label is<BR>> > an=20 appropriate choice for layer-2 label encoding.<BR>> ><BR>> = > There=20 are some other proposals assuming new assignment of Ethernet<BR>> = >=20 Length/Type value (e.g. new TPID in VLAN tag) and re-definition = of<BR>>=20 > information fields placed between 802.3 MAC header and IP = packet.<BR>>=20 > In this case, only the format of VLAN tag or extended VLAN = tag<BR>>=20 > is borrowed, however, inside the switching hardware, the<BR>> = >=20 filter and relay,<BR>> > etc. are totally different new = dataplane=20 switch.<BR>> ><BR>> > I do not see such approach is a = GMPLS=20 implementation for Ethernet<BR>> > because the core switching = technique=20 is not 802.1 Ethernet<BR>> > bridge at all.<BR>> ><BR>> = >=20 The other option we may consider is using MAC address filed<BR>> as = below.<BR>> ><BR>> ><BR>> >=20 +-------+-------+-------+-------+-------+-------+<BR>> > | 1byte = | 2byte=20 | 3byte | 4byte | 5byte | 6byte |<BR>> >=20 +-------+-------+-------+-------+-------+-------+<BR>> >=20 +-----------------------+-----------------------+<BR>> > | = OUI=20 Prefix (=3DGMPLS) | DA-label (24bit) = |<BR>> >=20 +-----------------------+-----------------------+<BR>> > | = OUI=20 Prefix (=3DGMPLS) | SA-label (24bit) = |<BR>> >=20 +---------------+-------+-----------------------+<BR>> > | = Length/Type=20 |<BR>> > +---------------+<BR>> ><BR>> = ><BR>> >=20 IEEE is designated by the ISO Council to act as the<BR>> > = registration=20 authority<BR>> > for the higher three-octet of OUI number in the = MAC=20 address<BR>> > to be used by manufacturer. Ethernet manufacturer = may=20 generate<BR>> > global unique MAC address using the OUI prefix = and=20 address block of<BR>> > lower three-octet (24bit). Taking = advantage of=20 the<BR>> addressing scheme,<BR>> > GMPLS may use the lower=20 three-octet exclusively if a unique<BR>> > OUI number<BR>> = > is=20 reserved for the protocol. With this labeling scheme, GMPLS = will<BR>> >=20 control MAC forwarding entry, not VLAN table.<BR>> ><BR>> = > All=20 Ethernet frames controlled by GMPLS will have identical<BR>> > = OUI=20 number<BR>> > that they can easily be distinguished from other = Ethernet=20 frames.<BR>> > In principle, the label lookup hardware is = identical to=20 MAC lookup<BR>> > hardware in this labeling scheme. Therefore = GMPLS=20 implemented<BR>> > switch may still function as normal Ethernet = bridge=20 to the frames<BR>> > that OUI number is not GMPLS. This also = facilitates=20 GMPLS<BR>> implemented<BR>> > switches being deployed in = operating=20 Ethernet with minimum<BR>> > service disruption.<BR>> = ><BR>>=20 > Note also that above proposed label encoding method is=20 transparent<BR>> > to the use of Ethernet Length/Type field. = End-user=20 device may use<BR>> > the Length/Type field as defined in IEEE = 802.3=20 protocol.<BR>> > It also allows network operators configure VLAN = for=20 their own<BR>> > purpose.<BR>> > When IEEE 802.1p is used = in=20 conjunction with L2-LSP, the priority<BR>> > field of VLAN tag = can also=20 be used for imposing consistent TE<BR>> > policy in<BR>> > = legacy=20 switches and GMPLS switches.<BR>> ><BR>> > Any way, my = conclusion=20 is, there are other options we may consider,<BR>> > and this = issue is=20 still open to discuss.<BR>> ><BR>> > Thanks,<BR>> = ><BR>>=20 > Sincerely,<BR>> ><BR>> > Jaihyung<BR>> = ><BR>>=20 ><BR>> ><BR>> ><BR>> ><BR>> > -----??=20 ???-----<BR>> > ?? ??: "Heiles Juergen"=20 <juergen.heiles@siemens.com><BR>> > ?? ??: 2005-07-22 ??=20 10:23:43<BR>> > ?? ??: "Loa Andersson" <loa@pi.se>,=20 "richard.spencer@bt.com"<BR>> > = <richard.spencer@bt.com><BR>>=20 > ??: "per@defero.se" <per@defero.se>, = "ccamp@ops.ietf.org"<BR>>=20 > <ccamp@ops.ietf.org><BR>> > ??: RE: Frameformat in a = l2cs=20 gmpls rnvironment.<BR>> ><BR>> ><BR>> ><BR>> = ><BR>>=20 ><BR>> ><BR>> > Loa,<BR>> ><BR>> > I = interpret the=20 ID as a proposal to use GMPLS for VLAN setup.<BR>> > So GMPLS = and VLAN=20 to not compete. The VLAN is at the data<BR>> > plane and GMPLS = at the=20 control plane. The question is how and<BR>> > should different = control=20 plane techniques like GMPLS and<BR>> > (GVRP and STB) work=20 together?<BR>> ><BR>> > Regards<BR>> ><BR>> >=20 Juergen<BR>> ><BR>> ><BR>> > > -----Original=20 Message-----<BR>> > > From: owner-ccamp@ops.ietf.org<BR>> = >=20 > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa = Andersson<BR>> >=20 > Sent: Friday, July 22, 2005 2:53 PM<BR>> > > To:=20 richard.spencer@bt.com<BR>> > > Cc: per@defero.se;=20 ccamp@ops.ietf.org<BR>> > > Subject: Re: Frameformat in a = l2cs gmpls=20 rnvironment.<BR>> > ><BR>> > > Richard,<BR>> > = ><BR>> > > I agree to most of this. Since we have doubts = about the=20 viability<BR>> > > of taking GMPLS all the way to end-user or = enterprise I think it<BR>> > > would be good, from a wg = perspective,=20 if we agreed to solve the<BR>> > > core network problems=20 first.<BR>> > ><BR>> > > Do you have any comment on = the=20 requirement to run both VLANs and<BR>> > > GMPLS on the same=20 switch?<BR>> > ><BR>> > > /Loa<BR>> > = ><BR>>=20 > > richard.spencer@bt.com wrote:<BR>> > > > = Regardless of=20 whether or not a switch is directly connected<BR>> > > = > to=20 hosts, it must be able to forward packets using the<BR>> > >=20 connectionless<BR>> > > > Ethernet data plane. This = is due=20 to the fundamental<BR>> > > requirement that<BR>> > = >=20 > for GMPLS switches to be able to exchange control<BR>> = >=20 information with<BR>> > > > each other, a data plane = for=20 control traffic must be present.<BR>> > > > This is = akin to=20 using the IP data plane for MPLS<BR>> > signalling in an<BR>> = >=20 > > IP/MPLS network. An alternative would be to use a=20 static<BR>> > reserved<BR>> > > > L2-LSP for = control=20 traffic in the same way that reserved<BR>> > VPI/VCIs<BR>> = > >=20 > are used for PNNI signalling in ATM.<BR>> > > = ><BR>>=20 > > > Regarding connecting hosts to GMPLS switches, I<BR>> = personally don't<BR>> > > > think extending L2-LSPs = into the=20 enterprise/home network is<BR>> > > > commercially = viable.=20 However, if you do want to use<BR>> > GMPLS switches<BR>> = > >=20 > in the home/enterprise network and for some reason<BR>> = don't=20 want to<BR>> > > > extend L2-LSPs down to the host = then you=20 will not be<BR>> performing<BR>> > > > normal = Ethernet=20 Mac address switching anyway. Instead you will<BR>> > > = >=20 need some kind of policy on the switch that maps<BR>>=20 connectionless</FONT><BR><FONT face=3DMonospace,Courier>> > > = > Ethernet packets (e.g. based on MAC src/dest, 802.1p,=20 </FONT><BR><FONT face=3DMonospace,Courier>> VLAN) to a<BR>> > = >=20 > L2-LSP. This is because multiple L2-LSPs to the = same<BR>>=20 destination<BR>> > > > (e.g. a gateway router) may = exist for=20 different services/flows<BR>> > > > (e.g. video = download,=20 VoIP call, etc.).<BR>> > > ><BR>> > > >=20 Regards,<BR>> > > > Richard<BR>> > > ><BR>> = >=20 > ><BR>> > > >>-----Original Message-----<BR>> = >=20 > >>From: owner-ccamp@ops.ietf.org<BR>> >=20 [mailto:owner-ccamp@ops.ietf.org]On<BR>> > > >>Behalf = Of Par=20 Mattsson<BR>> > > >>Sent: 22 July 2005 11:42<BR>> = > >=20 >>To: Loa Andersson<BR>> > > >>Cc:=20 ccamp@ops.ietf.org<BR>> > > >>Subject: Re: Frameformat = in a=20 l2cs gmpls rnvironment.<BR>> > > >><BR>> > >=20 >><BR>> > > >><BR>> > > >>>Per = and=20 Dimitri,<BR>> > > >>><BR>> > > = >>>I would=20 like to come down stronger than that, for me it is<BR>> > >=20 >>>a very strong requirement that the same switch can = handle<BR>>=20 > > >>>both VLANs and GMPLs trafic correctly. I can't = dsee how=20 that<BR>> > > >>>could be done if using the VLAN = tpid to=20 indicate GMPLS<BR>> > > >>>traffic.<BR>> > = >=20 >><BR>> > > >>If you ever want that same switch = to handle=20 traffic for a<BR>> > directly<BR>> > > = >>connected=20 host (not to uncommen) you would want that to<BR>> > use=20 normal<BR>> > > >>ethernet macaddress switching. So of = course=20 you do not want<BR>> > > to have to<BR>> > > = >>choose=20 between vlan and gmpls, you would want both at the<BR>> > same=20 time.<BR>> > > >><BR>> > > = >>/per<BR>> >=20 > >><BR>> > > >><BR>> > > = >><BR>>=20 > > >>>/Loa<BR>> > > >>><BR>> > = >=20 >>>Par Mattsson wrote:<BR>> > > >>><BR>> = >=20 > >>>>>hi par, one of the possibilities that has = been=20 considered<BR>> > > >><BR>> > > >>to = cope=20 with<BR>> > > >><BR>> > > = >>>>>this=20 requirement is to use a dedicated TPID for the<BR>> > >=20 >><BR>> > > >>Ethernet labeled<BR>> > >=20 >><BR>> > > >>>>>frames; this would = allow=20 differentiated processing with<BR>> > > non-labeled<BR>> = > >=20 >>>>>framesthanks.<BR>> > > = >>>><BR>>=20 > > >>>><BR>> > > >>>>That = seems to=20 make more sence. If that frame is to be<BR>> > sized like = a<BR>> >=20 > >>>>802.1q<BR>> > > >>>>frame. = There is=20 not that much space left to a label. Or is<BR>> > > = >><BR>>=20 > > >>the demand to<BR>> > > >><BR>> = > >=20 >>>>use jumboframes ?<BR>> > > = >>>>Has there=20 been any discussion on labelstacking, and<BR>> > > mainly = where=20 to<BR>> > > >>>>place<BR>> > >=20 >>>>the information?<BR>> > > = >>>><BR>>=20 > > >>>>Regards.<BR>> > >=20 >>>>Per<BR>> > > >>>><BR>> > = >=20 >>>><BR>> > > >>>><BR>> > >=20 >>><BR>> > > >>><BR>> > >=20 >>>--<BR>> > > >>>Loa Andersson<BR>> = > >=20 >>><BR>> > > >>>Principal Networking=20 Architect<BR>> > > >>>Acreo AB = =20 phone:=20 +46 8 632 77 14<BR>> > > >>>Isafjordsgatan 22 = =20 mobile: +46 739 = 81 21=20 64<BR>> > > >>>Kista, Sweden = =20 email:<BR>> >=20 loa.andersson@acreo.se<BR>> > > >>> = =20 = =20 = loa@pi.se<BR>> >=20 > >>><BR>> > > >><BR>> > >=20 >><BR>> > > >><BR>> > > ><BR>> = >=20 ><BR>> > ><BR>> > > --<BR>> > > Loa=20 Andersson<BR>> > ><BR>> > > Principal Networking=20 Architect<BR>> > > Acreo AB = =20 phone: +46 8 = 632 77=20 14<BR>> > > Isafjordsgatan 22 = =20 mobile: +46 739 81 21 64<BR>> > > = Kista,=20 Sweden = =20 email: loa.andersson@acreo.se<BR>> > > = =20 = =20 =20 loa@pi.se<BR>> > ><BR>> ><BR>> ><BR>> = ><BR>>=20 ><BR>><BR>><BR>></FONT></P></BLOCKQUOTE></BODY></HTML> ------_=_NextPart_001_01C59273.E41176C9-- Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 04:29:23 +0000 Message-Id: <5.1.1.9.2.20050727123813.057639a0@mailsv4.y.ecl.ntt.co.jp> Date: Wed, 27 Jul 2005 13:28:20 +0900 To: Michael.Dueser@t-systems.com, ccamp@ops.ietf.org From: Wataru Imajuku <imajuku.wataru@lab.ntt.co.jp> Subject: Re: LCAS and GMPLS Mime-Version: 1.0 Content-Type: text/plain; charset="ISO-2022-JP"; format=flowed Content-Transfer-Encoding: 7bit Hi, Mr. Dueser and all Thank you for your comment. As you pointed out, the GFP/LCAS/VCAT functionality is temination capability. But, it can easily migrate with non-GFP/LCAS/VCAT TDM network, The one of ID's requirements includes the extension for the routing protocol in order to advertise such termination capability in each node. I believe this routing extension helps smooth migration amongst GFP/LCAS/VCAT capable and non capable networks. In the -00.txt version, the ID suggests the extension to Interface Switching Capability Descriptor (ISCD) as a temporal suggestion. Because, I understand that the meaning of interface switching capability descriptor includes both transit capability and termination capability of LSPs. However, the explicit addition of termination capability to the ISCD may be one of choice, considering the more generic node architeture as discussed in Dimitri's MRN-extensions draft. I hope the discussion in Paris meeting also for above issue . Best Regards, Wataru Hi, >some comments re I-D draft-imajuku-ccamp-gmpls-vcat-lagr-req-00: > >- LCAS was not intended for grouping of VC-X-yvs, but for synchronization >of the VC-X which is added or dropped. And it's bi-directional. >- Only the terminating nodes need to have the full GFP+LCAS+VCAT functionality. >- You need to be careful when using the verbs may, shoul and must - are >they conform to IETF convention? > >I believe in general this is an important I-D also wrt to the ITU-T >liaison and for implementation! I believe we can also discuss details >during the IETF next week. > >Michael Dueser --------------------------------- Wataru Imajuku Senior Research Engineer @NTT Network Innovation Labs. TEL +81-46-859-4315 FAX +81-46-859-5541 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 27 Jul 2005 03:02:48 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: quoted-printable Subject: RE: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet Date: Tue, 26 Jul 2005 20:00:53 -0700 Message-ID: <75B2A84D9323BC4CA3977CF378CE75EB59FAC8@xmb-sjc-21e.amer.cisco.com> Thread-Topic: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet Thread-Index: AcWRp4QlwPC80+CYQx66A8lU64uuxQApt/vg From: "Ali Sajassi \(sajassi\)" <sajassi@cisco.com> To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles Juergen" <juergen.heiles@siemens.com>, <per@defero.se> Cc: <ccamp@ops.ietf.org> =20 Jaihyung, Please refer to my comments inline ... [snip] > >=20 > >> 3. use lower 3 bytes of MAC address for L2 label encoding. > >=20 > >This approach has several issues: > >1) how does MAC addresses get distributed among the bridges ?=20 > >How does this distribution works during a link or node failure ?=20 > >How does it recover from a link or node failure ? > >=20 > =20 > =20 > Ali, > The questions are really implementation dependent. > Can we defer this discussion after Paris meeting? > However, I see the solution will not be so different than=20 > what RSVP-TE has been done for other transport network. >=20 These questions are not implementation dependent. IEEE bridges have a well established procedures for handling link and node failures based on RSTP/MSTP and "independent" interaction between these procedures and RSVP-TE is questionable at best. If you think this is not the case, then I encourage you to specify how this interaction works. =20 =20 > =20 > >2) Will you assign a separate provider MAC for each customer MAC.=20 > >If so, then how do you address MAC scalability.=20 > >If not, then how do you take care of 1-to-N mapping between provider=20 > >MAC address and customer MAC addresses and how do you handle=20 > multipoint=20 > >connections. > =20 > =20 > I think you are asking the issue of L2 label stacking or LSP=20 > aggregation, or perhaps some VPN application similar to 802.1ad. > There are some opinions on whether label stacking is=20 > necessary in L2SC domain. This is actually the question=20 > Adrian keep asking us. Since you are using provider MAC address as a label, then by label stacking, I presume you mean encapsulating customers MAC addresses inside the provider MAC. If you want to go to that extend, then why not just simply use IEEE 802.1ah that provide both stacking of MAC addresses as well as it provides you with a 20-bit tag for service instance identification. > I personally think that we can achieve LSP aggregation with=20 > the aid of MPLS shim label, however I think this is too=20 > premature discussion. > All I can say at the moment is, lower 3 octet of MAC address=20 > gives us 16M size of label space and this is enough to=20 > implement link-local, site-local or even some domain-local LSPs. > Between the L2-LSP boundary, it needs further discussion. >=20 It is one thing to have the address space, it is another thing to have proper management and control protocols to manage that space. The big question is what type of control and management protocols you intend to use for that space and how it interacts with the existing IEEE control and management protocols. Also, if it is one-to-one mapping, then we are back to square one regarding MAC address scalability. =20 > =20 > >=20 > >3) how do handle customer FCS retention when doing this MAC=20 > translation=20 > >stuff ?? > >=20 > =20 > =20 > I haven't seen any application that requires retention of=20 > original FCS image. > 802.1D specification recommends recalculation of FCS when=20 > there is any change in MAC. I see no problem if FCS is=20 > recalculated at the label swapping switch. Please refer to all the discussions on the PWE3 mailing list regarding FCS retention. There are certain applications that require it and that's why there is a draft for it. > =20 > =20 > >> 802.1Q bridge forwards Ethernet frames using two dataplane tables > >> - MAC forwarding table and VLAN forwarding table. > >> Bridge control protocols, such as GARP, GVRP, GMRP,=20 > manipulate one of=20 > >> the two dataplane entities. > >=20 > >This is not correct. Bridges don't have different databases for MAC=20 > >forwarding and VLAN forwarding. They have filtering datebase(s) with=20 > >corresponding IDs (FIDs). VLAN IDs get mapped to FIDs based on=20 > >operation mode (IVL versus SVL). Then once a filtering database is=20 > >identified, then a MAC address lookup is performed for forwarding.=20 > >Bridge control protocols use a designated MAC addressed and based on=20 > >these MAC addressed, the bridge knows how to process the frame. > =20 > =20 > oh, come on . > I explained it in conceptual level and you re-described it in=20 > implementation level. The contextual meaning is same. > I believe you don't deny the existence of two types of=20 > filtering databases, for MAC and VID, and the control of GARP=20 > over these databases.=20 > This is unnecessary argument, I think. > =20 No, I didn't describe the implementation but instead I described it as specified in .1D and .1Q. However, I see what you were trying to say and I think it was just a poor choice of words. =20 [snip]=20 > =20 > If we define the scope of L2 label as site-local or=20 > domain-local, we don't need such label swapping operation. > However, if we want link-local label, and find out that it is=20 > very useful > (* actually I do think it'll be very useful *), I think we=20 > need to consult such modification with IEEE. > Technically speaking, I don't think replacing the label value=20 > after normal MAC lookup will serious change hardware architecture. > Also, it doesn't break conformance rule. > Did any 802.1 spec. say that input MAC must be identical to=20 > output MAC? There is no MAC address swapping or translation in .1D, .1Q, or its future standards. If there is MAC address translation, then you can run into quite a few unwanted side affects. [snip] > =20 > =20 > [snip] > >=20 > >How does the redundancy work in here. Are you using bridge=20 > protocol in > >conjunction with GMPLS protocol. If you do, then there are bunch of > >issues.=20 > =20 > =20 > Yes I do.=20 > I intend GMPLS implemented switches being deployed in=20 > operating network > as normal Ethernet switch that perform bridge protocols. > When they detect (or snoop) RSVP-TE message, they establish L2-LSP > and selectively forward those GMPLS prefixed frames via the L2-LSP. > The message path that RSVP-TE is exchanged, thus L2-LSPs are=20 > established, does not necessarily be identical to spanning tree path > because L2-LSPs can be established via shortest path. > I welcome further discussion on this issue, but I hope to continue it=20 > after Paris meeting, if you may. If RSVP-TE traffic uses link-state based protocols and bridged traffic uses distance-based protocols (e.g., RSTP) and these traffic operate independently from each other and both types of traffic come through the same port, then how does a standard IEEE bridge know how to discern these packets and which packets to take action on ? - e.g., what IEEE procedures do you recommend to be used ?? If you are talking about different physical ports, then you are talking about different independent networks with different set of links and topology for each which is not the intent here. Cheers, Ali =20 > =20 > =20 > Thanks > =20 > Jaihyung > =20 >=20 > -----?? ???-----=20 > ?? ??: "Ali Sajassi (sajassi)" <sajassi@cisco.com>=20 > ?? ??: 2005-07-26 ?? 8:13:56=20 > ?? ??: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles=20 > Juergen" <juergen.heiles@siemens.com>, "per@defero.se"=20 > <per@defero.se>=20 > ??: "ccamp@ops.ietf.org" <ccamp@ops.ietf.org>=20 > ??: RE: Frameformat in a l2cs gmpls rnvironment - Issues w/=20 > GMPLS Ethernet=20 >=20 >=20 >=20 >=20 >=20 >=20 >=20 >=20 > Jaihyung,=20 >=20 > It seems to me that your primary objective is to use the existing=20 > Ethernet switches with new control plane (GMPLS) to setup TE=20 > paths for=20 > different applications. And you mentioned three approaches for doing=20 > this in your email. However, all three of them have issues=20 > and are not=20 > viable. Lets go over them ...=20 >=20 > >=20 > > In below mail, I discussed about three proposed approaches.=20 > > =20 > > =20 > > 1. using VLAN tag as it is for L2 label encoding.=20 >=20 > As you know this approach has scalability issue and you are=20 > limited to=20 > 4K services in the network and if you try to use the upcoming IEEE=20 > 802.1ad standard to make VLANs, interface specific, then you=20 > will need=20 > new switches and not existing Ethernet switches. Even if you=20 > use these=20 > newer switches, you will still be limited to 4K services per=20 > interface=20 > and it would defeat the objective of using existing switches.=20 >=20 > >=20 > > 2. defining new protocol ID (TPID) and borrow VLAN tag format,=20 > > ?or extended VLAN tag to implement label swapping.=20 > >=20 >=20 > This approach requires new data plane functionality to be implemented=20 > (which defeats the objective of using existing switches).=20 >=20 > > 3. use lower 3 bytes of MAC address for L2 label encoding.=20 > > =20 >=20 > This approach has several issues:=20 > 1) how does MAC addresses get distributed among the bridges ?=20 > How does=20 > this distribution works during a link or node failure ? How does it=20 > recover from a link or node failure ?=20 >=20 > 2) Will you assign a separate provider MAC for each customer=20 > MAC. If so,=20 > then how do you address MAC scalability. If not, then how do you take=20 > care of 1-to-N mapping between provider MAC address and customer MAC=20 > addresses and how do you handle multipoint connections.=20 >=20 > 3) how do handle customer FCS retention when doing this MAC=20 > translation=20 > stuff ??=20 >=20 > >=20 > > 802.1Q bridge forwards Ethernet frames using two dataplane tables=20 > > - MAC forwarding table and VLAN forwarding table.=20 > > Bridge control protocols, such as GARP, GVRP, GMRP,=20 > > manipulate one of the two dataplane entities.=20 >=20 > This is not correct. Bridges don't have different databases for MAC=20 > forwarding and VLAN forwarding. They have filtering datebase(s) with=20 > corresponding IDs (FIDs). VLAN IDs get mapped to FIDs based=20 > on operation=20 > mode (IVL versus SVL). Then once a filtering database is identified,=20 > then a MAC address lookup is performed for forwarding. Bridge control=20 > protocols use a designated MAC addressed and based on these MAC=20 > addressed, the bridge knows how to process the frame.=20 >=20 > >=20 > > Similarly, option 1 and 3 are about which one of two=20 > > dataplane entities=20 > > GMPLS protocol should control on behalf of bridge control=20 > protocols.=20 > > The two proposals do not intend to modify bridge behavior=20 > > seriously, such as MAC learning, aging, filtering.=20 > > Therefore, the approaches 1 and 3 are in the scope of CCAMP.=20 >=20 > If that's the intention, you may find out soon you get more than what=20 > you asked for :-)=20 >=20 > >=20 > > (% note however, I would regard implementing label swapping=20 > > function is acceptable change considering the Internal Sublayer=20 > > design of 802.1D bridge.)=20 > >=20 >=20 > The only way of doing this is using VID translation table=20 > specified in=20 > IEEE 802.1ad which requires new data-plane functionality and thus=20 > existing bridges cannot be used !!=20 >=20 > Also if going the route of tag swapping, you may consider using IEEE=20 > 802.1ah. As the editor of .1ah, I can go over the details=20 > when I get a=20 > chance.=20 > =20 > >=20 > > However, option 2 is about what TPID code we will=20 > > choose to implement new switching function.=20 > > The switch of this implementation should have new hardware=20 > > design in addition to normal bride hardware.=20 > > I think such proposal as option 2 is out of CCAMP scope.=20 >=20 > Yes.=20 >=20 > >=20 > > Overall objective is improving scalability, traffic=20 > engineering (TE)=20 > > characteristics of 802.1 bridge that it can be reliable, manageable=20 > > enough to replace some core routers.=20 >=20 > Easier said than done :-)=20 >=20 > > The switching technique you mentioned above, such as=20 > > Ethernet over MPLS as defined by PWE3 and L2VPN,=20 > > are all actually router based technology, however this=20 > > work is based on simple bridge architecture.=20 > > Cost-effectiveness is the key differentiator.=20 > >=20 >=20 > The devils are in the details and once you worked out the=20 > details, you=20 > will see the issues.=20 > =20 > >=20 > > In access or enterprise network, capability of providing end-to-end=20 > > L2-LSP will enable service providers policing, measuring, charging=20 > > application flows using Ethernet based network. This will=20 > eventually=20 > > improve income structure and introduce new session based commercial=20 > > service. I have been discussing this aspect in mail thread of=20 > > title 'End-to-end L2-LSP'. Please read the mail thread and comment=20 > > on the discussion.=20 > > =20 > >=20 > > >You mentioned to use the MAC address or part of it as a label.=20 > > >The TRILL working group is defining shortest path routing=20 > > for Ethernet.=20 > > >How would this fit together?=20 > > =20 > >=20 > > TRILL employs new encapsulation layer outside of Ethernet frames.=20 > > However in my proposal of option-3, I simply use the=20 > Ethernet header=20 > > to encode GMPLS label.=20 > >=20 > > The overall format of rbridge is =20 > > [Ethernet][r-tag][Ethernet][ data..]=20 > > It is not LSP based technology but a connectionless routing bridge.=20 > > TTL count is very important in rbridge because frames are routed=20 > > hop-by-hop and it may loop.=20 > >=20 > > Whereas in option-3 above, Ethernet frames that have certain=20 > > OUI prefix in MAC address are forwarded via a path that=20 > > GMPLS protocol has configured on MAC forwarding table.=20 > > =20 >=20 > How does the redundancy work in here. Are you using bridge=20 > protocol in=20 > conjunction with GMPLS protocol. If you do, then there are bunch of=20 > issues. If you don't, then what is it used and how does it=20 > interact with=20 > bridging protocols ??=20 >=20 > -Ali=20 >=20 >=20 > >=20 > > Thanks=20 > > =20 > > Jaihyung=20 > > =20 > > =20 > >=20 > > -----?? ???-----=20 > > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com>=20 > > ?? ??: 2005-07-25 ?? 6:44:34=20 > > ?? ??: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles=20 > > Juergen" <juergen.heiles@siemens.com>, "per@defero.se"=20 > > <per@defero.se>=20 > > ??: "ccamp@ops.ietf.org" <ccamp@ops.ietf.org>=20 > > ??: RE: Frameformat in a l2cs gmpls rnvironment.=20 > >=20 > >=20 > >=20 > >=20 > >=20 > >=20 > > Dear Jaihyung,=20 > >=20 > > you mention several options that could be used as a label.=20 > > Most of them have strong impacts on the data plane and=20 > > introduce new layer networks. Is this in the scope of CCAMP?=20 > > What is the goal, to introduce a control plane for a existing=20 > > L2 technology or introduce a new L2 technology? What would be=20 > > the benefit of such a new L2 co switching technology compared=20 > > to Ethernet over MPLS as defined by PWE3 and L2VPN?=20 > > You mentioned to use the MAC address or part of it as a=20 > > label. The TRILL working group is defining shortest path=20 > > routing for Ethernet. How would this fit together?=20 > >=20 > >=20 > > Regards=20 > >=20 > > Juergen=20 > > =20 > >=20 > >=20 > >=20 > >=20 > > > -----Original Message-----=20 > > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr]=20 > > > Sent: Saturday, July 23, 2005 6:19 AM=20 > > > To: Heiles Juergen; per@defero.se=20 > > > Cc: ccamp@ops.ietf.org=20 > > > Subject: RE: Frameformat in a l2cs gmpls rnvironment.=20 > > >=20 > > >=20 > > > Dear Juergen, Par and all,=20 > > >=20 > > > The proposed framework of L2SC was not intended to suggest=20 > > > any specific solution. It is still an open question.=20 > > > As Adrian noted, it only aims to present requirements=20 > > > as in scenarios that may drive people toward next step in many=20 > > > viable solutions.=20 > > >=20 > > > In fact, there are other options we may also consider for L2=20 > > > label encoding.=20 > > > IEEE 802.1 Ethernet bridges forward frames based on 48bits of=20 > > > MAC address,=20 > > > and additionally using VLAN tag.=20 > > >=20 > > > When the purpose of GMPLS control over Ethernet is not to=20 > > > create new dataplane=20 > > > but to utilize IEEE 802.1 bridge architecture, we may=20 > > consider using=20 > > > one of the two forwarding methods, MAC forwarding or VLAN=20 > > forwarding.=20 > > > (and perhaps any other combination of fields in MAC, but I'll=20 > > > not discuss it)=20 > > > =20 > > > Use of VLAN ID for label encoding may automate VLAN configuration=20 > > > using IP protocols. However, GMPLS protocol cannot use the=20 > > > field exclusively=20 > > > because public/private operators already use VLAN for various=20 > > > purpose.=20 > > > There is a potential conflict with existing use of VLAN and=20 > > > GMPLS use of VLAN label. =20 > > > Furthermore, scalability of VLAN ID has been frequently noted=20 > > > as weakness=20 > > > because the size of VLAN ID is at most 4096 (12bit).=20 > > >=20 > > > The scalability may be improved if the scope of VLAN label is=20 > > > confined to=20 > > > link-local, and some additional swapping function of VLAN ID=20 > > > is introduced in=20 > > > Internal Sublayer [802.1D] of GMPLS implemented switch.=20 > > > However, this will only be effective when the configuration=20 > > > of network=20 > > > is mesh structure that multiple LSP paths exist. If the=20 > > > configuration of network=20 > > > is star or tree shape, as normal configuration of access network,=20 > > > LSPs concentrate in root node and total number of=20 > > > LSPs that the network can hold still be limited by available=20 > > > label space=20 > > > at a few root links.=20 > > >=20 > > > For these reasons, I do not think any form of VLAN ID label is=20 > > > an appropriate choice for layer-2 label encoding.=20 > > > =20 > > > There are some other proposals assuming new assignment of=20 > Ethernet=20 > > > Length/Type value (e.g. new TPID in VLAN tag) and=20 > re-definition of=20 > > > information fields placed between 802.3 MAC header and IP packet.=20 > > > In this case, only the format of VLAN tag or extended VLAN tag=20 > > > is borrowed, however, inside the switching hardware, the=20 > > > filter and relay,=20 > > > etc. are totally different new dataplane switch.=20 > > >=20 > > > I do not see such approach is a GMPLS implementation for Ethernet=20 > > > because the core switching technique is not 802.1 Ethernet=20 > > > bridge at all.=20 > > > =20 > > > The other option we may consider is using MAC address filed=20 > > as below.=20 > > > =20 > > >=20 > > > +-------+-------+-------+-------+-------+-------+=20 > > > | 1byte | 2byte | 3byte | 4byte | 5byte | 6byte |=20 > > > +-------+-------+-------+-------+-------+-------+=20 > > > +-----------------------+-----------------------+=20 > > > | OUI Prefix (=3DGMPLS) | DA-label (24bit) |=20 > > > +-----------------------+-----------------------+=20 > > > | OUI Prefix (=3DGMPLS) | SA-label (24bit) |=20 > > > +---------------+-------+-----------------------+=20 > > > | Length/Type |=20 > > > +---------------+=20 > > > =20 > > >=20 > > > IEEE is designated by the ISO Council to act as the=20 > > > registration authority=20 > > > for the higher three-octet of OUI number in the MAC address=20 > > > to be used by manufacturer. Ethernet manufacturer may generate=20 > > > global unique MAC address using the OUI prefix and=20 > address block of=20 > > > lower three-octet (24bit). Taking advantage of the=20 > > addressing scheme,=20 > > > GMPLS may use the lower three-octet exclusively if a unique=20 > > > OUI number=20 > > > is reserved for the protocol. With this labeling scheme,=20 > GMPLS will=20 > > > control MAC forwarding entry, not VLAN table.=20 > > > =20 > > > All Ethernet frames controlled by GMPLS will have identical=20 > > > OUI number=20 > > > that they can easily be distinguished from other Ethernet frames.=20 > > > In principle, the label lookup hardware is identical to=20 > MAC lookup=20 > > > hardware in this labeling scheme. Therefore GMPLS implemented=20 > > > switch may still function as normal Ethernet bridge to the frames=20 > > > that OUI number is not GMPLS. This also facilitates GMPLS=20 > > implemented=20 > > > switches being deployed in operating Ethernet with minimum=20 > > > service disruption.=20 > > > =20 > > > Note also that above proposed label encoding method is=20 > transparent=20 > > > to the use of Ethernet Length/Type field. End-user device may use=20 > > > the Length/Type field as defined in IEEE 802.3 protocol.=20 > > > It also allows network operators configure VLAN for their own=20 > > > purpose.=20 > > > When IEEE 802.1p is used in conjunction with L2-LSP, the priority=20 > > > field of VLAN tag can also be used for imposing consistent TE=20 > > > policy in=20 > > > legacy switches and GMPLS switches.=20 > > > =20 > > > Any way, my conclusion is, there are other options we may=20 > consider,=20 > > > and this issue is still open to discuss.=20 > > > =20 > > > Thanks,=20 > > > =20 > > > Sincerely,=20 > > > =20 > > > Jaihyung=20 > > > =20 > > > =20 > > > =20 > > > =20 > > >=20 > > > -----?? ???-----=20 > > > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com>=20 > > > ?? ??: 2005-07-22 ?? 10:23:43=20 > > > ?? ??: "Loa Andersson" <loa@pi.se>, "richard.spencer@bt.com"=20 > > > <richard.spencer@bt.com>=20 > > > ??: "per@defero.se" <per@defero.se>, "ccamp@ops.ietf.org"=20 > > > <ccamp@ops.ietf.org>=20 > > > ??: RE: Frameformat in a l2cs gmpls rnvironment.=20 > > >=20 > > >=20 > > >=20 > > >=20 > > >=20 > > >=20 > > > Loa,=20 > > >=20 > > > I interpret the ID as a proposal to use GMPLS for VLAN setup.=20 > > > So GMPLS and VLAN to not compete. The VLAN is at the data=20 > > > plane and GMPLS at the control plane. The question is how and=20 > > > should different control plane techniques like GMPLS and=20 > > > (GVRP and STB) work together?=20 > > >=20 > > > Regards=20 > > >=20 > > > Juergen=20 > > >=20 > > >=20 > > > > -----Original Message-----=20 > > > > From: owner-ccamp@ops.ietf.org=20 > > > > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson=20 > > > > Sent: Friday, July 22, 2005 2:53 PM=20 > > > > To: richard.spencer@bt.com=20 > > > > Cc: per@defero.se; ccamp@ops.ietf.org=20 > > > > Subject: Re: Frameformat in a l2cs gmpls rnvironment.=20 > > > >=20 > > > > Richard,=20 > > > >=20 > > > > I agree to most of this. Since we have doubts about the=20 > viability=20 > > > > of taking GMPLS all the way to end-user or enterprise I=20 > think it=20 > > > > would be good, from a wg perspective, if we agreed to solve the=20 > > > > core network problems first.=20 > > > >=20 > > > > Do you have any comment on the requirement to run both=20 > VLANs and=20 > > > > GMPLS on the same switch?=20 > > > >=20 > > > > /Loa=20 > > > >=20 > > > > richard.spencer@bt.com wrote:=20 > > > > > Regardless of whether or not a switch is directly connected=20 > > > > > to hosts, it must be able to forward packets using the=20 > > > > connectionless=20 > > > > > Ethernet data plane. This is due to the fundamental=20 > > > > requirement that=20 > > > > > for GMPLS switches to be able to exchange control=20 > > > information with=20 > > > > > each other, a data plane for control traffic must be=20 > present.=20 > > > > > This is akin to using the IP data plane for MPLS=20 > > > signalling in an=20 > > > > > IP/MPLS network. An alternative would be to use a static=20 > > > reserved=20 > > > > > L2-LSP for control traffic in the same way that reserved=20 > > > VPI/VCIs=20 > > > > > are used for PNNI signalling in ATM.=20 > > > > >=20 > > > > > Regarding connecting hosts to GMPLS switches, I=20 > > personally don't=20 > > > > > think extending L2-LSPs into the enterprise/home network is=20 > > > > > commercially viable. However, if you do want to use=20 > > > GMPLS switches=20 > > > > > in the home/enterprise network and for some reason=20 > > don't want to=20 > > > > > extend L2-LSPs down to the host then you will not be=20 > > performing=20 > > > > > normal Ethernet Mac address switching anyway.=20 > Instead you will=20 > > > > > need some kind of policy on the switch that maps=20 > > connectionless=20 > > > > > Ethernet packets (e.g. based on MAC src/dest, 802.1p,=20 > > VLAN) to a=20 > > > > > L2-LSP. This is because multiple L2-LSPs to the same=20 > > destination=20 > > > > > (e.g. a gateway router) may exist for different=20 > services/flows=20 > > > > > (e.g. video download, VoIP call, etc.).=20 > > > > >=20 > > > > > Regards,=20 > > > > > Richard=20 > > > > >=20 > > > > >=20 > > > > >>-----Original Message-----=20 > > > > >>From: owner-ccamp@ops.ietf.org=20 > > > [mailto:owner-ccamp@ops.ietf.org]On=20 > > > > >>Behalf Of Par Mattsson=20 > > > > >>Sent: 22 July 2005 11:42=20 > > > > >>To: Loa Andersson=20 > > > > >>Cc: ccamp@ops.ietf.org=20 > > > > >>Subject: Re: Frameformat in a l2cs gmpls rnvironment.=20 > > > > >>=20 > > > > >>=20 > > > > >>=20 > > > > >>>Per and Dimitri,=20 > > > > >>>=20 > > > > >>>I would like to come down stronger than that, for me it is=20 > > > > >>>a very strong requirement that the same switch can handle=20 > > > > >>>both VLANs and GMPLs trafic correctly. I can't dsee how that=20 > > > > >>>could be done if using the VLAN tpid to indicate GMPLS=20 > > > > >>>traffic.=20 > > > > >>=20 > > > > >>If you ever want that same switch to handle traffic for a=20 > > > directly=20 > > > > >>connected host (not to uncommen) you would want that to=20 > > > use normal=20 > > > > >>ethernet macaddress switching. So of course you do not want=20 > > > > to have to=20 > > > > >>choose between vlan and gmpls, you would want both at the=20 > > > same time.=20 > > > > >>=20 > > > > >>/per=20 > > > > >>=20 > > > > >>=20 > > > > >>=20 > > > > >>>/Loa=20 > > > > >>>=20 > > > > >>>Par Mattsson wrote:=20 > > > > >>>=20 > > > > >>>>>hi par, one of the possibilities that has been considered=20 > > > > >>=20 > > > > >>to cope with=20 > > > > >>=20 > > > > >>>>>this requirement is to use a dedicated TPID for the=20 > > > > >>=20 > > > > >>Ethernet labeled=20 > > > > >>=20 > > > > >>>>>frames; this would allow differentiated processing with=20 > > > > non-labeled=20 > > > > >>>>>framesthanks.=20 > > > > >>>>=20 > > > > >>>>=20 > > > > >>>>That seems to make more sence. If that frame is to be=20 > > > sized like a=20 > > > > >>>>802.1q=20 > > > > >>>>frame. There is not that much space left to a label. Or is=20 > > > > >>=20 > > > > >>the demand to=20 > > > > >>=20 > > > > >>>>use jumboframes ?=20 > > > > >>>>Has there been any discussion on labelstacking, and=20 > > > > mainly where to=20 > > > > >>>>place=20 > > > > >>>>the information?=20 > > > > >>>>=20 > > > > >>>>Regards.=20 > > > > >>>>Per=20 > > > > >>>>=20 > > > > >>>>=20 > > > > >>>>=20 > > > > >>>=20 > > > > >>>=20 > > > > >>>--=20 > > > > >>>Loa Andersson=20 > > > > >>>=20 > > > > >>>Principal Networking Architect=20 > > > > >>>Acreo AB phone: +46 8 632 77 14=20 > > > > >>>Isafjordsgatan 22 mobile: +46 739 81 21 64=20 > > > > >>>Kista, Sweden email: =20 > > > loa.andersson@acreo.se=20 > > > > >>> loa@pi.se=20 > > > > >>>=20 > > > > >>=20 > > > > >>=20 > > > > >>=20 > > > > >=20 > > > >=20 > > > >=20 > > > > --=20 > > > > Loa Andersson=20 > > > >=20 > > > > Principal Networking Architect=20 > > > > Acreo AB phone: +46 8 632 77 14=20 > > > > Isafjordsgatan 22 mobile: +46 739 81 21 64=20 > > > > Kista, Sweden email: =20 > loa.andersson@acreo.se=20 > > > > loa@pi.se=20 > > > >=20 > > >=20 > > >=20 > > >=20 > > >=20 > >=20 > >=20 > >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 26 Jul 2005 23:22:05 +0000 To: "Ali Sajassi \(sajassi\)" <sajassi@cisco.com> Cc: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles Juergen" <juergen.heiles@siemens.com>, <per@defero.se>, <ccamp@ops.ietf.org> MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: RE: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet Date: Wed, 27 Jul 2005 01:20:45 +0200 Message-ID: <OF2F971F53.760A58CB-ONC125704A.00803D87-C125704A.00803E6B@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmFsaTwvRk9OVD48L1A+PFA+PEZPTlQg RkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPnRvIGJlIGNsZWFyIGhlcmUgeW91IGFyZSBoYW5naW5n IG9uIHVzaW5nIHRoZSB0ZXJtICZxdW90O2V4aXN0aW5nIEV0aGVybmV0IHN3aXRjaGVzJnF1b3Q7 IHdoaWxlIHRoZSBkb2N1bWVudCB0aGF0IGhhcyBiZWVuIGF0IHRoZSBzb3VyY2Ugb2YgdGhlIGRp c2N1c3Npb24gZG9lcyBub3Qgc3RhdGUgJnF1b3Q7R01QTFMgbXVzdCBiZSBkZXBsb3lhYmxlIG9u IGV4aXN0aW5nIEV0aGVybmV0IHN3aXRjaGVzJnF1b3Q7IHRoaXMgbWF5IGJlIHRoZSBjYXNlIGJ1 dCB0aGVyZSBhcmUgbGltaXRhdGlvbnMgKGFuZCB0aGVyZSBhcmUgcGVyZmVjdGx5IHdlbGwgdW5k ZXJzdG9vZCwgZG9uJ3Qgd29ycnkpIGJlc2lkZSB0aGlzIHlvdSBzaG91bGQgYWxzbyB0ZWxsIHVz IHdoYXQgZG8geW91IGV4YWN0bHkgbWVhbiB3aXRoIHRoZSB0ZXJtICZxdW90O2V4aXN0aW5nJnF1 b3Q7IC0gYnV0IHRoaXMgaXMgYSBzaWRlIGlzc3VlIC08L0ZPTlQ+PC9QPjxQPjxGT05UIEZBQ0U9 Ik1vbm9zcGFjZSxDb3VyaWVyIj50aGlzIHNhaWQsIHRoaXMgZG9lcyBOT1QgbWVhbiB0aGF0IGEg R01QTFMgRXRoZXJuZXQgc3dpdGNoIChlaXRoZXIgZXhpc3Rpbmcgb3IgZXZlbiAmcXVvdDtleHRl bmRlZCZxdW90Oykgd291bGQgbmVlZCB0byBzdXBwb3J0IGEgbmV3IEV0aGVybmV0IDgwMi4zIGZy YW1lIGZvcndhcmRpbmcgcGFyYWRpZ20sIHRoZSBkaWZmZXJlbmNlIGlzIHN1YnRsZSBidXQgaGFz IHRvIGJlIHVuZGVybGluZWQ8L0ZPTlQ+PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3Vy aWVyIj5ub3RlOiB0aGVyZSBpcyBhIHNsb3Qgb2YgMjAgbWludXRlcyBkdXJpbmcgdGhlIG5leHQg Q0NBTVAgV0cgbWVldGluZyB3aGVyZSB0aGlzIGRvY3VtZW50IGFuZCBhbGwgcmVsYXRlZCBpc3N1 ZXMgd2lsbCBiZSBoZWF2aWx5IGRpc2N1c3NlZCAuLi4gc28gaSB3ZWxjb21lIHlvdSB0byBqb2lu IHRoZSBmZXN0aXZpdGllcyA7LSk8L0ZPTlQ+PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxD b3VyaWVyIj50aGFua3MsPC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmll ciI+LSBkaW1pdHJpLjwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIi PiZuYnNwOyA8L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTI+PEI+JnF1b3Q7QWxpIFNhamFzc2kg XChzYWphc3NpXCkmcXVvdDsgJmx0O3NhamFzc2lAY2lzY28uY29tJmd0OzwvQj48L0ZPTlQ+PEJS PjxGT05UIFNJWkU9Mj5TZW50IGJ5OiBvd25lci1jY2FtcEBvcHMuaWV0Zi5vcmc8L0ZPTlQ+PEJS PjxGT05UIFNJWkU9Mj4wNy8yNS8yMDA1IDE2OjEzIE1TVDwvRk9OVD48QlI+PEJSPiA8Rk9OVCBT SVpFPTI+VG86PC9GT05UPiA8Rk9OVCBTSVpFPTI+JnF1b3Q7Q0hPLCBKQUkgSFlVTkcmcXVvdDsg Jmx0O2phaWh5dW5nQGV0cmkucmUua3ImZ3Q7LCAmcXVvdDtIZWlsZXMgSnVlcmdlbiZxdW90OyAm bHQ7anVlcmdlbi5oZWlsZXNAc2llbWVucy5jb20mZ3Q7LCAmbHQ7cGVyQGRlZmVyby5zZSZndDs8 L0ZPTlQ+PEJSPiA8Rk9OVCBTSVpFPTI+Y2M6PC9GT05UPiA8Rk9OVCBTSVpFPTI+Jmx0O2NjYW1w QG9wcy5pZXRmLm9yZyZndDs8L0ZPTlQ+PEJSPiA8Rk9OVCBTSVpFPTI+YmNjOjwvRk9OVD4gPEJS PiA8Rk9OVCBTSVpFPTI+U3ViamVjdDo8L0ZPTlQ+IDxGT05UIFNJWkU9Mj5SRTogRnJhbWVmb3Jt YXQgaW4gYSBsMmNzIGdtcGxzIHJudmlyb25tZW50IC0gSXNzdWVzIHcvIEdNUExTIEV0aGVybmV0 PC9GT05UPjxCUj4gPEJSPjxCUj48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIi PkphaWh5dW5nLDxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5J dCBzZWVtcyB0byBtZSB0aGF0IHlvdXIgcHJpbWFyeSBvYmplY3RpdmUgaXMgdG8gdXNlIHRoZSBl eGlzdGluZzxCUj5FdGhlcm5ldCBzd2l0Y2hlcyB3aXRoIG5ldyBjb250cm9sIHBsYW5lIChHTVBM UykgdG8gc2V0dXAgVEUgcGF0aHMgZm9yPEJSPmRpZmZlcmVudCBhcHBsaWNhdGlvbnMuIEFuZCB5 b3UgbWVudGlvbmVkIHRocmVlIGFwcHJvYWNoZXMgZm9yIGRvaW5nPEJSPnRoaXMgaW4geW91ciBl bWFpbC4gSG93ZXZlciwgYWxsIHRocmVlIG9mIHRoZW0gaGF2ZSBpc3N1ZXMgYW5kIGFyZSBub3Q8 QlI+dmlhYmxlLiBMZXRzIGdvIG92ZXIgdGhlbSAuLi48QlI+PC9GT05UPjxCUj48Rk9OVCBGQUNF PSJNb25vc3BhY2UsQ291cmllciI+Jmd0OzxCUj4mZ3Q7IEluIGJlbG93IG1haWwsIEkgZGlzY3Vz c2VkIGFib3V0IHRocmVlIHByb3Bvc2VkIGFwcHJvYWNoZXMuPEJSPiZndDs8QlI+Jmd0OzxCUj4m Z3Q7IDEuIHVzaW5nIFZMQU4gdGFnIGFzIGl0IGlzIGZvciBMMiBsYWJlbCBlbmNvZGluZy48QlI+ PC9GT05UPjxCUj48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+QXMgeW91IGtub3cgdGhp cyBhcHByb2FjaCBoYXMgc2NhbGFiaWxpdHkgaXNzdWUgYW5kIHlvdSBhcmUgbGltaXRlZCB0bzxC Uj40SyBzZXJ2aWNlcyBpbiB0aGUgbmV0d29yayBhbmQgaWYgeW91IHRyeSB0byB1c2UgdGhlIHVw Y29taW5nIElFRUU8QlI+ODAyLjFhZCBzdGFuZGFyZCB0byBtYWtlIFZMQU5zLCBpbnRlcmZhY2Ug c3BlY2lmaWMsIHRoZW4geW91IHdpbGwgbmVlZDxCUj5uZXcgc3dpdGNoZXMgYW5kIG5vdCBleGlz dGluZyBFdGhlcm5ldCBzd2l0Y2hlcy4gRXZlbiBpZiB5b3UgdXNlIHRoZXNlPEJSPm5ld2VyIHN3 aXRjaGVzLCB5b3Ugd2lsbCBzdGlsbCBiZSBsaW1pdGVkIHRvIDRLIHNlcnZpY2VzIHBlciBpbnRl cmZhY2U8QlI+YW5kIGl0IHdvdWxkIGRlZmVhdCB0aGUgb2JqZWN0aXZlIG9mIHVzaW5nIGV4aXN0 aW5nIHN3aXRjaGVzLjxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVy Ij4mZ3Q7PEJSPiZndDsgMi4gZGVmaW5pbmcgbmV3IHByb3RvY29sIElEIChUUElEKSBhbmQgYm9y cm93IFZMQU4gdGFnIGZvcm1hdCw8QlI+Jmd0OyA/b3IgZXh0ZW5kZWQgVkxBTiB0YWcgdG8gaW1w bGVtZW50IGxhYmVsIHN3YXBwaW5nLjxCUj4mZ3Q7PEJSPjwvRk9OVD48QlI+PEZPTlQgRkFDRT0i TW9ub3NwYWNlLENvdXJpZXIiPlRoaXMgYXBwcm9hY2ggcmVxdWlyZXMgbmV3IGRhdGEgcGxhbmUg ZnVuY3Rpb25hbGl0eSB0byBiZSBpbXBsZW1lbnRlZDxCUj4od2hpY2ggZGVmZWF0cyB0aGUgb2Jq ZWN0aXZlIG9mIHVzaW5nIGV4aXN0aW5nIHN3aXRjaGVzKS48QlI+PC9GT05UPjxCUj48Rk9OVCBG QUNFPSJNb25vc3BhY2UsQ291cmllciI+Jmd0OyAzLiB1c2UgbG93ZXIgMyBieXRlcyBvZiBNQUMg YWRkcmVzcyBmb3IgTDIgbGFiZWwgZW5jb2RpbmcuPEJSPiZndDs8QlI+PC9GT05UPjxCUj48Rk9O VCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+VGhpcyBhcHByb2FjaCBoYXMgc2V2ZXJhbCBpc3N1 ZXM6PEJSPjEpIGhvdyBkb2VzIE1BQyBhZGRyZXNzZXMgZ2V0IGRpc3RyaWJ1dGVkIGFtb25nIHRo ZSBicmlkZ2VzID8gSG93IGRvZXM8QlI+dGhpcyBkaXN0cmlidXRpb24gd29ya3MgZHVyaW5nIGEg bGluayBvciBub2RlIGZhaWx1cmUgPyBIb3cgZG9lcyBpdDxCUj5yZWNvdmVyIGZyb20gYSBsaW5r IG9yIG5vZGUgZmFpbHVyZSA/PEJSPjwvRk9OVD48QlI+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENv dXJpZXIiPjIpIFdpbGwgeW91IGFzc2lnbiBhIHNlcGFyYXRlIHByb3ZpZGVyIE1BQyBmb3IgZWFj aCBjdXN0b21lciBNQUMuIElmIHNvLDxCUj50aGVuIGhvdyBkbyB5b3UgYWRkcmVzcyBNQUMgc2Nh bGFiaWxpdHkuIElmIG5vdCwgdGhlbiBob3cgZG8geW91IHRha2U8QlI+Y2FyZSBvZiAxLXRvLU4g bWFwcGluZyBiZXR3ZWVuIHByb3ZpZGVyIE1BQyBhZGRyZXNzIGFuZCBjdXN0b21lciBNQUM8QlI+ YWRkcmVzc2VzIGFuZCBob3cgZG8geW91IGhhbmRsZSBtdWx0aXBvaW50IGNvbm5lY3Rpb25zLjxC Uj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj4zKSBob3cgZG8gaGFu ZGxlIGN1c3RvbWVyIEZDUyByZXRlbnRpb24gd2hlbiBkb2luZyB0aGlzIE1BQyB0cmFuc2xhdGlv bjxCUj5zdHVmZiA/PzxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVy Ij4mZ3Q7PEJSPiZndDsgODAyLjFRIGJyaWRnZSBmb3J3YXJkcyBFdGhlcm5ldCBmcmFtZXMgdXNp bmcgdHdvIGRhdGFwbGFuZSB0YWJsZXM8QlI+Jmd0OyAtIE1BQyBmb3J3YXJkaW5nIHRhYmxlIGFu ZCBWTEFOIGZvcndhcmRpbmcgdGFibGUuPEJSPiZndDsgQnJpZGdlIGNvbnRyb2wgcHJvdG9jb2xz LCBzdWNoIGFzIEdBUlAsIEdWUlAsIEdNUlAsPEJSPiZndDsgbWFuaXB1bGF0ZSBvbmUgb2YgdGhl IHR3byBkYXRhcGxhbmUgZW50aXRpZXMuPEJSPjwvRk9OVD48QlI+PEZPTlQgRkFDRT0iTW9ub3Nw YWNlLENvdXJpZXIiPlRoaXMgaXMgbm90IGNvcnJlY3QuIEJyaWRnZXMgZG9uJ3QgaGF2ZSBkaWZm ZXJlbnQgZGF0YWJhc2VzIGZvciBNQUM8QlI+Zm9yd2FyZGluZyBhbmQgVkxBTiBmb3J3YXJkaW5n LiBUaGV5IGhhdmUgZmlsdGVyaW5nIGRhdGViYXNlKHMpIHdpdGg8QlI+Y29ycmVzcG9uZGluZyBJ RHMgKEZJRHMpLiBWTEFOIElEcyBnZXQgbWFwcGVkIHRvIEZJRHMgYmFzZWQgb24gb3BlcmF0aW9u PEJSPm1vZGUgKElWTCB2ZXJzdXMgU1ZMKS4gVGhlbiBvbmNlIGEgZmlsdGVyaW5nIGRhdGFiYXNl IGlzIGlkZW50aWZpZWQsPEJSPnRoZW4gYSBNQUMgYWRkcmVzcyBsb29rdXAgaXMgcGVyZm9ybWVk IGZvciBmb3J3YXJkaW5nLiBCcmlkZ2UgY29udHJvbDxCUj5wcm90b2NvbHMgdXNlIGEgZGVzaWdu YXRlZCBNQUMgYWRkcmVzc2VkIGFuZCBiYXNlZCBvbiB0aGVzZSBNQUM8QlI+YWRkcmVzc2VkLCB0 aGUgYnJpZGdlIGtub3dzIGhvdyB0byBwcm9jZXNzIHRoZSBmcmFtZS48QlI+PC9GT05UPjxCUj48 Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+Jmd0OzxCUj4mZ3Q7IFNpbWlsYXJseSwgb3B0 aW9uIDEgYW5kIDMgYXJlIGFib3V0IHdoaWNoIG9uZSBvZiB0d288QlI+Jmd0OyBkYXRhcGxhbmUg ZW50aXRpZXM8QlI+Jmd0OyBHTVBMUyBwcm90b2NvbCBzaG91bGQgY29udHJvbCBvbiBiZWhhbGYg b2YgYnJpZGdlIGNvbnRyb2wgcHJvdG9jb2xzLjxCUj4mZ3Q7IFRoZSB0d28gcHJvcG9zYWxzIGRv IG5vdCBpbnRlbmQgdG8gbW9kaWZ5IGJyaWRnZSBiZWhhdmlvcjxCUj4mZ3Q7IHNlcmlvdXNseSwg c3VjaCBhcyBNQUMgbGVhcm5pbmcsIGFnaW5nLCBmaWx0ZXJpbmcuPEJSPiZndDsgVGhlcmVmb3Jl LCB0aGUgYXBwcm9hY2hlcyAxIGFuZCAzIGFyZSBpbiB0aGUgc2NvcGUgb2YgQ0NBTVAuPEJSPjwv Rk9OVD48QlI+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPklmIHRoYXQncyB0aGUgaW50 ZW50aW9uLCB5b3UgbWF5IGZpbmQgb3V0IHNvb24geW91IGdldCBtb3JlIHRoYW4gd2hhdDxCUj55 b3UgYXNrZWQgZm9yIDotKTxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3Vy aWVyIj4mZ3Q7PEJSPiZndDsgKCUgbm90ZSBob3dldmVyLCBJIHdvdWxkIHJlZ2FyZCBpbXBsZW1l bnRpbmcgbGFiZWwgc3dhcHBpbmc8QlI+Jmd0OyBmdW5jdGlvbiBpcyBhY2NlcHRhYmxlIGNoYW5n ZSBjb25zaWRlcmluZyB0aGUgSW50ZXJuYWwgU3VibGF5ZXI8QlI+Jmd0OyBkZXNpZ24gb2YgODAy LjFEIGJyaWRnZS4pPEJSPiZndDs8QlI+PC9GT05UPjxCUj48Rk9OVCBGQUNFPSJNb25vc3BhY2Us Q291cmllciI+VGhlIG9ubHkgd2F5IG9mIGRvaW5nIHRoaXMgaXMgdXNpbmcgVklEIHRyYW5zbGF0 aW9uIHRhYmxlIHNwZWNpZmllZCBpbjxCUj5JRUVFIDgwMi4xYWQgd2hpY2ggcmVxdWlyZXMgbmV3 IGRhdGEtcGxhbmUgZnVuY3Rpb25hbGl0eSBhbmQgdGh1czxCUj5leGlzdGluZyBicmlkZ2VzIGNh bm5vdCBiZSB1c2VkICEhPEJSPjwvRk9OVD48QlI+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJp ZXIiPkFsc28gaWYgZ29pbmcgdGhlIHJvdXRlIG9mIHRhZyBzd2FwcGluZywgeW91IG1heSBjb25z aWRlciB1c2luZyBJRUVFPEJSPjgwMi4xYWguIEFzIHRoZSBlZGl0b3Igb2YgLjFhaCwgSSBjYW4g Z28gb3ZlciB0aGUgZGV0YWlscyB3aGVuIEkgZ2V0IGE8QlI+Y2hhbmNlLjxCUj48L0ZPTlQ+PEJS PjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj4mZ3Q7PEJSPiZndDsgSG93ZXZlciwgb3B0 aW9uIDIgaXMgYWJvdXQgd2hhdCBUUElEIGNvZGUgd2Ugd2lsbDxCUj4mZ3Q7IGNob29zZSB0byBp bXBsZW1lbnQgbmV3IHN3aXRjaGluZyBmdW5jdGlvbi48QlI+Jmd0OyBUaGUgc3dpdGNoIG9mIHRo aXMgaW1wbGVtZW50YXRpb24gc2hvdWxkIGhhdmUgbmV3IGhhcmR3YXJlPEJSPiZndDsgZGVzaWdu IGluIGFkZGl0aW9uIHRvIG5vcm1hbCBicmlkZSBoYXJkd2FyZS48QlI+Jmd0OyBJIHRoaW5rIHN1 Y2ggcHJvcG9zYWwgYXMgb3B0aW9uIDIgaXMgb3V0IG9mIENDQU1QIHNjb3BlLjxCUj48L0ZPTlQ+ PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5ZZXMuPEJSPjwvRk9OVD48QlI+PEZP TlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPiZndDs8QlI+Jmd0OyBPdmVyYWxsIG9iamVjdGl2 ZSBpcyBpbXByb3Zpbmcgc2NhbGFiaWxpdHksIHRyYWZmaWMgZW5naW5lZXJpbmcgKFRFKTxCUj4m Z3Q7IGNoYXJhY3RlcmlzdGljcyBvZiA4MDIuMSBicmlkZ2UgdGhhdCBpdCBjYW4gYmUgcmVsaWFi bGUsIG1hbmFnZWFibGU8QlI+Jmd0OyBlbm91Z2ggdG8gcmVwbGFjZSBzb21lIGNvcmUgcm91dGVy cy48QlI+PC9GT05UPjxCUj48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+RWFzaWVyIHNh aWQgdGhhbiBkb25lIDotKTxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3Vy aWVyIj4mZ3Q7IFRoZSBzd2l0Y2hpbmcgdGVjaG5pcXVlIHlvdSBtZW50aW9uZWQgYWJvdmUsIHN1 Y2ggYXM8QlI+Jmd0OyBFdGhlcm5ldCBvdmVyIE1QTFMgYXMgZGVmaW5lZCBieSBQV0UzIGFuZCBM MlZQTiw8QlI+Jmd0OyBhcmUgYWxsIGFjdHVhbGx5IHJvdXRlciBiYXNlZCB0ZWNobm9sb2d5LCBo b3dldmVyIHRoaXM8QlI+Jmd0OyB3b3JrIGlzIGJhc2VkIG9uIHNpbXBsZSBicmlkZ2UgYXJjaGl0 ZWN0dXJlLjxCUj4mZ3Q7IENvc3QtZWZmZWN0aXZlbmVzcyBpcyB0aGUga2V5IGRpZmZlcmVudGlh dG9yLjxCUj4mZ3Q7PEJSPjwvRk9OVD48QlI+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIi PlRoZSBkZXZpbHMgYXJlIGluIHRoZSBkZXRhaWxzIGFuZCBvbmNlIHlvdSB3b3JrZWQgb3V0IHRo ZSBkZXRhaWxzLCB5b3U8QlI+d2lsbCBzZWUgdGhlIGlzc3Vlcy48QlI+PC9GT05UPjxCUj48Rk9O VCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+Jmd0OzxCUj4mZ3Q7IEluIGFjY2VzcyBvciBlbnRl cnByaXNlIG5ldHdvcmssIGNhcGFiaWxpdHkgb2YgcHJvdmlkaW5nIGVuZC10by1lbmQ8QlI+Jmd0 OyBMMi1MU1Agd2lsbCBlbmFibGUgc2VydmljZSBwcm92aWRlcnMgcG9saWNpbmcsIG1lYXN1cmlu ZywgY2hhcmdpbmc8QlI+Jmd0OyBhcHBsaWNhdGlvbiBmbG93cyB1c2luZyBFdGhlcm5ldCBiYXNl ZCBuZXR3b3JrLiBUaGlzIHdpbGwgZXZlbnR1YWxseTxCUj4mZ3Q7IGltcHJvdmUgaW5jb21lIHN0 cnVjdHVyZSBhbmQgaW50cm9kdWNlIG5ldyBzZXNzaW9uIGJhc2VkIGNvbW1lcmNpYWw8QlI+Jmd0 OyBzZXJ2aWNlLiBJIGhhdmUgYmVlbiBkaXNjdXNzaW5nIHRoaXMgYXNwZWN0IGluIG1haWwgdGhy ZWFkIG9mPEJSPiZndDsgdGl0bGUgJ0VuZC10by1lbmQgTDItTFNQJy4gJm5ic3A7UGxlYXNlIHJl YWQgdGhlIG1haWwgdGhyZWFkIGFuZCBjb21tZW50PEJSPiZndDsgb24gdGhlIGRpc2N1c3Npb24u PEJSPiZndDs8QlI+Jmd0OzxCUj4mZ3Q7ICZndDtZb3UgbWVudGlvbmVkIHRvIHVzZSB0aGUgTUFD IGFkZHJlc3Mgb3IgcGFydCBvZiBpdCBhcyBhIGxhYmVsLjxCUj4mZ3Q7ICZndDtUaGUgVFJJTEwg d29ya2luZyBncm91cCBpcyBkZWZpbmluZyBzaG9ydGVzdCBwYXRoIHJvdXRpbmc8QlI+Jmd0OyBm b3IgRXRoZXJuZXQuPEJSPiZndDsgJmd0O0hvdyB3b3VsZCB0aGlzIGZpdCB0b2dldGhlcj88QlI+ Jmd0OzxCUj4mZ3Q7PEJSPiZndDsgVFJJTEwgZW1wbG95cyBuZXcgZW5jYXBzdWxhdGlvbiBsYXll ciBvdXRzaWRlIG9mIEV0aGVybmV0IGZyYW1lcy48QlI+Jmd0OyBIb3dldmVyIGluIG15IHByb3Bv c2FsIG9mIG9wdGlvbi0zLCBJIHNpbXBseSB1c2UgdGhlIEV0aGVybmV0IGhlYWRlcjxCUj4mZ3Q7 IHRvIGVuY29kZSBHTVBMUyBsYWJlbC48QlI+Jmd0OzxCUj4mZ3Q7IFRoZSBvdmVyYWxsIGZvcm1h dCBvZiByYnJpZGdlIGlzPEJSPiZndDsgW0V0aGVybmV0XVtyLXRhZ11bRXRoZXJuZXRdWyBkYXRh Li5dPEJSPiZndDsgSXQgaXMgbm90IExTUCBiYXNlZCB0ZWNobm9sb2d5IGJ1dCBhIGNvbm5lY3Rp b25sZXNzIHJvdXRpbmcgYnJpZGdlLjxCUj4mZ3Q7IFRUTCBjb3VudCBpcyB2ZXJ5IGltcG9ydGFu dCBpbiByYnJpZGdlIGJlY2F1c2UgZnJhbWVzIGFyZSByb3V0ZWQ8QlI+Jmd0OyBob3AtYnktaG9w IGFuZCBpdCBtYXkgbG9vcC48QlI+Jmd0OzxCUj4mZ3Q7IFdoZXJlYXMgaW4gb3B0aW9uLTMgYWJv dmUsIEV0aGVybmV0IGZyYW1lcyB0aGF0IGhhdmUgY2VydGFpbjxCUj4mZ3Q7IE9VSSBwcmVmaXgg aW4gTUFDIGFkZHJlc3MgYXJlIGZvcndhcmRlZCB2aWEgYSBwYXRoIHRoYXQ8QlI+Jmd0OyBHTVBM UyBwcm90b2NvbCBoYXMgY29uZmlndXJlZCBvbiBNQUMgZm9yd2FyZGluZyB0YWJsZS48QlI+Jmd0 OzxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5Ib3cgZG9lcyB0 aGUgcmVkdW5kYW5jeSB3b3JrIGluIGhlcmUuIEFyZSB5b3UgdXNpbmcgYnJpZGdlIHByb3RvY29s IGluPEJSPmNvbmp1bmN0aW9uIHdpdGggR01QTFMgcHJvdG9jb2wuIElmIHlvdSBkbywgdGhlbiB0 aGVyZSBhcmUgYnVuY2ggb2Y8QlI+aXNzdWVzLiBJZiB5b3UgZG9uJ3QsIHRoZW4gd2hhdCBpcyBp dCB1c2VkIGFuZCBob3cgZG9lcyBpdCBpbnRlcmFjdCB3aXRoPEJSPmJyaWRnaW5nIHByb3RvY29s cyA/PzxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj4tQWxpPEJS PjwvRk9OVD48QlI+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj4mZ3Q7PEJSPiZn dDsgVGhhbmtzPEJSPiZndDs8QlI+Jmd0OyBKYWloeXVuZzxCUj4mZ3Q7PEJSPiZndDs8QlI+Jmd0 OzxCUj4mZ3Q7IC0tLS0tPz8gPz8/LS0tLS08QlI+Jmd0OyA/PyA/PzogJnF1b3Q7SGVpbGVzIEp1 ZXJnZW4mcXVvdDsgJmx0O2p1ZXJnZW4uaGVpbGVzQHNpZW1lbnMuY29tJmd0OzxCUj4mZ3Q7ID8/ ID8/OiAyMDA1LTA3LTI1ID8/IDY6NDQ6MzQ8QlI+Jmd0OyA/PyA/PzogJnF1b3Q7Q0hPLCBKQUkg SFlVTkcmcXVvdDsgJmx0O2phaWh5dW5nQGV0cmkucmUua3ImZ3Q7LCAmcXVvdDtIZWlsZXM8QlI+ Jmd0OyBKdWVyZ2VuJnF1b3Q7ICZsdDtqdWVyZ2VuLmhlaWxlc0BzaWVtZW5zLmNvbSZndDssICZx dW90O3BlckBkZWZlcm8uc2UmcXVvdDs8QlI+Jmd0OyAmbHQ7cGVyQGRlZmVyby5zZSZndDs8QlI+ Jmd0OyA/PzogJnF1b3Q7Y2NhbXBAb3BzLmlldGYub3JnJnF1b3Q7ICZsdDtjY2FtcEBvcHMuaWV0 Zi5vcmcmZ3Q7PEJSPiZndDsgPz86IFJFOiBGcmFtZWZvcm1hdCBpbiBhIGwyY3MgZ21wbHMgcm52 aXJvbm1lbnQuPEJSPiZndDs8QlI+Jmd0OzxCUj4mZ3Q7PEJSPiZndDs8QlI+Jmd0OzxCUj4mZ3Q7 PEJSPiZndDsgRGVhciBKYWloeXVuZyw8QlI+Jmd0OzxCUj4mZ3Q7IHlvdSBtZW50aW9uIHNldmVy YWwgb3B0aW9ucyB0aGF0IGNvdWxkIGJlIHVzZWQgYXMgYSBsYWJlbC48QlI+Jmd0OyBNb3N0IG9m IHRoZW0gaGF2ZSBzdHJvbmcgaW1wYWN0cyBvbiB0aGUgZGF0YSBwbGFuZSBhbmQ8QlI+Jmd0OyBp bnRyb2R1Y2UgbmV3IGxheWVyIG5ldHdvcmtzLiBJcyB0aGlzIGluIHRoZSBzY29wZSBvZiBDQ0FN UD88QlI+Jmd0OyBXaGF0IGlzIHRoZSBnb2FsLCB0byBpbnRyb2R1Y2UgYSBjb250cm9sIHBsYW5l IGZvciBhIGV4aXN0aW5nPEJSPiZndDsgTDIgdGVjaG5vbG9neSBvciBpbnRyb2R1Y2UgYSBuZXcg TDIgdGVjaG5vbG9neT8gV2hhdCB3b3VsZCBiZTxCUj4mZ3Q7IHRoZSBiZW5lZml0IG9mIHN1Y2gg YSBuZXcgTDIgY28gc3dpdGNoaW5nIHRlY2hub2xvZ3kgY29tcGFyZWQ8QlI+Jmd0OyB0byBFdGhl cm5ldCBvdmVyIE1QTFMgYXMgZGVmaW5lZCBieSBQV0UzIGFuZCBMMlZQTj88QlI+Jmd0OyBZb3Ug bWVudGlvbmVkIHRvIHVzZSB0aGUgTUFDIGFkZHJlc3Mgb3IgcGFydCBvZiBpdCBhcyBhPEJSPiZn dDsgbGFiZWwuIFRoZSBUUklMTCB3b3JraW5nIGdyb3VwIGlzIGRlZmluaW5nIHNob3J0ZXN0IHBh dGg8QlI+Jmd0OyByb3V0aW5nIGZvciBFdGhlcm5ldC4gSG93IHdvdWxkIHRoaXMgZml0IHRvZ2V0 aGVyPzxCUj4mZ3Q7PEJSPiZndDs8QlI+Jmd0OyBSZWdhcmRzPEJSPiZndDs8QlI+Jmd0OyBKdWVy Z2VuPEJSPiZndDs8QlI+Jmd0OzxCUj4mZ3Q7PEJSPiZndDs8QlI+Jmd0OzxCUj4mZ3Q7ICZndDsg LS0tLS1PcmlnaW5hbCBNZXNzYWdlLS0tLS08QlI+Jmd0OyAmZ3Q7IEZyb206IENITywgSkFJIEhZ VU5HIFttYWlsdG86amFpaHl1bmdAZXRyaS5yZS5rcl08QlI+Jmd0OyAmZ3Q7IFNlbnQ6IFNhdHVy ZGF5LCBKdWx5IDIzLCAyMDA1IDY6MTkgQU08QlI+Jmd0OyAmZ3Q7IFRvOiBIZWlsZXMgSnVlcmdl bjsgcGVyQGRlZmVyby5zZTxCUj4mZ3Q7ICZndDsgQ2M6IGNjYW1wQG9wcy5pZXRmLm9yZzxCUj4m Z3Q7ICZndDsgU3ViamVjdDogUkU6IEZyYW1lZm9ybWF0IGluIGEgbDJjcyBnbXBscyBybnZpcm9u bWVudC48QlI+Jmd0OyAmZ3Q7PEJSPiZndDsgJmd0OzxCUj4mZ3Q7ICZndDsgRGVhciBKdWVyZ2Vu LCBQYXIgYW5kIGFsbCw8QlI+Jmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyBUaGUgcHJvcG9zZWQgZnJh bWV3b3JrIG9mIEwyU0Mgd2FzIG5vdCBpbnRlbmRlZCB0byBzdWdnZXN0PEJSPiZndDsgJmd0OyBh bnkgc3BlY2lmaWMgc29sdXRpb24uIEl0IGlzIHN0aWxsIGFuIG9wZW4gcXVlc3Rpb24uPEJSPiZn dDsgJmd0OyBBcyBBZHJpYW4gbm90ZWQsIGl0IG9ubHkgYWltcyB0byBwcmVzZW50IHJlcXVpcmVt ZW50czxCUj4mZ3Q7ICZndDsgYXMgaW4gc2NlbmFyaW9zIHRoYXQgbWF5IGRyaXZlIHBlb3BsZSB0 b3dhcmQgbmV4dCBzdGVwIGluIG1hbnk8QlI+Jmd0OyAmZ3Q7IHZpYWJsZSBzb2x1dGlvbnMuPEJS PiZndDsgJmd0OzxCUj4mZ3Q7ICZndDsgSW4gZmFjdCwgdGhlcmUgYXJlIG90aGVyIG9wdGlvbnMg d2UgbWF5IGFsc28gY29uc2lkZXIgZm9yIEwyPEJSPiZndDsgJmd0OyBsYWJlbCBlbmNvZGluZy48 QlI+Jmd0OyAmZ3Q7IElFRUUgODAyLjEgRXRoZXJuZXQgYnJpZGdlcyBmb3J3YXJkIGZyYW1lcyBi YXNlZCBvbiA0OGJpdHMgb2Y8QlI+Jmd0OyAmZ3Q7IE1BQyBhZGRyZXNzLDxCUj4mZ3Q7ICZndDsg YW5kIGFkZGl0aW9uYWxseSB1c2luZyBWTEFOIHRhZy48QlI+Jmd0OyAmZ3Q7PEJSPiZndDsgJmd0 OyBXaGVuIHRoZSBwdXJwb3NlIG9mIEdNUExTIGNvbnRyb2wgb3ZlciBFdGhlcm5ldCBpcyBub3Qg dG88QlI+Jmd0OyAmZ3Q7IGNyZWF0ZSBuZXcgZGF0YXBsYW5lPEJSPiZndDsgJmd0OyBidXQgdG8g dXRpbGl6ZSBJRUVFIDgwMi4xIGJyaWRnZSBhcmNoaXRlY3R1cmUsIHdlIG1heTxCUj4mZ3Q7IGNv bnNpZGVyIHVzaW5nPEJSPiZndDsgJmd0OyBvbmUgb2YgdGhlIHR3byBmb3J3YXJkaW5nIG1ldGhv ZHMsIE1BQyBmb3J3YXJkaW5nIG9yIFZMQU48QlI+Jmd0OyBmb3J3YXJkaW5nLjxCUj4mZ3Q7ICZn dDsgKGFuZCBwZXJoYXBzIGFueSBvdGhlciBjb21iaW5hdGlvbiBvZiBmaWVsZHMgaW4gTUFDLCBi dXQgSSdsbDxCUj4mZ3Q7ICZndDsgbm90IGRpc2N1c3MgaXQpPEJSPiZndDsgJmd0OzxCUj4mZ3Q7 ICZndDsgVXNlIG9mIFZMQU4gSUQgZm9yIGxhYmVsIGVuY29kaW5nIG1heSBhdXRvbWF0ZSBWTEFO IGNvbmZpZ3VyYXRpb248QlI+Jmd0OyAmZ3Q7IHVzaW5nIElQIHByb3RvY29scy4gSG93ZXZlciwg R01QTFMgcHJvdG9jb2wgY2Fubm90IHVzZSB0aGU8QlI+Jmd0OyAmZ3Q7IGZpZWxkIGV4Y2x1c2l2 ZWx5PEJSPiZndDsgJmd0OyBiZWNhdXNlIHB1YmxpYy9wcml2YXRlIG9wZXJhdG9ycyBhbHJlYWR5 IHVzZSBWTEFOIGZvciB2YXJpb3VzPEJSPiZndDsgJmd0OyBwdXJwb3NlLjxCUj4mZ3Q7ICZndDsg VGhlcmUgaXMgYSBwb3RlbnRpYWwgY29uZmxpY3Qgd2l0aCBleGlzdGluZyB1c2Ugb2YgVkxBTiBh bmQ8QlI+Jmd0OyAmZ3Q7IEdNUExTIHVzZSBvZiBWTEFOIGxhYmVsLjxCUj4mZ3Q7ICZndDsgRnVy dGhlcm1vcmUsIHNjYWxhYmlsaXR5IG9mIFZMQU4gSUQgaGFzIGJlZW4gZnJlcXVlbnRseSBub3Rl ZDxCUj4mZ3Q7ICZndDsgYXMgd2Vha25lc3M8QlI+Jmd0OyAmZ3Q7IGJlY2F1c2UgdGhlIHNpemUg b2YgVkxBTiBJRCBpcyBhdCBtb3N0IDQwOTYgKDEyYml0KS48QlI+Jmd0OyAmZ3Q7PEJSPiZndDsg Jmd0OyBUaGUgc2NhbGFiaWxpdHkgbWF5IGJlIGltcHJvdmVkIGlmIHRoZSBzY29wZSBvZiBWTEFO IGxhYmVsIGlzPEJSPiZndDsgJmd0OyBjb25maW5lZCB0bzxCUj4mZ3Q7ICZndDsgbGluay1sb2Nh bCwgYW5kIHNvbWUgYWRkaXRpb25hbCBzd2FwcGluZyBmdW5jdGlvbiBvZiBWTEFOIElEPEJSPiZn dDsgJmd0OyBpcyBpbnRyb2R1Y2VkIGluPEJSPiZndDsgJmd0OyBJbnRlcm5hbCBTdWJsYXllciBb ODAyLjFEXSBvZiBHTVBMUyBpbXBsZW1lbnRlZCBzd2l0Y2guPEJSPiZndDsgJmd0OyBIb3dldmVy LCB0aGlzIHdpbGwgb25seSBiZSBlZmZlY3RpdmUgd2hlbiB0aGUgY29uZmlndXJhdGlvbjxCUj4m Z3Q7ICZndDsgb2YgbmV0d29yazxCUj4mZ3Q7ICZndDsgaXMgbWVzaCBzdHJ1Y3R1cmUgdGhhdCBt dWx0aXBsZSBMU1AgcGF0aHMgZXhpc3QuIElmIHRoZTxCUj4mZ3Q7ICZndDsgY29uZmlndXJhdGlv biBvZiBuZXR3b3JrPEJSPiZndDsgJmd0OyBpcyBzdGFyIG9yIHRyZWUgc2hhcGUsIGFzIG5vcm1h bCBjb25maWd1cmF0aW9uIG9mIGFjY2VzcyBuZXR3b3JrLDxCUj4mZ3Q7ICZndDsgTFNQcyBjb25j ZW50cmF0ZSBpbiByb290IG5vZGUgYW5kIHRvdGFsIG51bWJlciBvZjxCUj4mZ3Q7ICZndDsgTFNQ cyB0aGF0IHRoZSBuZXR3b3JrIGNhbiBob2xkIHN0aWxsIGJlIGxpbWl0ZWQgYnkgYXZhaWxhYmxl PEJSPiZndDsgJmd0OyBsYWJlbCBzcGFjZTxCUj4mZ3Q7ICZndDsgYXQgYSBmZXcgcm9vdCBsaW5r cy48QlI+Jmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyBGb3IgdGhlc2UgcmVhc29ucywgSSBkbyBub3Qg dGhpbmsgYW55IGZvcm0gb2YgVkxBTiBJRCBsYWJlbCBpczxCUj4mZ3Q7ICZndDsgYW4gYXBwcm9w cmlhdGUgY2hvaWNlIGZvciBsYXllci0yIGxhYmVsIGVuY29kaW5nLjxCUj4mZ3Q7ICZndDs8QlI+ Jmd0OyAmZ3Q7IFRoZXJlIGFyZSBzb21lIG90aGVyIHByb3Bvc2FscyBhc3N1bWluZyBuZXcgYXNz aWdubWVudCBvZiBFdGhlcm5ldDxCUj4mZ3Q7ICZndDsgTGVuZ3RoL1R5cGUgdmFsdWUgKGUuZy4g bmV3IFRQSUQgaW4gVkxBTiB0YWcpIGFuZCByZS1kZWZpbml0aW9uIG9mPEJSPiZndDsgJmd0OyBp bmZvcm1hdGlvbiBmaWVsZHMgcGxhY2VkIGJldHdlZW4gODAyLjMgTUFDIGhlYWRlciBhbmQgSVAg cGFja2V0LjxCUj4mZ3Q7ICZndDsgSW4gdGhpcyBjYXNlLCBvbmx5IHRoZSBmb3JtYXQgb2YgVkxB TiB0YWcgb3IgZXh0ZW5kZWQgVkxBTiB0YWc8QlI+Jmd0OyAmZ3Q7IGlzIGJvcnJvd2VkLCBob3dl dmVyLCBpbnNpZGUgdGhlIHN3aXRjaGluZyBoYXJkd2FyZSwgdGhlPEJSPiZndDsgJmd0OyBmaWx0 ZXIgYW5kIHJlbGF5LDxCUj4mZ3Q7ICZndDsgZXRjLiBhcmUgdG90YWxseSBkaWZmZXJlbnQgbmV3 IGRhdGFwbGFuZSBzd2l0Y2guPEJSPiZndDsgJmd0OzxCUj4mZ3Q7ICZndDsgSSBkbyBub3Qgc2Vl IHN1Y2ggYXBwcm9hY2ggaXMgYSBHTVBMUyBpbXBsZW1lbnRhdGlvbiBmb3IgRXRoZXJuZXQ8QlI+ Jmd0OyAmZ3Q7IGJlY2F1c2UgdGhlIGNvcmUgc3dpdGNoaW5nIHRlY2huaXF1ZSBpcyBub3QgODAy LjEgRXRoZXJuZXQ8QlI+Jmd0OyAmZ3Q7IGJyaWRnZSBhdCBhbGwuPEJSPiZndDsgJmd0OzxCUj4m Z3Q7ICZndDsgVGhlIG90aGVyIG9wdGlvbiB3ZSBtYXkgY29uc2lkZXIgaXMgdXNpbmcgTUFDIGFk ZHJlc3MgZmlsZWQ8QlI+Jmd0OyBhcyBiZWxvdy48QlI+Jmd0OyAmZ3Q7PEJSPiZndDsgJmd0OzxC Uj4mZ3Q7ICZndDsgKy0tLS0tLS0rLS0tLS0tLSstLS0tLS0tKy0tLS0tLS0rLS0tLS0tLSstLS0t LS0tKzxCUj4mZ3Q7ICZndDsgfCAxYnl0ZSB8IDJieXRlIHwgM2J5dGUgfCA0Ynl0ZSB8IDVieXRl IHwgNmJ5dGUgfDxCUj4mZ3Q7ICZndDsgKy0tLS0tLS0rLS0tLS0tLSstLS0tLS0tKy0tLS0tLS0r LS0tLS0tLSstLS0tLS0tKzxCUj4mZ3Q7ICZndDsgKy0tLS0tLS0tLS0tLS0tLS0tLS0tLS0tKy0t LS0tLS0tLS0tLS0tLS0tLS0tLS0tKzxCUj4mZ3Q7ICZndDsgfCAmbmJzcDtPVUkgUHJlZml4ICg9 R01QTFMpICZuYnNwO3wgJm5ic3A7IERBLWxhYmVsICgyNGJpdCkgJm5ic3A7ICZuYnNwO3w8QlI+ Jmd0OyAmZ3Q7ICstLS0tLS0tLS0tLS0tLS0tLS0tLS0tLSstLS0tLS0tLS0tLS0tLS0tLS0tLS0t LSs8QlI+Jmd0OyAmZ3Q7IHwgJm5ic3A7T1VJIFByZWZpeCAoPUdNUExTKSAmbmJzcDt8ICZuYnNw OyBTQS1sYWJlbCAoMjRiaXQpICZuYnNwOyAmbmJzcDt8PEJSPiZndDsgJmd0OyArLS0tLS0tLS0t LS0tLS0tKy0tLS0tLS0rLS0tLS0tLS0tLS0tLS0tLS0tLS0tLS0rPEJSPiZndDsgJmd0OyB8IExl bmd0aC9UeXBlICZuYnNwOyB8PEJSPiZndDsgJmd0OyArLS0tLS0tLS0tLS0tLS0tKzxCUj4mZ3Q7 ICZndDs8QlI+Jmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyBJRUVFIGlzIGRlc2lnbmF0ZWQgYnkgdGhl IElTTyBDb3VuY2lsIHRvIGFjdCBhcyB0aGU8QlI+Jmd0OyAmZ3Q7IHJlZ2lzdHJhdGlvbiBhdXRo b3JpdHk8QlI+Jmd0OyAmZ3Q7IGZvciB0aGUgaGlnaGVyIHRocmVlLW9jdGV0IG9mIE9VSSBudW1i ZXIgaW4gdGhlIE1BQyBhZGRyZXNzPEJSPiZndDsgJmd0OyB0byBiZSB1c2VkIGJ5IG1hbnVmYWN0 dXJlci4gRXRoZXJuZXQgbWFudWZhY3R1cmVyIG1heSBnZW5lcmF0ZTxCUj4mZ3Q7ICZndDsgZ2xv YmFsIHVuaXF1ZSBNQUMgYWRkcmVzcyB1c2luZyB0aGUgT1VJIHByZWZpeCBhbmQgYWRkcmVzcyBi bG9jayBvZjxCUj4mZ3Q7ICZndDsgbG93ZXIgdGhyZWUtb2N0ZXQgKDI0Yml0KS4gVGFraW5nIGFk dmFudGFnZSBvZiB0aGU8QlI+Jmd0OyBhZGRyZXNzaW5nIHNjaGVtZSw8QlI+Jmd0OyAmZ3Q7IEdN UExTIG1heSB1c2UgdGhlIGxvd2VyIHRocmVlLW9jdGV0IGV4Y2x1c2l2ZWx5IGlmIGEgdW5pcXVl PEJSPiZndDsgJmd0OyBPVUkgbnVtYmVyPEJSPiZndDsgJmd0OyBpcyByZXNlcnZlZCBmb3IgdGhl IHByb3RvY29sLiBXaXRoIHRoaXMgbGFiZWxpbmcgc2NoZW1lLCBHTVBMUyB3aWxsPEJSPiZndDsg Jmd0OyBjb250cm9sIE1BQyBmb3J3YXJkaW5nIGVudHJ5LCBub3QgVkxBTiB0YWJsZS48QlI+Jmd0 OyAmZ3Q7PEJSPiZndDsgJmd0OyBBbGwgRXRoZXJuZXQgZnJhbWVzIGNvbnRyb2xsZWQgYnkgR01Q TFMgd2lsbCBoYXZlIGlkZW50aWNhbDxCUj4mZ3Q7ICZndDsgT1VJIG51bWJlcjxCUj4mZ3Q7ICZn dDsgdGhhdCB0aGV5IGNhbiBlYXNpbHkgYmUgZGlzdGluZ3Vpc2hlZCBmcm9tIG90aGVyIEV0aGVy bmV0IGZyYW1lcy48QlI+Jmd0OyAmZ3Q7IEluIHByaW5jaXBsZSwgdGhlIGxhYmVsIGxvb2t1cCBo YXJkd2FyZSBpcyBpZGVudGljYWwgdG8gTUFDIGxvb2t1cDxCUj4mZ3Q7ICZndDsgaGFyZHdhcmUg aW4gdGhpcyBsYWJlbGluZyBzY2hlbWUuIFRoZXJlZm9yZSBHTVBMUyBpbXBsZW1lbnRlZDxCUj4m Z3Q7ICZndDsgc3dpdGNoIG1heSBzdGlsbCBmdW5jdGlvbiBhcyBub3JtYWwgRXRoZXJuZXQgYnJp ZGdlIHRvIHRoZSBmcmFtZXM8QlI+Jmd0OyAmZ3Q7IHRoYXQgT1VJIG51bWJlciBpcyBub3QgR01Q TFMuIFRoaXMgYWxzbyBmYWNpbGl0YXRlcyBHTVBMUzxCUj4mZ3Q7IGltcGxlbWVudGVkPEJSPiZn dDsgJmd0OyBzd2l0Y2hlcyBiZWluZyBkZXBsb3llZCBpbiBvcGVyYXRpbmcgRXRoZXJuZXQgd2l0 aCBtaW5pbXVtPEJSPiZndDsgJmd0OyBzZXJ2aWNlIGRpc3J1cHRpb24uPEJSPiZndDsgJmd0OzxC Uj4mZ3Q7ICZndDsgTm90ZSBhbHNvIHRoYXQgYWJvdmUgcHJvcG9zZWQgbGFiZWwgZW5jb2Rpbmcg bWV0aG9kIGlzIHRyYW5zcGFyZW50PEJSPiZndDsgJmd0OyB0byB0aGUgdXNlIG9mIEV0aGVybmV0 IExlbmd0aC9UeXBlIGZpZWxkLiBFbmQtdXNlciBkZXZpY2UgbWF5IHVzZTxCUj4mZ3Q7ICZndDsg dGhlIExlbmd0aC9UeXBlIGZpZWxkIGFzIGRlZmluZWQgaW4gSUVFRSA4MDIuMyBwcm90b2NvbC48 QlI+Jmd0OyAmZ3Q7IEl0IGFsc28gYWxsb3dzIG5ldHdvcmsgb3BlcmF0b3JzIGNvbmZpZ3VyZSBW TEFOIGZvciB0aGVpciBvd248QlI+Jmd0OyAmZ3Q7IHB1cnBvc2UuPEJSPiZndDsgJmd0OyBXaGVu IElFRUUgODAyLjFwIGlzIHVzZWQgaW4gY29uanVuY3Rpb24gd2l0aCBMMi1MU1AsIHRoZSBwcmlv cml0eTxCUj4mZ3Q7ICZndDsgZmllbGQgb2YgVkxBTiB0YWcgY2FuIGFsc28gYmUgdXNlZCBmb3Ig aW1wb3NpbmcgY29uc2lzdGVudCBURTxCUj4mZ3Q7ICZndDsgcG9saWN5IGluPEJSPiZndDsgJmd0 OyBsZWdhY3kgc3dpdGNoZXMgYW5kIEdNUExTIHN3aXRjaGVzLjxCUj4mZ3Q7ICZndDs8QlI+Jmd0 OyAmZ3Q7IEFueSB3YXksIG15IGNvbmNsdXNpb24gaXMsIHRoZXJlIGFyZSBvdGhlciBvcHRpb25z IHdlIG1heSBjb25zaWRlciw8QlI+Jmd0OyAmZ3Q7IGFuZCB0aGlzIGlzc3VlIGlzIHN0aWxsIG9w ZW4gdG8gZGlzY3Vzcy48QlI+Jmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyBUaGFua3MsPEJSPiZndDsg Jmd0OzxCUj4mZ3Q7ICZndDsgU2luY2VyZWx5LDxCUj4mZ3Q7ICZndDs8QlI+Jmd0OyAmZ3Q7IEph aWh5dW5nPEJSPiZndDsgJmd0OzxCUj4mZ3Q7ICZndDs8QlI+Jmd0OyAmZ3Q7PEJSPiZndDsgJmd0 OzxCUj4mZ3Q7ICZndDs8QlI+Jmd0OyAmZ3Q7IC0tLS0tPz8gPz8/LS0tLS08QlI+Jmd0OyAmZ3Q7 ID8/ID8/OiAmcXVvdDtIZWlsZXMgSnVlcmdlbiZxdW90OyAmbHQ7anVlcmdlbi5oZWlsZXNAc2ll bWVucy5jb20mZ3Q7PEJSPiZndDsgJmd0OyA/PyA/PzogMjAwNS0wNy0yMiA/PyAxMDoyMzo0MzxC Uj4mZ3Q7ICZndDsgPz8gPz86ICZxdW90O0xvYSBBbmRlcnNzb24mcXVvdDsgJmx0O2xvYUBwaS5z ZSZndDssICZxdW90O3JpY2hhcmQuc3BlbmNlckBidC5jb20mcXVvdDs8QlI+Jmd0OyAmZ3Q7ICZs dDtyaWNoYXJkLnNwZW5jZXJAYnQuY29tJmd0OzxCUj4mZ3Q7ICZndDsgPz86ICZxdW90O3BlckBk ZWZlcm8uc2UmcXVvdDsgJmx0O3BlckBkZWZlcm8uc2UmZ3Q7LCAmcXVvdDtjY2FtcEBvcHMuaWV0 Zi5vcmcmcXVvdDs8QlI+Jmd0OyAmZ3Q7ICZsdDtjY2FtcEBvcHMuaWV0Zi5vcmcmZ3Q7PEJSPiZn dDsgJmd0OyA/PzogUkU6IEZyYW1lZm9ybWF0IGluIGEgbDJjcyBnbXBscyBybnZpcm9ubWVudC48 QlI+Jmd0OyAmZ3Q7PEJSPiZndDsgJmd0OzxCUj4mZ3Q7ICZndDs8QlI+Jmd0OyAmZ3Q7PEJSPiZn dDsgJmd0OzxCUj4mZ3Q7ICZndDs8QlI+Jmd0OyAmZ3Q7IExvYSw8QlI+Jmd0OyAmZ3Q7PEJSPiZn dDsgJmd0OyBJIGludGVycHJldCB0aGUgSUQgYXMgYSBwcm9wb3NhbCB0byB1c2UgR01QTFMgZm9y IFZMQU4gc2V0dXAuPEJSPiZndDsgJmd0OyBTbyBHTVBMUyBhbmQgVkxBTiB0byBub3QgY29tcGV0 ZS4gVGhlIFZMQU4gaXMgYXQgdGhlIGRhdGE8QlI+Jmd0OyAmZ3Q7IHBsYW5lIGFuZCBHTVBMUyBh dCB0aGUgY29udHJvbCBwbGFuZS4gVGhlIHF1ZXN0aW9uIGlzIGhvdyBhbmQ8QlI+Jmd0OyAmZ3Q7 IHNob3VsZCBkaWZmZXJlbnQgY29udHJvbCBwbGFuZSB0ZWNobmlxdWVzIGxpa2UgR01QTFMgYW5k PEJSPiZndDsgJmd0OyAoR1ZSUCBhbmQgU1RCKSB3b3JrIHRvZ2V0aGVyPzxCUj4mZ3Q7ICZndDs8 QlI+Jmd0OyAmZ3Q7IFJlZ2FyZHM8QlI+Jmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyBKdWVyZ2VuPEJS PiZndDsgJmd0OzxCUj4mZ3Q7ICZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsgLS0tLS1PcmlnaW5hbCBN ZXNzYWdlLS0tLS08QlI+Jmd0OyAmZ3Q7ICZndDsgRnJvbTogb3duZXItY2NhbXBAb3BzLmlldGYu b3JnPEJSPiZndDsgJmd0OyAmZ3Q7IFttYWlsdG86b3duZXItY2NhbXBAb3BzLmlldGYub3JnXSBP biBCZWhhbGYgT2YgTG9hIEFuZGVyc3NvbjxCUj4mZ3Q7ICZndDsgJmd0OyBTZW50OiBGcmlkYXks IEp1bHkgMjIsIDIwMDUgMjo1MyBQTTxCUj4mZ3Q7ICZndDsgJmd0OyBUbzogcmljaGFyZC5zcGVu Y2VyQGJ0LmNvbTxCUj4mZ3Q7ICZndDsgJmd0OyBDYzogcGVyQGRlZmVyby5zZTsgY2NhbXBAb3Bz LmlldGYub3JnPEJSPiZndDsgJmd0OyAmZ3Q7IFN1YmplY3Q6IFJlOiBGcmFtZWZvcm1hdCBpbiBh IGwyY3MgZ21wbHMgcm52aXJvbm1lbnQuPEJSPiZndDsgJmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyAm Z3Q7IFJpY2hhcmQsPEJSPiZndDsgJmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyAmZ3Q7IEkgYWdyZWUg dG8gbW9zdCBvZiB0aGlzLiBTaW5jZSB3ZSBoYXZlIGRvdWJ0cyBhYm91dCB0aGUgdmlhYmlsaXR5 PEJSPiZndDsgJmd0OyAmZ3Q7IG9mIHRha2luZyBHTVBMUyBhbGwgdGhlIHdheSB0byBlbmQtdXNl ciBvciBlbnRlcnByaXNlIEkgdGhpbmsgaXQ8QlI+Jmd0OyAmZ3Q7ICZndDsgd291bGQgYmUgZ29v ZCwgZnJvbSBhIHdnIHBlcnNwZWN0aXZlLCBpZiB3ZSBhZ3JlZWQgdG8gc29sdmUgdGhlPEJSPiZn dDsgJmd0OyAmZ3Q7IGNvcmUgbmV0d29yayBwcm9ibGVtcyBmaXJzdC48QlI+Jmd0OyAmZ3Q7ICZn dDs8QlI+Jmd0OyAmZ3Q7ICZndDsgRG8geW91IGhhdmUgYW55IGNvbW1lbnQgb24gdGhlIHJlcXVp cmVtZW50IHRvIHJ1biBib3RoIFZMQU5zIGFuZDxCUj4mZ3Q7ICZndDsgJmd0OyBHTVBMUyBvbiB0 aGUgc2FtZSBzd2l0Y2g/PEJSPiZndDsgJmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyAmZ3Q7IC9Mb2E8 QlI+Jmd0OyAmZ3Q7ICZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsgcmljaGFyZC5zcGVuY2VyQGJ0LmNv bSB3cm90ZTo8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyBSZWdhcmRsZXNzIG9mIHdoZXRoZXIgb3Ig bm90IGEgc3dpdGNoIGlzIGRpcmVjdGx5IGNvbm5lY3RlZDxCUj4mZ3Q7ICZndDsgJmd0OyAmbmJz cDsmZ3Q7IHRvIGhvc3RzLCBpdCBtdXN0IGJlIGFibGUgdG8gZm9yd2FyZCBwYWNrZXRzIHVzaW5n IHRoZTxCUj4mZ3Q7ICZndDsgJmd0OyBjb25uZWN0aW9ubGVzczxCUj4mZ3Q7ICZndDsgJmd0OyAm bmJzcDsmZ3Q7IEV0aGVybmV0IGRhdGEgcGxhbmUuIFRoaXMgaXMgZHVlIHRvIHRoZSBmdW5kYW1l bnRhbDxCUj4mZ3Q7ICZndDsgJmd0OyByZXF1aXJlbWVudCB0aGF0PEJSPiZndDsgJmd0OyAmZ3Q7 ICZuYnNwOyZndDsgZm9yIEdNUExTIHN3aXRjaGVzIHRvIGJlIGFibGUgdG8gZXhjaGFuZ2UgY29u dHJvbDxCUj4mZ3Q7ICZndDsgaW5mb3JtYXRpb24gd2l0aDxCUj4mZ3Q7ICZndDsgJmd0OyAmbmJz cDsmZ3Q7IGVhY2ggb3RoZXIsIGEgZGF0YSBwbGFuZSBmb3IgY29udHJvbCB0cmFmZmljIG11c3Qg YmUgcHJlc2VudC48QlI+Jmd0OyAmZ3Q7ICZndDsgJm5ic3A7Jmd0OyBUaGlzIGlzIGFraW4gdG8g dXNpbmcgdGhlIElQIGRhdGEgcGxhbmUgZm9yIE1QTFM8QlI+Jmd0OyAmZ3Q7IHNpZ25hbGxpbmcg aW4gYW48QlI+Jmd0OyAmZ3Q7ICZndDsgJm5ic3A7Jmd0OyBJUC9NUExTIG5ldHdvcmsuIEFuIGFs dGVybmF0aXZlIHdvdWxkIGJlIHRvIHVzZSBhIHN0YXRpYzxCUj4mZ3Q7ICZndDsgcmVzZXJ2ZWQ8 QlI+Jmd0OyAmZ3Q7ICZndDsgJm5ic3A7Jmd0OyBMMi1MU1AgZm9yIGNvbnRyb2wgdHJhZmZpYyBp biB0aGUgc2FtZSB3YXkgdGhhdCByZXNlcnZlZDxCUj4mZ3Q7ICZndDsgVlBJL1ZDSXM8QlI+Jmd0 OyAmZ3Q7ICZndDsgJm5ic3A7Jmd0OyBhcmUgdXNlZCBmb3IgUE5OSSBzaWduYWxsaW5nIGluIEFU TS48QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OzxCUj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7IFJlZ2FyZGlu ZyBjb25uZWN0aW5nIGhvc3RzIHRvIEdNUExTIHN3aXRjaGVzLCBJPEJSPiZndDsgcGVyc29uYWxs eSBkb24ndDxCUj4mZ3Q7ICZndDsgJmd0OyAmbmJzcDsmZ3Q7IHRoaW5rIGV4dGVuZGluZyBMMi1M U1BzIGludG8gdGhlIGVudGVycHJpc2UvaG9tZSBuZXR3b3JrIGlzPEJSPiZndDsgJmd0OyAmZ3Q7 ICZuYnNwOyZndDsgY29tbWVyY2lhbGx5IHZpYWJsZS4gSG93ZXZlciwgaWYgeW91IGRvIHdhbnQg dG8gdXNlPEJSPiZndDsgJmd0OyBHTVBMUyBzd2l0Y2hlczxCUj4mZ3Q7ICZndDsgJmd0OyAmbmJz cDsmZ3Q7IGluIHRoZSBob21lL2VudGVycHJpc2UgbmV0d29yayBhbmQgZm9yIHNvbWUgcmVhc29u PEJSPiZndDsgZG9uJ3Qgd2FudCB0bzxCUj4mZ3Q7ICZndDsgJmd0OyAmbmJzcDsmZ3Q7IGV4dGVu ZCBMMi1MU1BzIGRvd24gdG8gdGhlIGhvc3QgdGhlbiB5b3Ugd2lsbCBub3QgYmU8QlI+Jmd0OyBw ZXJmb3JtaW5nPEJSPiZndDsgJmd0OyAmZ3Q7ICZuYnNwOyZndDsgbm9ybWFsIEV0aGVybmV0IE1h YyBhZGRyZXNzIHN3aXRjaGluZyBhbnl3YXkuIEluc3RlYWQgeW91IHdpbGw8QlI+Jmd0OyAmZ3Q7 ICZndDsgJm5ic3A7Jmd0OyBuZWVkIHNvbWUga2luZCBvZiBwb2xpY3kgb24gdGhlIHN3aXRjaCB0 aGF0IG1hcHM8QlI+Jmd0OyBjb25uZWN0aW9ubGVzczwvRk9OVD48QlI+PEZPTlQgRkFDRT0iTW9u b3NwYWNlLENvdXJpZXIiPiZndDsgJmd0OyAmZ3Q7ICZuYnNwOyZndDsgRXRoZXJuZXQgcGFja2V0 cyAoZS5nLiBiYXNlZCBvbiBNQUMgc3JjL2Rlc3QsIDgwMi4xcCwgPC9GT05UPjxCUj48Rk9OVCBG QUNFPSJNb25vc3BhY2UsQ291cmllciI+Jmd0OyBWTEFOKSB0byBhPEJSPiZndDsgJmd0OyAmZ3Q7 ICZuYnNwOyZndDsgTDItTFNQLiBUaGlzIGlzIGJlY2F1c2UgbXVsdGlwbGUgTDItTFNQcyB0byB0 aGUgc2FtZTxCUj4mZ3Q7IGRlc3RpbmF0aW9uPEJSPiZndDsgJmd0OyAmZ3Q7ICZuYnNwOyZndDsg KGUuZy4gYSBnYXRld2F5IHJvdXRlcikgbWF5IGV4aXN0IGZvciBkaWZmZXJlbnQgc2VydmljZXMv Zmxvd3M8QlI+Jmd0OyAmZ3Q7ICZndDsgJm5ic3A7Jmd0OyAoZS5nLiB2aWRlbyBkb3dubG9hZCwg Vm9JUCBjYWxsLCBldGMuKS48QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OzxCUj4mZ3Q7ICZndDsgJmd0 OyAmZ3Q7IFJlZ2FyZHMsPEJSPiZndDsgJmd0OyAmZ3Q7ICZndDsgUmljaGFyZDxCUj4mZ3Q7ICZn dDsgJmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyAmZ3Q7ICZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0 OyZndDstLS0tLU9yaWdpbmFsIE1lc3NhZ2UtLS0tLTxCUj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0 O0Zyb206IG93bmVyLWNjYW1wQG9wcy5pZXRmLm9yZzxCUj4mZ3Q7ICZndDsgW21haWx0bzpvd25l ci1jY2FtcEBvcHMuaWV0Zi5vcmddT248QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDtCZWhhbGYg T2YgUGFyIE1hdHRzc29uPEJSPiZndDsgJmd0OyAmZ3Q7ICZndDsmZ3Q7U2VudDogMjIgSnVseSAy MDA1IDExOjQyPEJSPiZndDsgJmd0OyAmZ3Q7ICZndDsmZ3Q7VG86IExvYSBBbmRlcnNzb248QlI+ Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDtDYzogY2NhbXBAb3BzLmlldGYub3JnPEJSPiZndDsgJmd0 OyAmZ3Q7ICZndDsmZ3Q7U3ViamVjdDogUmU6IEZyYW1lZm9ybWF0IGluIGEgbDJjcyBnbXBscyBy bnZpcm9ubWVudC48QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsg Jmd0OyZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0 OyZndDsmZ3Q7UGVyIGFuZCBEaW1pdHJpLDxCUj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0OyZndDs8 QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDsmZ3Q7SSB3b3VsZCBsaWtlIHRvIGNvbWUgZG93biBz dHJvbmdlciB0aGFuIHRoYXQsIGZvciBtZSBpdCBpczxCUj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0 OyZndDthIHZlcnkgc3Ryb25nIHJlcXVpcmVtZW50IHRoYXQgdGhlIHNhbWUgc3dpdGNoIGNhbiBo YW5kbGU8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDsmZ3Q7Ym90aCBWTEFOcyBhbmQgR01QTHMg dHJhZmljIGNvcnJlY3RseS4gSSBjYW4ndCBkc2VlIGhvdyB0aGF0PEJSPiZndDsgJmd0OyAmZ3Q7 ICZndDsmZ3Q7Jmd0O2NvdWxkIGJlIGRvbmUgaWYgdXNpbmcgdGhlIFZMQU4gdHBpZCB0byBpbmRp Y2F0ZSBHTVBMUzxCUj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0OyZndDt0cmFmZmljLjxCUj4mZ3Q7 ICZndDsgJmd0OyAmZ3Q7Jmd0OzxCUj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0O0lmIHlvdSBldmVy IHdhbnQgdGhhdCBzYW1lIHN3aXRjaCB0byBoYW5kbGUgdHJhZmZpYyBmb3IgYTxCUj4mZ3Q7ICZn dDsgZGlyZWN0bHk8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDtjb25uZWN0ZWQgJm5ic3A7aG9z dCAobm90IHRvIHVuY29tbWVuKSB5b3Ugd291bGQgd2FudCB0aGF0IHRvPEJSPiZndDsgJmd0OyB1 c2Ugbm9ybWFsPEJSPiZndDsgJmd0OyAmZ3Q7ICZndDsmZ3Q7ZXRoZXJuZXQgbWFjYWRkcmVzcyBz d2l0Y2hpbmcuIFNvIG9mIGNvdXJzZSB5b3UgZG8gbm90IHdhbnQ8QlI+Jmd0OyAmZ3Q7ICZndDsg dG8gaGF2ZSB0bzxCUj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0O2Nob29zZSBiZXR3ZWVuIHZsYW4g YW5kIGdtcGxzLCB5b3Ugd291bGQgd2FudCBib3RoIGF0IHRoZTxCUj4mZ3Q7ICZndDsgc2FtZSB0 aW1lLjxCUj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0OzxCUj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0 Oy9wZXI8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZn dDs8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDsm Z3Q7L0xvYTxCUj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0OyZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsg Jmd0OyZndDsmZ3Q7UGFyIE1hdHRzc29uIHdyb3RlOjxCUj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0 OyZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDsmZ3Q7Jmd0OyZndDtoaSBwYXIsIG9uZSBv ZiB0aGUgcG9zc2liaWxpdGllcyB0aGF0IGhhcyBiZWVuIGNvbnNpZGVyZWQ8QlI+Jmd0OyAmZ3Q7 ICZndDsgJmd0OyZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDt0byBjb3BlIHdpdGg8QlI+ Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDsmZ3Q7Jmd0 OyZndDt0aGlzIHJlcXVpcmVtZW50IGlzIHRvIHVzZSBhIGRlZGljYXRlZCBUUElEIGZvciB0aGU8 QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDtFdGhl cm5ldCBsYWJlbGVkPEJSPiZndDsgJmd0OyAmZ3Q7ICZndDsmZ3Q7PEJSPiZndDsgJmd0OyAmZ3Q7 ICZndDsmZ3Q7Jmd0OyZndDsmZ3Q7ZnJhbWVzOyB0aGlzIHdvdWxkIGFsbG93IGRpZmZlcmVudGlh dGVkIHByb2Nlc3Npbmcgd2l0aDxCUj4mZ3Q7ICZndDsgJmd0OyBub24tbGFiZWxlZDxCUj4mZ3Q7 ICZndDsgJmd0OyAmZ3Q7Jmd0OyZndDsmZ3Q7Jmd0O2ZyYW1lc3RoYW5rcy48QlI+Jmd0OyAmZ3Q7 ICZndDsgJmd0OyZndDsmZ3Q7Jmd0OzxCUj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0OyZndDsmZ3Q7 PEJSPiZndDsgJmd0OyAmZ3Q7ICZndDsmZ3Q7Jmd0OyZndDtUaGF0IHNlZW1zIHRvIG1ha2UgbW9y ZSBzZW5jZS4gSWYgdGhhdCBmcmFtZSBpcyB0byBiZTxCUj4mZ3Q7ICZndDsgc2l6ZWQgbGlrZSBh PEJSPiZndDsgJmd0OyAmZ3Q7ICZndDsmZ3Q7Jmd0OyZndDs4MDIuMXE8QlI+Jmd0OyAmZ3Q7ICZn dDsgJmd0OyZndDsmZ3Q7Jmd0O2ZyYW1lLiBUaGVyZSBpcyBub3QgdGhhdCBtdWNoIHNwYWNlIGxl ZnQgdG8gYSBsYWJlbC4gT3IgaXM8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDs8QlI+Jmd0OyAm Z3Q7ICZndDsgJmd0OyZndDt0aGUgZGVtYW5kIHRvPEJSPiZndDsgJmd0OyAmZ3Q7ICZndDsmZ3Q7 PEJSPiZndDsgJmd0OyAmZ3Q7ICZndDsmZ3Q7Jmd0OyZndDt1c2UganVtYm9mcmFtZXMgPzxCUj4m Z3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0OyZndDsmZ3Q7SGFzIHRoZXJlIGJlZW4gYW55IGRpc2N1c3Np b24gb24gbGFiZWxzdGFja2luZywgYW5kPEJSPiZndDsgJmd0OyAmZ3Q7IG1haW5seSB3aGVyZSB0 bzxCUj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0OyZndDsmZ3Q7cGxhY2U8QlI+Jmd0OyAmZ3Q7ICZn dDsgJmd0OyZndDsmZ3Q7Jmd0O3RoZSBpbmZvcm1hdGlvbj88QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0 OyZndDsmZ3Q7Jmd0OzxCUj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0OyZndDsmZ3Q7UmVnYXJkcy48 QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDsmZ3Q7Jmd0O1BlcjxCUj4mZ3Q7ICZndDsgJmd0OyAm Z3Q7Jmd0OyZndDsmZ3Q7PEJSPiZndDsgJmd0OyAmZ3Q7ICZndDsmZ3Q7Jmd0OyZndDs8QlI+Jmd0 OyAmZ3Q7ICZndDsgJmd0OyZndDsmZ3Q7Jmd0OzxCUj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0OyZn dDs8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDsmZ3Q7PEJSPiZndDsgJmd0OyAmZ3Q7ICZndDsm Z3Q7Jmd0Oy0tPEJSPiZndDsgJmd0OyAmZ3Q7ICZndDsmZ3Q7Jmd0O0xvYSBBbmRlcnNzb248QlI+ Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDsmZ3Q7PEJSPiZndDsgJmd0OyAmZ3Q7ICZndDsmZ3Q7Jmd0 O1ByaW5jaXBhbCBOZXR3b3JraW5nIEFyY2hpdGVjdDxCUj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0 OyZndDtBY3JlbyBBQiAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAm bmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgcGhvbmU6ICZu YnNwOys0NiA4IDYzMiA3NyAxNDxCUj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0OyZndDtJc2Fmam9y ZHNnYXRhbiAyMiAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJz cDsgJm5ic3A7ICZuYnNwO21vYmlsZTogKzQ2IDczOSA4MSAyMSA2NDxCUj4mZ3Q7ICZndDsgJmd0 OyAmZ3Q7Jmd0OyZndDtLaXN0YSwgU3dlZGVuICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAm bmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDtlbWFpbDo8QlI+ Jmd0OyAmZ3Q7IGxvYS5hbmRlcnNzb25AYWNyZW8uc2U8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZn dDsmZ3Q7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAm bmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZu YnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwO2xvYUBwaS5zZTxC Uj4mZ3Q7ICZndDsgJmd0OyAmZ3Q7Jmd0OyZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDs8 QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDs8QlI+Jmd0OyAmZ3Q7ICZndDsgJmd0OyZndDs8QlI+ Jmd0OyAmZ3Q7ICZndDsgJmd0OzxCUj4mZ3Q7ICZndDsgJmd0OzxCUj4mZ3Q7ICZndDsgJmd0OzxC Uj4mZ3Q7ICZndDsgJmd0OyAtLTxCUj4mZ3Q7ICZndDsgJmd0OyBMb2EgQW5kZXJzc29uPEJSPiZn dDsgJmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyAmZ3Q7IFByaW5jaXBhbCBOZXR3b3JraW5nIEFyY2hp dGVjdDxCUj4mZ3Q7ICZndDsgJmd0OyBBY3JlbyBBQiAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJz cDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNw OyAmbmJzcDsgcGhvbmU6ICZuYnNwOys0NiA4IDYzMiA3NyAxNDxCUj4mZ3Q7ICZndDsgJmd0OyBJ c2Fmam9yZHNnYXRhbiAyMiAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNw OyAmbmJzcDsgJm5ic3A7ICZuYnNwO21vYmlsZTogKzQ2IDczOSA4MSAyMSA2NDxCUj4mZ3Q7ICZn dDsgJmd0OyBLaXN0YSwgU3dlZGVuICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsg Jm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDtlbWFpbDogJm5ic3A7bG9h LmFuZGVyc3NvbkBhY3Jlby5zZTxCUj4mZ3Q7ICZndDsgJmd0OyAmbmJzcDsgJm5ic3A7ICZuYnNw OyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7 ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsg Jm5ic3A7ICZuYnNwOyAmbmJzcDsgbG9hQHBpLnNlPEJSPiZndDsgJmd0OyAmZ3Q7PEJSPiZndDsg Jmd0OzxCUj4mZ3Q7ICZndDs8QlI+Jmd0OyAmZ3Q7PEJSPiZndDsgJmd0OzxCUj4mZ3Q7PEJSPiZn dDs8QlI+Jmd0OzwvRk9OVD48L1A+ Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 26 Jul 2005 17:03:45 +0000 Message-Id: <B946BD9AB30355458B1B9629C6A601BE0356F652@E8PBE.blf01.telekom.de> From: Michael.Dueser@t-systems.com To: ccamp@ops.ietf.org Subject: GMPLS VCAT and LAGR Date: Tue, 26 Jul 2005 19:02:22 +0200 MIME-Version: 1.0 Content-Type: text/plain Hi, some comments re I-D draft-imajuku-ccamp-gmpls-vcat-lagr-req-00: - LCAS was not intended for grouping of VC-X-yvs, but for synchronization of the VC-X which is added or dropped. And it's bi-directional. - Only the terminating nodes need to have the full GFP+LCAS+VCAT functionality. - You need to be careful when using the verbs may, shoul and must - are they conform to IETF convention? I believe in general this is an important I-D also wrt to the ITU-T liaison and for implementation! I believe we can also discuss details during the IETF next week. Michael Dueser Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 26 Jul 2005 14:45:08 +0000 Message-ID: <021401c591f0$cba2cd70$a3919ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <ccamp@ops.ietf.org> Subject: Online agenda Date: Tue, 26 Jul 2005 15:25:05 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Available at http://www.olddog.co.uk/63/ccamp.htm Will be updated as slides reach me. Cheers, Adrian Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 26 Jul 2005 08:44:42 +0000 Sensitivity: Subject: RE: Moving LSP ownership between control plane and management plane To: julien.meuric@francetelecom.com Cc: ccamp@ops.ietf.org Message-ID: <OF68AFF6B3.F7A72C8C-ONC125704A.002FD2F5-C125704A.002FE9EB@uk.marconicomms.com> From: "Diego Caviglia" <Diego.Caviglia@marconi.com> Date: Tue, 26 Jul 2005 10:43:01 +0200 MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii Hi Julien, I'll be in Paris next week and will be a pleasure to discuss this with you. Regards Diego "MEURIC Julien RD-CORE-LAN" <julien.meuric@francetelecom.com> on 25/07/2005 18.18.11 To: "Diego Caviglia" <Diego.Caviglia@marconi.com> cc: "ccamp" <ccamp@ops.ietf.org> Subject: RE: Moving LSP ownership between control plane and management plane Hello Diego, Adrian and all. In my view, I feel that moving LSP ownership between MP and CP are very useful functions for a carrier operational network, whether it be for migration to GMPLS or for network maintenance. As far as I understand, the I-D is a proposal for the signalling part and does not pretend solving every associated issues. Some work will probably be needed to implement the whole process, but this part looks interesting to me. I would be glad to discuss further this matter next week in Paris. Regards Julien Julien Meuric France Telecom Research & Development Core Network 2, avenue Pierre Marzin 22307 Lannion Cedex - France Tel : + 33 (0)2 96 05 28 28 Fax : + 33 (0)2 96 05 12 52 http://www.francetelecom.com/rd/ -----Original Message----- From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Diego Caviglia Hi Geroge, sorry but I don't see how the management effort is something related to this ID and more in general with CCAMP. There are already RFC ID that provide tools to signal LSP with ERO filled up to the Label level and of course this can be done only via management, I don't remeber any objections on that due to NMS impact. I'd like to discuss here (I mean in the CCAMP ML) about the ID, that is, if it useful and if could work and not about the difficulties it implies for the NMS. I'm open to discuss, and actually I did, NMS related issues on the ITU-T ML. Regards Diego George Newsome <gnewsome@ieee.org>@ops.ietf.org on 20/07/2005 11.03.00 Sent by: owner-ccamp@ops.ietf.org To: Adrian Farrel <adrian@olddog.co.uk> cc: ccamp@ops.ietf.org, Neil Harrison <neil.2.harrison@bt.com>, Dino Bramanti <Dino.Bramanti@marconi.com> Subject: Re: Moving LSP ownership between control plane and management plane Adrian, Dino, etc., This thread has rapidly got too long, and I don't have the time at the moment to argue the toss, so I'll reply as a note rather than intersperse remarks. On Adrian's question of which operators were at Holmdel - AT&T and T-Systems. On do operators think they want this - some do some don't. BT clearly don't see much value in fiddling with signaling. You ask whether this might be done in service, and there are certainly those who would like to see just that. Related questions are how many times in the life of the system might this be done for the same connection? Is it cost effective to develop a highly reliable solution that is used just once in the life of the system? However this is not so much a discussion of whether this could be useful, but a technical discussion of whether changing a bit (and a slew of procedures) in a signaling protocol actually provides the claimed facility. Because of the effort the management system needs to provide before signaling might be involved, changing the signaling protocol gives the illusion of success, yet does not provide the claimed service at all. Once the management systems have got as far as assigning control plane names, we can ask what happens to the cross connect object when it is deleted as part of the hand off procedure? The behavior of this object is that its deletion results in dropping the matrix connection, which is exactly not required during the change over. It is my considered opinion that fiddling with signaling merely obfuscates the significant difficulties, which are glossed over by stating that those difficulties occur at a different time. Regards George Adrian Farrel wrote: > Thanks George, > > Useful input. > > >>You asked for comments on this proposal, which was presented to ITU-T >>Q14 experts in January of this year. > > > Can you confirm that there were plenty of providers represented at the > meeting. > > >>There are so many issues involved in moving a connection from the >>management plane to the control plane that involve intensive management >>action, that it is not evident that signaling need be involved at all. > > > Interesting. It is certainly valuable to have a list of these issues out > in the open. I have made some comments in the appended meeting report. > > If the control plane is to take over the connection then we assume that > signaling state must be built so that the control plane could then modify > or tear down the connection. how would you propose that this be done > without involving signaling? > > >>As a personal opinion, it makes little sense to consider moving the >>connection. It makes more sense to replace the connection with one set >>up via the control plane, and to then delete the original management set >>up connection. > > > The question arrises of how this would be done in a network with limited > resources. That is, if it is possible to set up a parallel conneciton then > the method you describe would be fine, but if there are not enough > parallel resources then would you propose that a lambda (for example) > could be used for the management-established connection and for the > control-plane-owned connection at the same time? This might present some > difficulties as well. > > Do you have an opinion on my main question which is whether this tyoe of > change-over would ever be done in service? > > Thanks, > Adrian > > >>Excerpt of the meeting report of Q14 Experts, January 2005 >> >>WD06 (Marconi) "G.7713 Modification in order to support Circuit >>Migration" contains a modified version of draft G.7713 Revision as was >>presented in the Berlin meeting (WD 11, i.e., TD50/3 Dec.2004). This >>contribution addresses SP24 (PC ? SPC) of G.7713. It proposes to extend >>the concept of Call/Connection setup to setup/adopt and release to >>release/de-adopt. In Connection Adoption both the SNC are not actually >>created and the LC are not actually established due to the fact that the >>underlying physical resources are already in place. In the >>Call/Connection dis-adoption SNC-LC-SNC is not actually de-allocated, >>only the Control Plane information associated with then is removed and >>the ownership is moved to Management Plane. Attributes for indicating >>Adoption/Dis-adoption are proposed for the INNI Connection messages. >> >>The discussion was mainly on the Management plane and Control plane >>actions required before signalling is involved. The group noted that the >>following issues need to be addressed: >> >>- Naming of transport resources to the control plane. Before a call can >>be placed under signalling control, links that are involved in the >>connection need to be given control plane names. Without these, no >>explicit route can be formed to match the resources of the connection. > > > This is pretty obvious. But actually, this is an issue with moving the > resources from one plane to another (or having them present in both > planes). It is certainly a prerequisite for moving the connection, but it > is not functionally an element of moving the connection. > > The draft, as I read it, assumes that these operations have been completed > and then asks "how do I move the connection from one plane to the other?" > > >>- Resource state while under dual CP and MP views. After resources are >>given control plane names, the resources are still viewed by the >>management plane. It may be necessary to create a new state for the >>resource to indicate that the management plane cannot perform some >>actions on the connection points as the control plane is about to take > > over. > > Yes. There is certainly scope for a race condition here. I believe that > the I-D proposes to (try to) make the transfer an atomic action from the > point of view of the change-over. > > Note, however, that it is wrong to assume that a resource belongs only to > the control or to the management plane at any moment in time. This may be > true with respect to the ability to provision/control connections, but > resources that are available for use by the control plane can still be > managed through the management plane. Thus, you take a laser out of > service through the management plane, not through the control plane. > > >>- Role of discovery functions (esp. CELA). After control plane names >>are allocated, distributed control plane functions need to be associated >>and communication adjacencies formed. This too is a precursor to any >>signalling procedures in migrating from the MP to CP. > > > Note to CCAMP: CELA is Control Entity Logical Adjacency. > This too is "obvious". > But this discussion seems to be wide of the mark. Before deploying a > control plane, it is clear that this type of operation needs to take > place. Since (within the context of CCAMP) a control plane *is* going to > be deployed, we can assume this operation and move on to the discussion in > the draft. > > >>- Similarity of migration to synchronization after CP failure and >>subsequent recovery. Connection control procedures might be the same in >>migrating a call from the MP to the CP as the situation when the CP has >>failed and is recovering. Here, the network connections are already in >>place, but connection and call state need to be created to match it. >>Knowledge of the call and what the connection should be is obtained from >>the MP for migration, and from some reliable database in the recovery > > case. > > This is a valid point. > Control plane failure, however, is unlikely to be total. That is, it would > be unusual for more than one control plane link or node to fail. Thus, in > the control plane, the "reliable database" is distributed and recovery can > follow the procedures outlined in draft-ietf-ccamp-rsvp-restart-ext-03.txt > > It would certainly be worth considering whether the procedures used for > control plane recovery could be applied to transfer of ownership from the > management to control plane. at the moment, I suspect that they could not > because the restart extensions assume a specific period during which > signaling is associated with recovery. > > We would also need to look at the reverse process (transfer from CP to MP) > that is not possible to cover by CP restart procedures. > > >>- Call awareness of migration vs. connection being unaware of migration. >> When connection state is being created to match an existing >>connection, the connection controllers (CCs) do not require awareness of >>why this is happening as the context could be migration or recovery. A >>mechanism to create control state without affecting transport plane >>state is needed in the CCs. Call controllers though do need migration >>awareness as they need to obtain/derive call identification from the MP. > > > I disagree. > Connection controllers must be able to distinguish between resources that > are already in use and cannot be assigned to a new connection, and those > that are already in use and can be assigned to a new connection because > the connection is replacing an existing connection. If this were happening > entirely within the control plane we would use resource sharing, but when > the operation spans the control and management planes resource > ownership/sharing is more complex and the conneciotn controller needs to > be able to issue the approriate instrucitons for the assignment of the > resources. > On the other hand, call controllers do not need awareness of the details > of the realisation of the connections. > > >>In summary, the group decided that it is premature to consider >>signalling procedure before the above issues, amongst other, have been >>studied. However the contribution does enable the group to expose in a >>larger context the interaction between CP and MP for connection >>migration. The above issues were included in SP24 of the G.7713 Living >>List (see WD22). > > > Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 26 Jul 2005 06:02:23 +0000 Thread-Topic: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet thread-index: AcWRp4QlwPC80+CYQx66A8lU64uuxQ== Reply-To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> From: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> To: "Ali Sajassi \(sajassi\)" <sajassi@cisco.com>, "Heiles Juergen" <juergen.heiles@siemens.com>, <per@defero.se> Cc: <ccamp@ops.ietf.org> Bcc: Subject: RE: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet Date: Tue, 26 Jul 2005 15:01:51 +0900 Comment: GQ19@|@ZEk=E?,18?x, BcN1b<z:P<.F@, 4c4g Message-ID: <58c001c591a7$842d1480$8310fe81@email1> MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit Content-Class: urn:content-classes:message Dear Ali, Thank you your comments. please see below, >> 1. using VLAN tag as it is for L2 label encoding. > >As you know this approach has scalability issue and you are limited to >4K services in the network and if you try to use the upcoming IEEE >802.1ad standard to make VLANs, interface specific, then you will need >new switches and not existing Ethernet switches. Even if you use these >newer switches, you will still be limited to 4K services per interface >and it would defeat the objective of using existing switches. > I agree. That's the point I also want to make to those proponents of this approach. >> >> 2. defining new protocol ID (TPID) and borrow VLAN tag format, >> ?or extended VLAN tag to implement label swapping. >> > >This approach requires new data plane functionality to be implemented >(which defeats the objective of using existing switches). > >> 3. use lower 3 bytes of MAC address for L2 label encoding. > >This approach has several issues: >1) how does MAC addresses get distributed among the bridges ? >How does this distribution works during a link or node failure ? >How does it recover from a link or node failure ? > Ali, The questions are really implementation dependent. Can we defer this discussion after Paris meeting? However, I see the solution will not be so different than what RSVP-TE has been done for other transport network. >2) Will you assign a separate provider MAC for each customer MAC. >If so, then how do you address MAC scalability. >If not, then how do you take care of 1-to-N mapping between provider >MAC address and customer MAC addresses >and how do you handle multipoint connections. I think you are asking the issue of L2 label stacking or LSP aggregation, or perhaps some VPN application similar to 802.1ad. There are some opinions on whether label stacking is necessary in L2SC domain. This is actually the question Adrian keep asking us. I personally think that we can achieve LSP aggregation with the aid of MPLS shim label, however I think this is too premature discussion. All I can say at the moment is, lower 3 octet of MAC address gives us 16M size of label space and this is enough to implement link-local, site-local or even some domain-local LSPs. Between the L2-LSP boundary, it needs further discussion. > >3) how do handle customer FCS retention when doing this MAC translation >stuff ?? > I haven't seen any application that requires retention of original FCS image. 802.1D specification recommends recalculation of FCS when there is any change in MAC. I see no problem if FCS is recalculated at the label swapping switch. >> 802.1Q bridge forwards Ethernet frames using two dataplane tables >> - MAC forwarding table and VLAN forwarding table. >> Bridge control protocols, such as GARP, GVRP, GMRP, >> manipulate one of the two dataplane entities. > >This is not correct. Bridges don't have different databases for MAC >forwarding and VLAN forwarding. They have filtering datebase(s) with >corresponding IDs (FIDs). VLAN IDs get mapped to FIDs based on operation >mode (IVL versus SVL). Then once a filtering database is identified, >then a MAC address lookup is performed for forwarding. Bridge control >protocols use a designated MAC addressed and based on these MAC >addressed, the bridge knows how to process the frame. oh, come on . I explained it in conceptual level and you re-described it in implementation level. The contextual meaning is same. I believe you don't deny the existence of two types of filtering databases, for MAC and VID, and the control of GARP over these databases. This is unnecessary argument, I think. >> >> (% note however, I would regard implementing label swapping >> function is acceptable change considering the Internal Sublayer >> design of 802.1D bridge.) > >The only way of doing this is using VID translation table specified in >IEEE 802.1ad which requires new data-plane functionality and thus >existing bridges cannot be used !! > If we define the scope of L2 label as site-local or domain-local, we don't need such label swapping operation. However, if we want link-local label, and find out that it is very useful (* actually I do think it'll be very useful *), I think we need to consult such modification with IEEE. Technically speaking, I don't think replacing the label value after normal MAC lookup will serious change hardware architecture. Also, it doesn't break conformance rule. Did any 802.1 spec. say that input MAC must be identical to output MAC? >Also if going the route of tag swapping, you may consider using IEEE >802.1ah. As the editor of .1ah, I can go over the details when I get a >chance. thanks, hopefully, this will be far different approach than 802.1ah. >> >> However, option 2 is about what TPID code we will >> choose to implement new switching function. >> The switch of this implementation should have new hardware >> design in addition to normal bride hardware. >> I think such proposal as option 2 is out of CCAMP scope. >Yes. ===> I agree. [snip] > >How does the redundancy work in here. Are you using bridge protocol in >conjunction with GMPLS protocol. If you do, then there are bunch of >issues. Yes I do. I intend GMPLS implemented switches being deployed in operating network as normal Ethernet switch that perform bridge protocols. When they detect (or snoop) RSVP-TE message, they establish L2-LSP and selectively forward those GMPLS prefixed frames via the L2-LSP. The message path that RSVP-TE is exchanged, thus L2-LSPs are established, does not necessarily be identical to spanning tree path because L2-LSPs can be established via shortest path. I welcome further discussion on this issue, but I hope to continue it after Paris meeting, if you may. Thanks Jaihyung -----?? ???----- ?? ??: "Ali Sajassi (sajassi)" <sajassi@cisco.com> ?? ??: 2005-07-26 ?? 8:13:56 ?? ??: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles Juergen" <juergen.heiles@siemens.com>, "per@defero.se" <per@defero.se> ??: "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> ??: RE: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet Jaihyung, It seems to me that your primary objective is to use the existing Ethernet switches with new control plane (GMPLS) to setup TE paths for different applications. And you mentioned three approaches for doing this in your email. However, all three of them have issues and are not viable. Lets go over them ... > > In below mail, I discussed about three proposed approaches. > > > 1. using VLAN tag as it is for L2 label encoding. As you know this approach has scalability issue and you are limited to 4K services in the network and if you try to use the upcoming IEEE 802.1ad standard to make VLANs, interface specific, then you will need new switches and not existing Ethernet switches. Even if you use these newer switches, you will still be limited to 4K services per interface and it would defeat the objective of using existing switches. > > 2. defining new protocol ID (TPID) and borrow VLAN tag format, > ?or extended VLAN tag to implement label swapping. > This approach requires new data plane functionality to be implemented (which defeats the objective of using existing switches). > 3. use lower 3 bytes of MAC address for L2 label encoding. > This approach has several issues: 1) how does MAC addresses get distributed among the bridges ? How does this distribution works during a link or node failure ? How does it recover from a link or node failure ? 2) Will you assign a separate provider MAC for each customer MAC. If so, then how do you address MAC scalability. If not, then how do you take care of 1-to-N mapping between provider MAC address and customer MAC addresses and how do you handle multipoint connections. 3) how do handle customer FCS retention when doing this MAC translation stuff ?? > > 802.1Q bridge forwards Ethernet frames using two dataplane tables > - MAC forwarding table and VLAN forwarding table. > Bridge control protocols, such as GARP, GVRP, GMRP, > manipulate one of the two dataplane entities. This is not correct. Bridges don't have different databases for MAC forwarding and VLAN forwarding. They have filtering datebase(s) with corresponding IDs (FIDs). VLAN IDs get mapped to FIDs based on operation mode (IVL versus SVL). Then once a filtering database is identified, then a MAC address lookup is performed for forwarding. Bridge control protocols use a designated MAC addressed and based on these MAC addressed, the bridge knows how to process the frame. > > Similarly, option 1 and 3 are about which one of two > dataplane entities > GMPLS protocol should control on behalf of bridge control protocols. > The two proposals do not intend to modify bridge behavior > seriously, such as MAC learning, aging, filtering. > Therefore, the approaches 1 and 3 are in the scope of CCAMP. If that's the intention, you may find out soon you get more than what you asked for :-) > > (% note however, I would regard implementing label swapping > function is acceptable change considering the Internal Sublayer > design of 802.1D bridge.) > The only way of doing this is using VID translation table specified in IEEE 802.1ad which requires new data-plane functionality and thus existing bridges cannot be used !! Also if going the route of tag swapping, you may consider using IEEE 802.1ah. As the editor of .1ah, I can go over the details when I get a chance. > > However, option 2 is about what TPID code we will > choose to implement new switching function. > The switch of this implementation should have new hardware > design in addition to normal bride hardware. > I think such proposal as option 2 is out of CCAMP scope. Yes. > > Overall objective is improving scalability, traffic engineering (TE) > characteristics of 802.1 bridge that it can be reliable, manageable > enough to replace some core routers. Easier said than done :-) > The switching technique you mentioned above, such as > Ethernet over MPLS as defined by PWE3 and L2VPN, > are all actually router based technology, however this > work is based on simple bridge architecture. > Cost-effectiveness is the key differentiator. > The devils are in the details and once you worked out the details, you will see the issues. > > In access or enterprise network, capability of providing end-to-end > L2-LSP will enable service providers policing, measuring, charging > application flows using Ethernet based network. This will eventually > improve income structure and introduce new session based commercial > service. I have been discussing this aspect in mail thread of > title 'End-to-end L2-LSP'. Please read the mail thread and comment > on the discussion. > > > >You mentioned to use the MAC address or part of it as a label. > >The TRILL working group is defining shortest path routing > for Ethernet. > >How would this fit together? > > > TRILL employs new encapsulation layer outside of Ethernet frames. > However in my proposal of option-3, I simply use the Ethernet header > to encode GMPLS label. > > The overall format of rbridge is > [Ethernet][r-tag][Ethernet][ data..] > It is not LSP based technology but a connectionless routing bridge. > TTL count is very important in rbridge because frames are routed > hop-by-hop and it may loop. > > Whereas in option-3 above, Ethernet frames that have certain > OUI prefix in MAC address are forwarded via a path that > GMPLS protocol has configured on MAC forwarding table. > How does the redundancy work in here. Are you using bridge protocol in conjunction with GMPLS protocol. If you do, then there are bunch of issues. If you don't, then what is it used and how does it interact with bridging protocols ?? -Ali > > Thanks > > Jaihyung > > > > -----?? ???----- > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com> > ?? ??: 2005-07-25 ?? 6:44:34 > ?? ??: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles > Juergen" <juergen.heiles@siemens.com>, "per@defero.se" > <per@defero.se> > ??: "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> > ??: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > > Dear Jaihyung, > > you mention several options that could be used as a label. > Most of them have strong impacts on the data plane and > introduce new layer networks. Is this in the scope of CCAMP? > What is the goal, to introduce a control plane for a existing > L2 technology or introduce a new L2 technology? What would be > the benefit of such a new L2 co switching technology compared > to Ethernet over MPLS as defined by PWE3 and L2VPN? > You mentioned to use the MAC address or part of it as a > label. The TRILL working group is defining shortest path > routing for Ethernet. How would this fit together? > > > Regards > > Juergen > > > > > > > -----Original Message----- > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > > Sent: Saturday, July 23, 2005 6:19 AM > > To: Heiles Juergen; per@defero.se > > Cc: ccamp@ops.ietf.org > > Subject: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > Dear Juergen, Par and all, > > > > The proposed framework of L2SC was not intended to suggest > > any specific solution. It is still an open question. > > As Adrian noted, it only aims to present requirements > > as in scenarios that may drive people toward next step in many > > viable solutions. > > > > In fact, there are other options we may also consider for L2 > > label encoding. > > IEEE 802.1 Ethernet bridges forward frames based on 48bits of > > MAC address, > > and additionally using VLAN tag. > > > > When the purpose of GMPLS control over Ethernet is not to > > create new dataplane > > but to utilize IEEE 802.1 bridge architecture, we may > consider using > > one of the two forwarding methods, MAC forwarding or VLAN > forwarding. > > (and perhaps any other combination of fields in MAC, but I'll > > not discuss it) > > > > Use of VLAN ID for label encoding may automate VLAN configuration > > using IP protocols. However, GMPLS protocol cannot use the > > field exclusively > > because public/private operators already use VLAN for various > > purpose. > > There is a potential conflict with existing use of VLAN and > > GMPLS use of VLAN label. > > Furthermore, scalability of VLAN ID has been frequently noted > > as weakness > > because the size of VLAN ID is at most 4096 (12bit). > > > > The scalability may be improved if the scope of VLAN label is > > confined to > > link-local, and some additional swapping function of VLAN ID > > is introduced in > > Internal Sublayer [802.1D] of GMPLS implemented switch. > > However, this will only be effective when the configuration > > of network > > is mesh structure that multiple LSP paths exist. If the > > configuration of network > > is star or tree shape, as normal configuration of access network, > > LSPs concentrate in root node and total number of > > LSPs that the network can hold still be limited by available > > label space > > at a few root links. > > > > For these reasons, I do not think any form of VLAN ID label is > > an appropriate choice for layer-2 label encoding. > > > > There are some other proposals assuming new assignment of Ethernet > > Length/Type value (e.g. new TPID in VLAN tag) and re-definition of > > information fields placed between 802.3 MAC header and IP packet. > > In this case, only the format of VLAN tag or extended VLAN tag > > is borrowed, however, inside the switching hardware, the > > filter and relay, > > etc. are totally different new dataplane switch. > > > > I do not see such approach is a GMPLS implementation for Ethernet > > because the core switching technique is not 802.1 Ethernet > > bridge at all. > > > > The other option we may consider is using MAC address filed > as below. > > > > > > +-------+-------+-------+-------+-------+-------+ > > | 1byte | 2byte | 3byte | 4byte | 5byte | 6byte | > > +-------+-------+-------+-------+-------+-------+ > > +-----------------------+-----------------------+ > > | OUI Prefix (=GMPLS) | DA-label (24bit) | > > +-----------------------+-----------------------+ > > | OUI Prefix (=GMPLS) | SA-label (24bit) | > > +---------------+-------+-----------------------+ > > | Length/Type | > > +---------------+ > > > > > > IEEE is designated by the ISO Council to act as the > > registration authority > > for the higher three-octet of OUI number in the MAC address > > to be used by manufacturer. Ethernet manufacturer may generate > > global unique MAC address using the OUI prefix and address block of > > lower three-octet (24bit). Taking advantage of the > addressing scheme, > > GMPLS may use the lower three-octet exclusively if a unique > > OUI number > > is reserved for the protocol. With this labeling scheme, GMPLS will > > control MAC forwarding entry, not VLAN table. > > > > All Ethernet frames controlled by GMPLS will have identical > > OUI number > > that they can easily be distinguished from other Ethernet frames. > > In principle, the label lookup hardware is identical to MAC lookup > > hardware in this labeling scheme. Therefore GMPLS implemented > > switch may still function as normal Ethernet bridge to the frames > > that OUI number is not GMPLS. This also facilitates GMPLS > implemented > > switches being deployed in operating Ethernet with minimum > > service disruption. > > > > Note also that above proposed label encoding method is transparent > > to the use of Ethernet Length/Type field. End-user device may use > > the Length/Type field as defined in IEEE 802.3 protocol. > > It also allows network operators configure VLAN for their own > > purpose. > > When IEEE 802.1p is used in conjunction with L2-LSP, the priority > > field of VLAN tag can also be used for imposing consistent TE > > policy in > > legacy switches and GMPLS switches. > > > > Any way, my conclusion is, there are other options we may consider, > > and this issue is still open to discuss. > > > > Thanks, > > > > Sincerely, > > > > Jaihyung > > > > > > > > > > > > -----?? ???----- > > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com> > > ?? ??: 2005-07-22 ?? 10:23:43 > > ?? ??: "Loa Andersson" <loa@pi.se>, "richard.spencer@bt.com" > > <richard.spencer@bt.com> > > ??: "per@defero.se" <per@defero.se>, "ccamp@ops.ietf.org" > > <ccamp@ops.ietf.org> > > ??: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > > > > > > > > > Loa, > > > > I interpret the ID as a proposal to use GMPLS for VLAN setup. > > So GMPLS and VLAN to not compete. The VLAN is at the data > > plane and GMPLS at the control plane. The question is how and > > should different control plane techniques like GMPLS and > > (GVRP and STB) work together? > > > > Regards > > > > Juergen > > > > > > > -----Original Message----- > > > From: owner-ccamp@ops.ietf.org > > > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson > > > Sent: Friday, July 22, 2005 2:53 PM > > > To: richard.spencer@bt.com > > > Cc: per@defero.se; ccamp@ops.ietf.org > > > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > > > > Richard, > > > > > > I agree to most of this. Since we have doubts about the viability > > > of taking GMPLS all the way to end-user or enterprise I think it > > > would be good, from a wg perspective, if we agreed to solve the > > > core network problems first. > > > > > > Do you have any comment on the requirement to run both VLANs and > > > GMPLS on the same switch? > > > > > > /Loa > > > > > > richard.spencer@bt.com wrote: > > > > Regardless of whether or not a switch is directly connected > > > > to hosts, it must be able to forward packets using the > > > connectionless > > > > Ethernet data plane. This is due to the fundamental > > > requirement that > > > > for GMPLS switches to be able to exchange control > > information with > > > > each other, a data plane for control traffic must be present. > > > > This is akin to using the IP data plane for MPLS > > signalling in an > > > > IP/MPLS network. An alternative would be to use a static > > reserved > > > > L2-LSP for control traffic in the same way that reserved > > VPI/VCIs > > > > are used for PNNI signalling in ATM. > > > > > > > > Regarding connecting hosts to GMPLS switches, I > personally don't > > > > think extending L2-LSPs into the enterprise/home network is > > > > commercially viable. However, if you do want to use > > GMPLS switches > > > > in the home/enterprise network and for some reason > don't want to > > > > extend L2-LSPs down to the host then you will not be > performing > > > > normal Ethernet Mac address switching anyway. Instead you will > > > > need some kind of policy on the switch that maps > connectionless > > > > Ethernet packets (e.g. based on MAC src/dest, 802.1p, > VLAN) to a > > > > L2-LSP. This is because multiple L2-LSPs to the same > destination > > > > (e.g. a gateway router) may exist for different services/flows > > > > (e.g. video download, VoIP call, etc.). > > > > > > > > Regards, > > > > Richard > > > > > > > > > > > >>-----Original Message----- > > > >>From: owner-ccamp@ops.ietf.org > > [mailto:owner-ccamp@ops.ietf.org]On > > > >>Behalf Of Par Mattsson > > > >>Sent: 22 July 2005 11:42 > > > >>To: Loa Andersson > > > >>Cc: ccamp@ops.ietf.org > > > >>Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > >> > > > >> > > > >> > > > >>>Per and Dimitri, > > > >>> > > > >>>I would like to come down stronger than that, for me it is > > > >>>a very strong requirement that the same switch can handle > > > >>>both VLANs and GMPLs trafic correctly. I can't dsee how that > > > >>>could be done if using the VLAN tpid to indicate GMPLS > > > >>>traffic. > > > >> > > > >>If you ever want that same switch to handle traffic for a > > directly > > > >>connected host (not to uncommen) you would want that to > > use normal > > > >>ethernet macaddress switching. So of course you do not want > > > to have to > > > >>choose between vlan and gmpls, you would want both at the > > same time. > > > >> > > > >>/per > > > >> > > > >> > > > >> > > > >>>/Loa > > > >>> > > > >>>Par Mattsson wrote: > > > >>> > > > >>>>>hi par, one of the possibilities that has been considered > > > >> > > > >>to cope with > > > >> > > > >>>>>this requirement is to use a dedicated TPID for the > > > >> > > > >>Ethernet labeled > > > >> > > > >>>>>frames; this would allow differentiated processing with > > > non-labeled > > > >>>>>framesthanks. > > > >>>> > > > >>>> > > > >>>>That seems to make more sence. If that frame is to be > > sized like a > > > >>>>802.1q > > > >>>>frame. There is not that much space left to a label. Or is > > > >> > > > >>the demand to > > > >> > > > >>>>use jumboframes ? > > > >>>>Has there been any discussion on labelstacking, and > > > mainly where to > > > >>>>place > > > >>>>the information? > > > >>>> > > > >>>>Regards. > > > >>>>Per > > > >>>> > > > >>>> > > > >>>> > > > >>> > > > >>> > > > >>>-- > > > >>>Loa Andersson > > > >>> > > > >>>Principal Networking Architect > > > >>>Acreo AB phone: +46 8 632 77 14 > > > >>>Isafjordsgatan 22 mobile: +46 739 81 21 64 > > > >>>Kista, Sweden email: > > loa.andersson@acreo.se > > > >>> loa@pi.se > > > >>> > > > >> > > > >> > > > >> > > > > > > > > > > > > > -- > > > Loa Andersson > > > > > > Principal Networking Architect > > > Acreo AB phone: +46 8 632 77 14 > > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > > Kista, Sweden email: loa.andersson@acreo.se > > > loa@pi.se > > > > > > > > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 23:15:18 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: quoted-printable Subject: RE: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet Date: Mon, 25 Jul 2005 16:13:56 -0700 Message-ID: <75B2A84D9323BC4CA3977CF378CE75EB59F786@xmb-sjc-21e.amer.cisco.com> Thread-Topic: Frameformat in a l2cs gmpls rnvironment - Issues w/ GMPLS Ethernet Thread-Index: AcWRDY8OHVZh5WBvSbOBxtrHtj5d+wAWDnaA From: "Ali Sajassi \(sajassi\)" <sajassi@cisco.com> To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles Juergen" <juergen.heiles@siemens.com>, <per@defero.se> Cc: <ccamp@ops.ietf.org> Jaihyung, It seems to me that your primary objective is to use the existing Ethernet switches with new control plane (GMPLS) to setup TE paths for different applications. And you mentioned three approaches for doing this in your email. However, all three of them have issues and are not viable. Lets go over them ... =20 >=20 > In below mail, I discussed about three proposed approaches. > =20 > =20 > 1. using VLAN tag as it is for L2 label encoding. As you know this approach has scalability issue and you are limited to 4K services in the network and if you try to use the upcoming IEEE 802.1ad standard to make VLANs, interface specific, then you will need new switches and not existing Ethernet switches. Even if you use these newer switches, you will still be limited to 4K services per interface and it would defeat the objective of using existing switches. >=20 > 2. defining new protocol ID (TPID) and borrow VLAN tag format,=20 > ?or extended VLAN tag to implement label swapping. >=20 This approach requires new data plane functionality to be implemented (which defeats the objective of using existing switches). > 3. use lower 3 bytes of MAC address for L2 label encoding. > =20 This approach has several issues: 1) how does MAC addresses get distributed among the bridges ? How does this distribution works during a link or node failure ? How does it recover from a link or node failure ? 2) Will you assign a separate provider MAC for each customer MAC. If so, then how do you address MAC scalability. If not, then how do you take care of 1-to-N mapping between provider MAC address and customer MAC addresses and how do you handle multipoint connections. 3) how do handle customer FCS retention when doing this MAC translation stuff ?? >=20 > 802.1Q bridge forwards Ethernet frames using two dataplane tables=20 > - MAC forwarding table and VLAN forwarding table.=20 > Bridge control protocols, such as GARP, GVRP, GMRP,=20 > manipulate one of the two dataplane entities. This is not correct. Bridges don't have different databases for MAC forwarding and VLAN forwarding. They have filtering datebase(s) with corresponding IDs (FIDs). VLAN IDs get mapped to FIDs based on operation mode (IVL versus SVL). Then once a filtering database is identified, then a MAC address lookup is performed for forwarding. Bridge control protocols use a designated MAC addressed and based on these MAC addressed, the bridge knows how to process the frame.=20 >=20 > Similarly, option 1 and 3 are about which one of two=20 > dataplane entities=20 > GMPLS protocol should control on behalf of bridge control protocols.=20 > The two proposals do not intend to modify bridge behavior=20 > seriously, such as MAC learning, aging, filtering.=20 > Therefore, the approaches 1 and 3 are in the scope of CCAMP. If that's the intention, you may find out soon you get more than what you asked for :-) >=20 > (% note however, I would regard implementing label swapping=20 > function is acceptable change considering the Internal Sublayer=20 > design of 802.1D bridge.) > The only way of doing this is using VID translation table specified in IEEE 802.1ad which requires new data-plane functionality and thus existing bridges cannot be used !! Also if going the route of tag swapping, you may consider using IEEE 802.1ah. As the editor of .1ah, I can go over the details when I get a chance.=20 =20 >=20 > However, option 2 is about what TPID code we will=20 > choose to implement new switching function.=20 > The switch of this implementation should have new hardware=20 > design in addition to normal bride hardware.=20 > I think such proposal as option 2 is out of CCAMP scope.=20 Yes. >=20 > Overall objective is improving scalability, traffic engineering (TE)=20 > characteristics of 802.1 bridge that it can be reliable, manageable=20 > enough to replace some core routers. Easier said than done :-) > The switching technique you mentioned above, such as=20 > Ethernet over MPLS as defined by PWE3 and L2VPN, > are all actually router based technology, however this > work is based on simple bridge architecture. > Cost-effectiveness is the key differentiator. > The devils are in the details and once you worked out the details, you will see the issues. =20 >=20 > In access or enterprise network, capability of providing end-to-end=20 > L2-LSP will enable service providers policing, measuring, charging=20 > application flows using Ethernet based network. This will eventually=20 > improve income structure and introduce new session based commercial=20 > service. I have been discussing this aspect in mail thread of=20 > title 'End-to-end L2-LSP'. Please read the mail thread and comment > on the discussion. > =20 >=20 > >You mentioned to use the MAC address or part of it as a label.=20 > >The TRILL working group is defining shortest path routing=20 > for Ethernet.=20 > >How would this fit together? > =20 >=20 > TRILL employs new encapsulation layer outside of Ethernet frames.=20 > However in my proposal of option-3, I simply use the Ethernet header=20 > to encode GMPLS label.=20 >=20 > The overall format of rbridge is =20 > [Ethernet][r-tag][Ethernet][ data..]=20 > It is not LSP based technology but a connectionless routing bridge.=20 > TTL count is very important in rbridge because frames are routed=20 > hop-by-hop and it may loop. >=20 > Whereas in option-3 above, Ethernet frames that have certain=20 > OUI prefix in MAC address are forwarded via a path that=20 > GMPLS protocol has configured on MAC forwarding table. > =20 How does the redundancy work in here. Are you using bridge protocol in conjunction with GMPLS protocol. If you do, then there are bunch of issues. If you don't, then what is it used and how does it interact with bridging protocols ?? -Ali >=20 > Thanks > =20 > Jaihyung > =20 > =20 >=20 > -----?? ???-----=20 > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com>=20 > ?? ??: 2005-07-25 ?? 6:44:34=20 > ?? ??: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles=20 > Juergen" <juergen.heiles@siemens.com>, "per@defero.se"=20 > <per@defero.se>=20 > ??: "ccamp@ops.ietf.org" <ccamp@ops.ietf.org>=20 > ??: RE: Frameformat in a l2cs gmpls rnvironment.=20 >=20 >=20 >=20 >=20 >=20 >=20 > Dear Jaihyung,=20 >=20 > you mention several options that could be used as a label.=20 > Most of them have strong impacts on the data plane and=20 > introduce new layer networks. Is this in the scope of CCAMP?=20 > What is the goal, to introduce a control plane for a existing=20 > L2 technology or introduce a new L2 technology? What would be=20 > the benefit of such a new L2 co switching technology compared=20 > to Ethernet over MPLS as defined by PWE3 and L2VPN?=20 > You mentioned to use the MAC address or part of it as a=20 > label. The TRILL working group is defining shortest path=20 > routing for Ethernet. How would this fit together?=20 >=20 >=20 > Regards=20 >=20 > Juergen=20 > =20 >=20 >=20 >=20 >=20 > > -----Original Message-----=20 > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr]=20 > > Sent: Saturday, July 23, 2005 6:19 AM=20 > > To: Heiles Juergen; per@defero.se=20 > > Cc: ccamp@ops.ietf.org=20 > > Subject: RE: Frameformat in a l2cs gmpls rnvironment.=20 > >=20 > >=20 > > Dear Juergen, Par and all,=20 > >=20 > > The proposed framework of L2SC was not intended to suggest=20 > > any specific solution. It is still an open question.=20 > > As Adrian noted, it only aims to present requirements=20 > > as in scenarios that may drive people toward next step in many=20 > > viable solutions.=20 > >=20 > > In fact, there are other options we may also consider for L2=20 > > label encoding.=20 > > IEEE 802.1 Ethernet bridges forward frames based on 48bits of=20 > > MAC address,=20 > > and additionally using VLAN tag.=20 > >=20 > > When the purpose of GMPLS control over Ethernet is not to=20 > > create new dataplane=20 > > but to utilize IEEE 802.1 bridge architecture, we may=20 > consider using=20 > > one of the two forwarding methods, MAC forwarding or VLAN=20 > forwarding.=20 > > (and perhaps any other combination of fields in MAC, but I'll=20 > > not discuss it)=20 > > =20 > > Use of VLAN ID for label encoding may automate VLAN configuration=20 > > using IP protocols. However, GMPLS protocol cannot use the=20 > > field exclusively=20 > > because public/private operators already use VLAN for various=20 > > purpose.=20 > > There is a potential conflict with existing use of VLAN and=20 > > GMPLS use of VLAN label. =20 > > Furthermore, scalability of VLAN ID has been frequently noted=20 > > as weakness=20 > > because the size of VLAN ID is at most 4096 (12bit).=20 > >=20 > > The scalability may be improved if the scope of VLAN label is=20 > > confined to=20 > > link-local, and some additional swapping function of VLAN ID=20 > > is introduced in=20 > > Internal Sublayer [802.1D] of GMPLS implemented switch.=20 > > However, this will only be effective when the configuration=20 > > of network=20 > > is mesh structure that multiple LSP paths exist. If the=20 > > configuration of network=20 > > is star or tree shape, as normal configuration of access network,=20 > > LSPs concentrate in root node and total number of=20 > > LSPs that the network can hold still be limited by available=20 > > label space=20 > > at a few root links.=20 > >=20 > > For these reasons, I do not think any form of VLAN ID label is=20 > > an appropriate choice for layer-2 label encoding.=20 > > =20 > > There are some other proposals assuming new assignment of Ethernet=20 > > Length/Type value (e.g. new TPID in VLAN tag) and re-definition of=20 > > information fields placed between 802.3 MAC header and IP packet.=20 > > In this case, only the format of VLAN tag or extended VLAN tag=20 > > is borrowed, however, inside the switching hardware, the=20 > > filter and relay,=20 > > etc. are totally different new dataplane switch.=20 > >=20 > > I do not see such approach is a GMPLS implementation for Ethernet=20 > > because the core switching technique is not 802.1 Ethernet=20 > > bridge at all.=20 > > =20 > > The other option we may consider is using MAC address filed=20 > as below.=20 > > =20 > >=20 > > +-------+-------+-------+-------+-------+-------+=20 > > | 1byte | 2byte | 3byte | 4byte | 5byte | 6byte |=20 > > +-------+-------+-------+-------+-------+-------+=20 > > +-----------------------+-----------------------+=20 > > | OUI Prefix (=3DGMPLS) | DA-label (24bit) |=20 > > +-----------------------+-----------------------+=20 > > | OUI Prefix (=3DGMPLS) | SA-label (24bit) |=20 > > +---------------+-------+-----------------------+=20 > > | Length/Type |=20 > > +---------------+=20 > > =20 > >=20 > > IEEE is designated by the ISO Council to act as the=20 > > registration authority=20 > > for the higher three-octet of OUI number in the MAC address=20 > > to be used by manufacturer. Ethernet manufacturer may generate=20 > > global unique MAC address using the OUI prefix and address block of=20 > > lower three-octet (24bit). Taking advantage of the=20 > addressing scheme,=20 > > GMPLS may use the lower three-octet exclusively if a unique=20 > > OUI number=20 > > is reserved for the protocol. With this labeling scheme, GMPLS will=20 > > control MAC forwarding entry, not VLAN table.=20 > > =20 > > All Ethernet frames controlled by GMPLS will have identical=20 > > OUI number=20 > > that they can easily be distinguished from other Ethernet frames.=20 > > In principle, the label lookup hardware is identical to MAC lookup=20 > > hardware in this labeling scheme. Therefore GMPLS implemented=20 > > switch may still function as normal Ethernet bridge to the frames=20 > > that OUI number is not GMPLS. This also facilitates GMPLS=20 > implemented=20 > > switches being deployed in operating Ethernet with minimum=20 > > service disruption.=20 > > =20 > > Note also that above proposed label encoding method is transparent=20 > > to the use of Ethernet Length/Type field. End-user device may use=20 > > the Length/Type field as defined in IEEE 802.3 protocol.=20 > > It also allows network operators configure VLAN for their own=20 > > purpose.=20 > > When IEEE 802.1p is used in conjunction with L2-LSP, the priority=20 > > field of VLAN tag can also be used for imposing consistent TE=20 > > policy in=20 > > legacy switches and GMPLS switches.=20 > > =20 > > Any way, my conclusion is, there are other options we may consider,=20 > > and this issue is still open to discuss.=20 > > =20 > > Thanks,=20 > > =20 > > Sincerely,=20 > > =20 > > Jaihyung=20 > > =20 > > =20 > > =20 > > =20 > >=20 > > -----?? ???-----=20 > > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com>=20 > > ?? ??: 2005-07-22 ?? 10:23:43=20 > > ?? ??: "Loa Andersson" <loa@pi.se>, "richard.spencer@bt.com"=20 > > <richard.spencer@bt.com>=20 > > ??: "per@defero.se" <per@defero.se>, "ccamp@ops.ietf.org"=20 > > <ccamp@ops.ietf.org>=20 > > ??: RE: Frameformat in a l2cs gmpls rnvironment.=20 > >=20 > >=20 > >=20 > >=20 > >=20 > >=20 > > Loa,=20 > >=20 > > I interpret the ID as a proposal to use GMPLS for VLAN setup.=20 > > So GMPLS and VLAN to not compete. The VLAN is at the data=20 > > plane and GMPLS at the control plane. The question is how and=20 > > should different control plane techniques like GMPLS and=20 > > (GVRP and STB) work together?=20 > >=20 > > Regards=20 > >=20 > > Juergen=20 > >=20 > >=20 > > > -----Original Message-----=20 > > > From: owner-ccamp@ops.ietf.org=20 > > > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson=20 > > > Sent: Friday, July 22, 2005 2:53 PM=20 > > > To: richard.spencer@bt.com=20 > > > Cc: per@defero.se; ccamp@ops.ietf.org=20 > > > Subject: Re: Frameformat in a l2cs gmpls rnvironment.=20 > > >=20 > > > Richard,=20 > > >=20 > > > I agree to most of this. Since we have doubts about the viability=20 > > > of taking GMPLS all the way to end-user or enterprise I think it=20 > > > would be good, from a wg perspective, if we agreed to solve the=20 > > > core network problems first.=20 > > >=20 > > > Do you have any comment on the requirement to run both VLANs and=20 > > > GMPLS on the same switch?=20 > > >=20 > > > /Loa=20 > > >=20 > > > richard.spencer@bt.com wrote:=20 > > > > Regardless of whether or not a switch is directly connected=20 > > > > to hosts, it must be able to forward packets using the=20 > > > connectionless=20 > > > > Ethernet data plane. This is due to the fundamental=20 > > > requirement that=20 > > > > for GMPLS switches to be able to exchange control=20 > > information with=20 > > > > each other, a data plane for control traffic must be present.=20 > > > > This is akin to using the IP data plane for MPLS=20 > > signalling in an=20 > > > > IP/MPLS network. An alternative would be to use a static=20 > > reserved=20 > > > > L2-LSP for control traffic in the same way that reserved=20 > > VPI/VCIs=20 > > > > are used for PNNI signalling in ATM.=20 > > > >=20 > > > > Regarding connecting hosts to GMPLS switches, I=20 > personally don't=20 > > > > think extending L2-LSPs into the enterprise/home network is=20 > > > > commercially viable. However, if you do want to use=20 > > GMPLS switches=20 > > > > in the home/enterprise network and for some reason=20 > don't want to=20 > > > > extend L2-LSPs down to the host then you will not be=20 > performing=20 > > > > normal Ethernet Mac address switching anyway. Instead you will=20 > > > > need some kind of policy on the switch that maps=20 > connectionless=20 > > > > Ethernet packets (e.g. based on MAC src/dest, 802.1p,=20 > VLAN) to a=20 > > > > L2-LSP. This is because multiple L2-LSPs to the same=20 > destination=20 > > > > (e.g. a gateway router) may exist for different services/flows=20 > > > > (e.g. video download, VoIP call, etc.).=20 > > > >=20 > > > > Regards,=20 > > > > Richard=20 > > > >=20 > > > >=20 > > > >>-----Original Message-----=20 > > > >>From: owner-ccamp@ops.ietf.org=20 > > [mailto:owner-ccamp@ops.ietf.org]On=20 > > > >>Behalf Of Par Mattsson=20 > > > >>Sent: 22 July 2005 11:42=20 > > > >>To: Loa Andersson=20 > > > >>Cc: ccamp@ops.ietf.org=20 > > > >>Subject: Re: Frameformat in a l2cs gmpls rnvironment.=20 > > > >>=20 > > > >>=20 > > > >>=20 > > > >>>Per and Dimitri,=20 > > > >>>=20 > > > >>>I would like to come down stronger than that, for me it is=20 > > > >>>a very strong requirement that the same switch can handle=20 > > > >>>both VLANs and GMPLs trafic correctly. I can't dsee how that=20 > > > >>>could be done if using the VLAN tpid to indicate GMPLS=20 > > > >>>traffic.=20 > > > >>=20 > > > >>If you ever want that same switch to handle traffic for a=20 > > directly=20 > > > >>connected host (not to uncommen) you would want that to=20 > > use normal=20 > > > >>ethernet macaddress switching. So of course you do not want=20 > > > to have to=20 > > > >>choose between vlan and gmpls, you would want both at the=20 > > same time.=20 > > > >>=20 > > > >>/per=20 > > > >>=20 > > > >>=20 > > > >>=20 > > > >>>/Loa=20 > > > >>>=20 > > > >>>Par Mattsson wrote:=20 > > > >>>=20 > > > >>>>>hi par, one of the possibilities that has been considered=20 > > > >>=20 > > > >>to cope with=20 > > > >>=20 > > > >>>>>this requirement is to use a dedicated TPID for the=20 > > > >>=20 > > > >>Ethernet labeled=20 > > > >>=20 > > > >>>>>frames; this would allow differentiated processing with=20 > > > non-labeled=20 > > > >>>>>framesthanks.=20 > > > >>>>=20 > > > >>>>=20 > > > >>>>That seems to make more sence. If that frame is to be=20 > > sized like a=20 > > > >>>>802.1q=20 > > > >>>>frame. There is not that much space left to a label. Or is=20 > > > >>=20 > > > >>the demand to=20 > > > >>=20 > > > >>>>use jumboframes ?=20 > > > >>>>Has there been any discussion on labelstacking, and=20 > > > mainly where to=20 > > > >>>>place=20 > > > >>>>the information?=20 > > > >>>>=20 > > > >>>>Regards.=20 > > > >>>>Per=20 > > > >>>>=20 > > > >>>>=20 > > > >>>>=20 > > > >>>=20 > > > >>>=20 > > > >>>--=20 > > > >>>Loa Andersson=20 > > > >>>=20 > > > >>>Principal Networking Architect=20 > > > >>>Acreo AB phone: +46 8 632 77 14=20 > > > >>>Isafjordsgatan 22 mobile: +46 739 81 21 64=20 > > > >>>Kista, Sweden email: =20 > > loa.andersson@acreo.se=20 > > > >>> loa@pi.se=20 > > > >>>=20 > > > >>=20 > > > >>=20 > > > >>=20 > > > >=20 > > >=20 > > >=20 > > > --=20 > > > Loa Andersson=20 > > >=20 > > > Principal Networking Architect=20 > > > Acreo AB phone: +46 8 632 77 14=20 > > > Isafjordsgatan 22 mobile: +46 739 81 21 64=20 > > > Kista, Sweden email: loa.andersson@acreo.se=20 > > > loa@pi.se=20 > > >=20 > >=20 > >=20 > >=20 > >=20 >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 21:38:37 +0000 To: <richard.spencer@bt.com> Cc: <juergen.heiles@siemens.com>, <jaihyung@etri.re.kr>, <per@defero.se>, <ccamp@ops.ietf.org> From: Dimitri.Papadimitriou@alcatel.be Subject: RE: Frameformat in a l2cs gmpls rnvironment. MIME-Version: 1.0 Date: Mon, 25 Jul 2005 23:37:01 +0200 Message-ID: <OF2EC1F9FD.8BDA706E-ONC1257049.0076BE72-C1257049.0076BEE8@netfr.alcatel.fr> Content-type: text/plain; charset=us-ascii hi richard > I fully agree with Juergen's comments. The key question that's still not > clear in my mind here is, what is the real motivation for this work? motivations are detailed in section 4 - this said, i need some clarification from your side to provide a more detailed answer to your initial question > Is it to simply support dynamic VLAN provisioning for Ethernet based on > shortest/TE paths using GMPLS instead of using OSS? among other - yes > Or is it to develop an alternative CO-PS solution (to say ATM or MPLS) > using the Ethernet frame structure, regardless of whether or not the > Ethernet forwarding behaviour (and hardware) needs to be modified? this is not an objective however, and this is where the whole difficulty of the exercise stands, as the DT was not chartered to specifically detail the data plane behaviour/fowarding behaviour there is indeed a whole spectrum of such behaviour that can sustain these requirements - reason why this work item (even if we get a consensus as part of this WG) requires consultation of the IEEE as stated in section 7.1 > If the scope of the proposal does not prevent the Ethernet forwarding > behaviour being modified resulting in "the switching hardware, the filter > and relay, etc. are totally different new dataplane switch", then this seems > to be contradictory to an earlier comment from Dimitri that said "the > definition of any data plane behaviour is outside of the scope of the DT > charter and the document it produced". and this statement still holds as the DT work is concerned - note: some individuals have expressed opinions in terms of data plane support but this is not part of the scope of the present DT charter but these is a logical consequence of initiating such a work item within a community of engineers > This also raises questions regarding the role of the IEEE and what the > applications/goals/benefits of the new Ethernet forwarding behaviour will > be. indeed - this is the reason why the document has explicitly stated that consultation with the IEEE must be ensured to progress with this item > If on the other hand the goal is to use the Ethernet data plane as it stands > today based on MAC address learning/switching, but to replace manual/OSS > VLAN provisioning with a GMPLS CP, it seems to me that there is a > fundamental difference between this proposal and all other GMPLS solutions. > In all the GMPLS solutions I'm aware of, the label signalled via GMPLS (e.g. > ATM VPI/VCI, WDM wavelength, etc.) is also used for forwarding in the data > plane. However in this case, data plane forwarding is not based on a label > (VLAN ID) but on an address, i.e. the destination MAC address. Instead, the > label is used to restrict the broadcast domain (e.g. to a P2P path), which > (following MAC flooding/learning) results in the packets belonging to a > particular VLAN following the same path. Although the result is effectively > the same, the fact that forwarding is not based on the label signalled using > GMPLS may cause complications. For example, when a failure occurs, in > addition to re-routing or switching over to backup paths, the MAC > databases/tables of all effected nodes must be purged on a per VLAN basis. indeed, flushing and re-learning aspects must be addressed - and this is the kind of feedback we expect to receive from the IEEE > Looking at section 4, I agree that one obvious benefit (as mentioned in > section 4) would be the ability to dynamically calculate loop-free > shortest/TE paths, which would allow operators to switch off spanning tree. > However as Juergen points out, this is one of the goals for the TRILL WG > (although the solution proposed is different), so I would like to understand > what the common/opposing goals are. Another example from section 4 is the > point about improving scalability by introducing IP addressing with GMPLS > rather relying on MAC addresses, which are not hierarchical. This point is > only valid if the control plane is being used to carry Ethernet addresses. i do not understand this last statement - would you clarify ? - > So, this might be a valid point when comparing the TRILL solution to the > GMPLS solution for example, but is not valid when comparing a standard 802.1 > Ethernet network to a GMPLS L2-LSP network. The reason being that both > solutions are still based on data plane learning using MAC addresses and are > therefore both subject to the same data plane scaling issues, i.e. for the > solution to scale to large numbers of Ethernet customers, MAC-in-MAC > encapsulation will be required. indeed MAC-in-MAC is currently under definition to solve such kind of problems, however the present approach does not intend to be used for establishing an LSP per pair of MAC SA-DA when the #Ethernet terminating hosts >> #network nodes; also to clarify a specific point in the context of the present discussion, this work item is not intended to solve ALL the specific issues and problems related to the use of an Ethernet data plane in any operational conditions; this document has simply identified several scenario where the use of GMPLS protocol suite may be considered in order to facilitate (some of the) Ethernet network operations > Therefore, what will the benefits of using L2-LSPs with MAC-in-MAC be > compared to encapsulating Ethernet in MPLS PWs? some of the TRILL WG objectives are indeed overlapping with the present effort however the respective approaches are not identical; TRILL does not intend to make use of any signaling mechanism/source-routing protocol; TRILL is intended to specifically cover "CAMPUS" networks (even if its scope could be larger) by defining an "intermediate" technology between IP and Ethernet to circumvent the drawbacks of the latter (in particular wrt to 802.1d and related) concerning the differences between the present work item and Ethernet over PW/MPLS, i refer you to the introduction of the document under discussion i hope this clarifies (at least some of your doubts) thanks, - dimitri. Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 19:13:14 +0000 To: Loa Andersson <loa@pi.se> Cc: Dimitri.Papadimitriou@alcatel.be, =?iso-8859-1?Q?P=E4r_Mattsson?= <per@defero.se>, ccamp@ops.ietf.org MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: Re: dimitris question on VLAN vs label (L2SC) Date: Mon, 25 Jul 2005 21:11:53 +0200 Message-ID: <OFC60B8493.8AD6DC8D-ONC1257049.0069747C-C1257049.00697590@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmxvYSA8L0ZPTlQ+PC9QPjxQPjxGT05U IEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj4tIG9rIC0ganVzdCB3YW50ZWQgdG8gYmUgc3VyZSB0 aGF0IHRoaXMgcmVxdWlyZW1lbnQgaXMgY29ycmVjdGx5IHVuZGVyc3Rvb2Q8L0ZPTlQ+PC9QPjxQ PjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj50aGFua3MsPC9GT05UPjwvUD48UD48Rk9O VCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+LSBkaW1pdHJpLjwvRk9OVD48QlI+PEJSPjxGT05U IFNJWkU9Mj48Qj5Mb2EgQW5kZXJzc29uICZsdDtsb2FAcGkuc2UmZ3Q7PC9CPjwvRk9OVD48QlI+ PEZPTlQgU0laRT0yPjA3LzI1LzIwMDUgMTU6MDYgWkUyPC9GT05UPjxCUj48QlI+IDxGT05UIFNJ WkU9Mj5Ubzo8L0ZPTlQ+IDxGT05UIFNJWkU9Mj5EaW1pdHJpIFBBUEFESU1JVFJJT1UvQkUvQUxD QVRFTEBBTENBVEVMPC9GT05UPjxCUj4gPEZPTlQgU0laRT0yPmNjOjwvRk9OVD4gPEZPTlQgU0la RT0yPlAmYXVtbDtyIE1hdHRzc29uICZsdDtwZXJAZGVmZXJvLnNlJmd0OywgY2NhbXBAb3BzLmll dGYub3JnPC9GT05UPjxCUj4gPEZPTlQgU0laRT0yPmJjYzo8L0ZPTlQ+IDxCUj4gPEZPTlQgU0la RT0yPlN1YmplY3Q6PC9GT05UPiA8Rk9OVCBTSVpFPTI+ZGltaXRyaXMgcXVlc3Rpb24gb24gVkxB TiB2cyBsYWJlbCAoTDJTQyk8L0ZPTlQ+PEJSPiA8QlI+PEJSPjwvUD48UD48Rk9OVCBGQUNFPSJN b25vc3BhY2UsQ291cmllciI+RGltaXRyaSw8QlI+PC9GT05UPjxCUj48Rk9OVCBGQUNFPSJNb25v c3BhY2UsQ291cmllciI+dGhlIHNpbXBsZSBhbnN3ZXIgaXMgdGhhdCBJIG5lZWQgbXkgVkxBTnMs IHRoZXkgZ2l2ZSBtZSBhPEJSPmdvb2QsIHVuZGVyc3Rvb2QgYW5kIHJlbGlhYmxlIHNlcnZpY2Us IGV2ZW4gd2hlbiBJIGNhbiBzZXQgdXA8QlI+TFNQIHRocm91Z2ggYW4gRXRoZXJuZXQgbmV0d29y ay48QlI+PC9GT05UPjxCUj48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+QXMgSSBzZWUg aXQgVkxBTnNkIGFuZCBMU1BzIGlzIG5vdCB0aGUgc2FtZSwgaWYgd2UgdXNlIHRoZTxCUj5WTEFO IHRvIGVuY2Fwc3VsYXRlIExTUHMgd2UgZ2V0IGludG8gYSBwcm9ibGVtIG9mIGRpc3Rpbmd1aXNo aW5nPEJSPmJldHdlZW4gYSBWTEFOIGFuZCBhbiBMU1AuPEJSPjwvRk9OVD48QlI+PEZPTlQgRkFD RT0iTW9ub3NwYWNlLENvdXJpZXIiPlZMQU4gaXMgYSBzY2FyY2UgcmVzb3VyY2UsIDIwNDggcGVy IG5ldHdvcmsgKHllYWggLSBrbm93IEkgY2FuPEJSPmluY3JlYXNlIGl0IGJ5IG1hbmFnaW5nIHRo ZSBWTEFOIGJ1ZGdldCBjYXJlZnVsbHksIGJ1dCBpdCBpczxCUj5raW5kIG9mIGF3a3dhcmQpLiBJ IHRoaW5rIHdlIHNob3VsZCBhaW0gZm9yIHNvbWV0aGluZyBsaWtlIDIwMDA8QlI+bGFiZWxzIHBl ciBsaW5rIChJIHdvdWxkbid0IG9iamVjdCBpZiB3ZSBjYW4gZG8gYmV0dGVyKS48QlI+PC9GT05U PjxCUj48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+SSdkIGFsc28gc2F5IHRoYXQgdGhl IHRoZSBFdGhlcm5ldCBuZXR3b3JrIG5lZWRzIHRvIGJlIGFibGU8QlI+dG8gaGFuZGxlIGFsbCB0 eXBlcyBvZiB0cmFmZmljLCBnbXBscywgdGFnZ2VkIGFuZCB1bnRhZ2dlZDxCUj5vdmVyIHRoZSBz YW1lIHN3dGljaC4gRG9uJ3Qgc2VlIHRoYXQgYSBpbnRlcmZhY2Ugc2VwYXJhdGlvbiB3b3VsZDxC Uj5oZWxwIHNpbmNlIFZMQU4gYXJlIHBlciBwbGF0Zm9ybS48QlI+PC9GT05UPjxCUj48Rk9OVCBG QUNFPSJNb25vc3BhY2UsQ291cmllciI+RGlkIEkgYW5zd2VyIHlvdXIgcXVlc3Rpb25zPzxCUj48 L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj4vTG9hPEJSPjwvRk9OVD48 QlI+PEJSPjxCUj48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+RGltaXRyaS5QYXBhZGlt aXRyaW91QGFsY2F0ZWwuYmUgd3JvdGU6PEJSPiZndDsgbG9hLDxCUj4mZ3Q7PEJSPiZndDsgZHVy aW5nIHRoZSBkaXNjdXNzaW9uIGxvb3Agb25lIG9mIHRoZSBwb3NzaWJpbGl0aWVzIHRoYXQgY2Ft ZSBvdXQ8QlI+Jmd0OyBjb25zaXN0ZWQgaW4gdXNpbmcgYSBkZWRpY2F0ZWQgVFBJRCB2YWx1ZSAo ZGlmZmVyZW50IGZyb20gdGhlIHZhbHVlPEJSPiZndDsgMHg4MTAwIHVzZWQgaW4gLjFxKSwgdGhh dCB3b3VsZCBwcmVjZWRlIHRoZSBlbmNvZGVkIHZhbHVlIGFzIHBhcnQgb2YgdGhlPEJSPiZndDsg VENJPEJSPiZndDs8QlI+Jmd0OyB3b3VsZCB5b3UgcGxlYXNlIGNsYXJpZnkgeW91ciBzcGVjaWZp YyBjb25jZXJuID8gaSB3b3VsZCBsaWtlIGFsc28gdG88QlI+Jmd0OyBwb2ludCBvdXQgdGhhdCB0 aGVyZSBhcmUgc2V2ZXJhbCBjYXNlcyBoZXJlPEJSPiZndDs8QlI+Jmd0OyAxLiBib3RoIHR5cGUg b2YgdHJhZmZpYyBvbiB0aGUgc2FtZSBzd2l0Y2ggYnV0IG5vdCBuZWNlc3NhcmlseSBvbiB0aGU8 QlI+Jmd0OyBzYW1lIGludGVyZmFjZTxCUj4mZ3Q7PEJSPiZndDsgMi4gYm90aCB0eXBlIG9mIHRy YWZmaWMgb24gdGhlIHNhbWUgaW50ZXJmYWNlPEJSPiZndDs8QlI+Jmd0OyB3aXRoIHRoZSBmb2xs b3dpbmcgdmFyaWFudDxCUj4mZ3Q7PEJSPiZndDsgMmEuIGludGVyZmFjZSBtdXN0IGFjY29tbW9k YXRlIGFueSB0eXBlIG9mIHRyYWZmaWMgKHVuLS90YWdnZWQgYW5kIGxhYmVsZWQpPEJSPiZndDs8 QlI+Jmd0OyAyYi4gaW50ZXJmYWNlIG11c3QgYWNjb21tb2RhdGUgdGFnZ2VkIGFuZCBsYWJlbGVk IHR5cGUgb2YgdHJhZmZpYyBvbmx5PEJSPiZndDs8QlI+Jmd0OyB0aGFua3MsPEJSPiZndDs8QlI+ Jmd0OyAtIGRpbWl0cmkuPEJSPiZndDs8QlI+Jmd0OyAqTG9hIEFuZGVyc3NvbiAmbHQ7bG9hQHBp LnNlJmd0Oyo8QlI+Jmd0OyBTZW50IGJ5OiBvd25lci1jY2FtcEBvcHMuaWV0Zi5vcmc8QlI+Jmd0 OyAwNy8yMi8yMDA1IDEyOjA0IFpFMjxCUj4mZ3Q7PEJSPiZndDsgVG86IFAmYXVtbDtyIE1hdHRz c29uICZsdDtwZXJAZGVmZXJvLnNlJmd0OzxCUj4mZ3Q7IGNjOiBjY2FtcEBvcHMuaWV0Zi5vcmc8 QlI+Jmd0OyBiY2M6PEJSPiZndDsgU3ViamVjdDogUmU6IEZyYW1lZm9ybWF0IGluIGEgbDJjcyBn bXBscyBybnZpcm9ubWVudC48QlI+Jmd0OzxCUj4mZ3Q7PEJSPiZndDsgUGVyIGFuZCBEaW1pdHJp LDxCUj4mZ3Q7PEJSPiZndDsgSSB3b3VsZCBsaWtlIHRvIGNvbWUgZG93biBzdHJvbmdlciB0aGFu IHRoYXQsIGZvciBtZSBpdCBpczxCUj4mZ3Q7IGEgdmVyeSBzdHJvbmcgcmVxdWlyZW1lbnQgdGhh dCB0aGUgc2FtZSBzd2l0Y2ggY2FuIGhhbmRsZTxCUj4mZ3Q7IGJvdGggVkxBTnMgYW5kIEdNUExz IHRyYWZpYyBjb3JyZWN0bHkuIEkgY2FuJ3QgZHNlZSBob3cgdGhhdDxCUj4mZ3Q7IGNvdWxkIGJl IGRvbmUgaWYgdXNpbmcgdGhlIFZMQU4gdHBpZCB0byBpbmRpY2F0ZSBHTVBMUzxCUj4mZ3Q7IHRy YWZmaWMuPEJSPiZndDs8QlI+Jmd0OyAvTG9hPEJSPiZndDs8QlI+Jmd0OyBQJmF1bWw7ciBNYXR0 c3NvbiB3cm90ZTo8QlI+Jmd0OyAmbmJzcDsmZ3Q7Jmd0O2hpIHBhciwgb25lIG9mIHRoZSBwb3Nz aWJpbGl0aWVzIHRoYXQgaGFzIGJlZW4gY29uc2lkZXJlZCB0byBjb3BlIHdpdGg8QlI+Jmd0OyAm bmJzcDsmZ3Q7Jmd0O3RoaXMgcmVxdWlyZW1lbnQgaXMgdG8gdXNlIGEgZGVkaWNhdGVkIFRQSUQg Zm9yIHRoZSBFdGhlcm5ldCBsYWJlbGVkPEJSPiZndDsgJm5ic3A7Jmd0OyZndDtmcmFtZXM7IHRo aXMgd291bGQgYWxsb3cgZGlmZmVyZW50aWF0ZWQgcHJvY2Vzc2luZyB3aXRoIG5vbi1sYWJlbGVk PEJSPiZndDsgJm5ic3A7Jmd0OyZndDtmcmFtZXN0aGFua3MuPEJSPiZndDsgJm5ic3A7Jmd0OzxC Uj4mZ3Q7ICZuYnNwOyZndDs8QlI+Jmd0OyAmbmJzcDsmZ3Q7IFRoYXQgc2VlbXMgdG8gbWFrZSBt b3JlIHNlbmNlLiBJZiB0aGF0IGZyYW1lIGlzIHRvIGJlIHNpemVkIGxpa2UgYSA4MDIuMXE8QlI+ Jmd0OyAmbmJzcDsmZ3Q7IGZyYW1lLiBUaGVyZSBpcyBub3QgdGhhdCBtdWNoIHNwYWNlIGxlZnQg dG8gYSBsYWJlbC4gT3IgaXMgdGhlIGRlbWFuZCB0bzxCUj4mZ3Q7ICZuYnNwOyZndDsgdXNlIGp1 bWJvZnJhbWVzID88QlI+Jmd0OyAmbmJzcDsmZ3Q7IEhhcyB0aGVyZSBiZWVuIGFueSBkaXNjdXNz aW9uIG9uIGxhYmVsc3RhY2tpbmcsIGFuZCBtYWlubHkgd2hlcmUgdG8gcGxhY2U8QlI+Jmd0OyAm bmJzcDsmZ3Q7IHRoZSBpbmZvcm1hdGlvbj88QlI+Jmd0OyAmbmJzcDsmZ3Q7PEJSPiZndDsgJm5i c3A7Jmd0OyBSZWdhcmRzLjxCUj4mZ3Q7ICZuYnNwOyZndDsgUGVyPEJSPiZndDsgJm5ic3A7Jmd0 OzxCUj4mZ3Q7ICZuYnNwOyZndDs8QlI+Jmd0OyAmbmJzcDsmZ3Q7PEJSPiZndDs8QlI+Jmd0OzxC Uj4mZ3Q7IC0tPEJSPiZndDsgTG9hIEFuZGVyc3NvbjxCUj4mZ3Q7PEJSPiZndDsgUHJpbmNpcGFs IE5ldHdvcmtpbmcgQXJjaGl0ZWN0PEJSPiZndDsgQWNyZW8gQUIgJm5ic3A7ICZuYnNwOyAmbmJz cDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNw OyAmbmJzcDsgJm5ic3A7IHBob25lOiAmbmJzcDsrNDYgOCA2MzIgNzcgMTQ8QlI+Jmd0OyBJc2Fm am9yZHNnYXRhbiAyMiAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAm bmJzcDsgJm5ic3A7ICZuYnNwO21vYmlsZTogKzQ2IDczOSA4MSAyMSA2NDxCUj4mZ3Q7IEtpc3Rh LCBTd2VkZW4gJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7 ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwO2VtYWlsOiAmbmJzcDtsb2EuYW5kZXJzc29uQGFj cmVvLnNlPEJSPiZndDs8QlI+Jmd0OyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7 ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsg Jm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAm bmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZu YnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5i c3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyBsb2FA cGkuc2U8QlI+PC9GT05UPjxCUj48QlI+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPi0t PEJSPkxvYSBBbmRlcnNzb248QlI+PC9GT05UPjxCUj48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291 cmllciI+UHJpbmNpcGFsIE5ldHdvcmtpbmcgQXJjaGl0ZWN0PEJSPkFjcmVvIEFCICZuYnNwOyAm bmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZu YnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyBwaG9uZTogJm5ic3A7KzQ2IDggNjMyIDc3IDE0PEJS PklzYWZqb3Jkc2dhdGFuIDIyICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5i c3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7bW9iaWxlOiArNDYgNzM5IDgxIDIxIDY0PEJSPktpc3Rh LCBTd2VkZW4gJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7 ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwO2VtYWlsOiAmbmJzcDtsb2EuYW5kZXJzc29uQGFj cmVvLnNlPC9GT05UPjwvUD48VUw+PFVMPjxVTD48VUw+PFVMPjxVTD48VUw+PFVMPjxVTD48VUw+ PFVMPjxVTD48VUw+PFVMPjxVTD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+bG9hQHBp LnNlPC9GT05UPjwvVUw+ Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 16:18:43 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: RE: Moving LSP ownership between control plane and management plane Date: Mon, 25 Jul 2005 18:18:11 +0200 Message-ID: <7DBAFEC6A76F3E42817DF1EBE64CB02602718C6A@FTRDMEL2.rd.francetelecom.fr> Thread-Topic: Moving LSP ownership between control plane and management plane Thread-Index: AcWNDydF04j4tl/eRIW5sP2WNO2RbQEIyTNA From: "MEURIC Julien RD-CORE-LAN" <julien.meuric@francetelecom.com> To: "Diego Caviglia" <Diego.Caviglia@marconi.com> Cc: "ccamp" <ccamp@ops.ietf.org> Hello Diego, Adrian and all. In my view, I feel that moving LSP ownership between MP and CP are very useful functions for a carrier operational network, whether it be for migration to GMPLS or for network maintenance. As far as I understand, the I-D is a proposal for the signalling part and does not pretend solving every associated issues. Some work will probably be needed to implement the whole process, but this part looks interesting to me. I would be glad to discuss further this matter next week in Paris. Regards Julien=20 Julien Meuric France Telecom Research & Development Core Network 2, avenue Pierre Marzin 22307 Lannion Cedex - France Tel : + 33 (0)2 96 05 28 28 Fax : + 33 (0)2 96 05 12 52 http://www.francetelecom.com/rd/ -----Original Message----- From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Diego Caviglia Hi Geroge, sorry but I don't see how the management effort is something related to this ID and more in general with CCAMP. There are already RFC ID that provide tools to signal LSP with ERO filled up to the Label level and of course this can be done only via management, I don't remeber any objections on that due to NMS impact. I'd like to discuss here (I mean in the CCAMP ML) about the ID, that is, if it useful and if could work and not about the difficulties it implies for the NMS. I'm open to discuss, and actually I did, NMS related issues on the ITU-T ML. Regards Diego George Newsome <gnewsome@ieee.org>@ops.ietf.org on 20/07/2005 11.03.00 Sent by: owner-ccamp@ops.ietf.org To: Adrian Farrel <adrian@olddog.co.uk> cc: ccamp@ops.ietf.org, Neil Harrison <neil.2.harrison@bt.com>, Dino Bramanti <Dino.Bramanti@marconi.com> Subject: Re: Moving LSP ownership between control plane and management plane Adrian, Dino, etc., This thread has rapidly got too long, and I don't have the time at the moment to argue the toss, so I'll reply as a note rather than intersperse remarks. On Adrian's question of which operators were at Holmdel - AT&T and T-Systems. On do operators think they want this - some do some don't. BT clearly don't see much value in fiddling with signaling. You ask whether this might be done in service, and there are certainly those who would like to see just that. Related questions are how many times in the life of the system might this be done for the same connection? Is it cost effective to develop a highly reliable solution that is used just once in the life of the system? However this is not so much a discussion of whether this could be useful, but a technical discussion of whether changing a bit (and a slew of procedures) in a signaling protocol actually provides the claimed facility. Because of the effort the management system needs to provide before signaling might be involved, changing the signaling protocol gives the illusion of success, yet does not provide the claimed service at all. Once the management systems have got as far as assigning control plane names, we can ask what happens to the cross connect object when it is deleted as part of the hand off procedure? The behavior of this object is that its deletion results in dropping the matrix connection, which is exactly not required during the change over. It is my considered opinion that fiddling with signaling merely obfuscates the significant difficulties, which are glossed over by stating that those difficulties occur at a different time. Regards George Adrian Farrel wrote: > Thanks George, > > Useful input. > > >>You asked for comments on this proposal, which was presented to ITU-T >>Q14 experts in January of this year. > > > Can you confirm that there were plenty of providers represented at the > meeting. > > >>There are so many issues involved in moving a connection from the >>management plane to the control plane that involve intensive management >>action, that it is not evident that signaling need be involved at all. > > > Interesting. It is certainly valuable to have a list of these issues out > in the open. I have made some comments in the appended meeting report. > > If the control plane is to take over the connection then we assume that > signaling state must be built so that the control plane could then modify > or tear down the connection. how would you propose that this be done > without involving signaling? > > >>As a personal opinion, it makes little sense to consider moving the >>connection. It makes more sense to replace the connection with one set >>up via the control plane, and to then delete the original management set >>up connection. > > > The question arrises of how this would be done in a network with limited > resources. That is, if it is possible to set up a parallel conneciton then > the method you describe would be fine, but if there are not enough > parallel resources then would you propose that a lambda (for example) > could be used for the management-established connection and for the > control-plane-owned connection at the same time? This might present some > difficulties as well. > > Do you have an opinion on my main question which is whether this tyoe of > change-over would ever be done in service? > > Thanks, > Adrian > > >>Excerpt of the meeting report of Q14 Experts, January 2005 >> >>WD06 (Marconi) "G.7713 Modification in order to support Circuit >>Migration" contains a modified version of draft G.7713 Revision as was >>presented in the Berlin meeting (WD 11, i.e., TD50/3 Dec.2004). This >>contribution addresses SP24 (PC ? SPC) of G.7713. It proposes to extend >>the concept of Call/Connection setup to setup/adopt and release to >>release/de-adopt. In Connection Adoption both the SNC are not actually >>created and the LC are not actually established due to the fact that the >>underlying physical resources are already in place. In the >>Call/Connection dis-adoption SNC-LC-SNC is not actually de-allocated, >>only the Control Plane information associated with then is removed and >>the ownership is moved to Management Plane. Attributes for indicating >>Adoption/Dis-adoption are proposed for the INNI Connection messages. >> >>The discussion was mainly on the Management plane and Control plane >>actions required before signalling is involved. The group noted that the >>following issues need to be addressed: >> >>- Naming of transport resources to the control plane. Before a call can >>be placed under signalling control, links that are involved in the >>connection need to be given control plane names. Without these, no >>explicit route can be formed to match the resources of the connection. > > > This is pretty obvious. But actually, this is an issue with moving the > resources from one plane to another (or having them present in both > planes). It is certainly a prerequisite for moving the connection, but it > is not functionally an element of moving the connection. > > The draft, as I read it, assumes that these operations have been completed > and then asks "how do I move the connection from one plane to the other?" > > >>- Resource state while under dual CP and MP views. After resources are >>given control plane names, the resources are still viewed by the >>management plane. It may be necessary to create a new state for the >>resource to indicate that the management plane cannot perform some >>actions on the connection points as the control plane is about to take > > over. > > Yes. There is certainly scope for a race condition here. I believe that > the I-D proposes to (try to) make the transfer an atomic action from the > point of view of the change-over. > > Note, however, that it is wrong to assume that a resource belongs only to > the control or to the management plane at any moment in time. This may be > true with respect to the ability to provision/control connections, but > resources that are available for use by the control plane can still be > managed through the management plane. Thus, you take a laser out of > service through the management plane, not through the control plane. > > >>- Role of discovery functions (esp. CELA). After control plane names >>are allocated, distributed control plane functions need to be associated >>and communication adjacencies formed. This too is a precursor to any >>signalling procedures in migrating from the MP to CP. > > > Note to CCAMP: CELA is Control Entity Logical Adjacency. > This too is "obvious". > But this discussion seems to be wide of the mark. Before deploying a > control plane, it is clear that this type of operation needs to take > place. Since (within the context of CCAMP) a control plane *is* going to > be deployed, we can assume this operation and move on to the discussion in > the draft. > > >>- Similarity of migration to synchronization after CP failure and >>subsequent recovery. Connection control procedures might be the same in >>migrating a call from the MP to the CP as the situation when the CP has >>failed and is recovering. Here, the network connections are already in >>place, but connection and call state need to be created to match it. >>Knowledge of the call and what the connection should be is obtained from >>the MP for migration, and from some reliable database in the recovery > > case. > > This is a valid point. > Control plane failure, however, is unlikely to be total. That is, it would > be unusual for more than one control plane link or node to fail. Thus, in > the control plane, the "reliable database" is distributed and recovery can > follow the procedures outlined in draft-ietf-ccamp-rsvp-restart-ext-03.txt > > It would certainly be worth considering whether the procedures used for > control plane recovery could be applied to transfer of ownership from the > management to control plane. at the moment, I suspect that they could not > because the restart extensions assume a specific period during which > signaling is associated with recovery. > > We would also need to look at the reverse process (transfer from CP to MP) > that is not possible to cover by CP restart procedures. > > >>- Call awareness of migration vs. connection being unaware of migration. >> When connection state is being created to match an existing >>connection, the connection controllers (CCs) do not require awareness of >>why this is happening as the context could be migration or recovery. A >>mechanism to create control state without affecting transport plane >>state is needed in the CCs. Call controllers though do need migration >>awareness as they need to obtain/derive call identification from the MP. > > > I disagree. > Connection controllers must be able to distinguish between resources that > are already in use and cannot be assigned to a new connection, and those > that are already in use and can be assigned to a new connection because > the connection is replacing an existing connection. If this were happening > entirely within the control plane we would use resource sharing, but when > the operation spans the control and management planes resource > ownership/sharing is more complex and the conneciotn controller needs to > be able to issue the approriate instrucitons for the assignment of the > resources. > On the other hand, call controllers do not need awareness of the details > of the realisation of the connections. > > >>In summary, the group decided that it is premature to consider >>signalling procedure before the above issues, amongst other, have been >>studied. However the contribution does enable the group to expose in a >>larger context the interaction between CP and MP for connection >>migration. The above issues were included in SP24 of the G.7713 Living >>List (see WD22). > > > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 15:46:21 +0000 Content-Class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Mon, 25 Jul 2005 16:44:55 +0100 Message-ID: <B5E87B043D4C514389141E2661D255EC0A835AEC@i2km41-ukdy.domain1.systemhost.net> Thread-Topic: Frameformat in a l2cs gmpls rnvironment. Thread-Index: AcWQ/aViET2i5RVMS6exO5FlWh/ssgAAKGwg From: <richard.spencer@bt.com> To: <juergen.heiles@siemens.com>, <jaihyung@etri.re.kr>, <per@defero.se> Cc: <ccamp@ops.ietf.org> I fully agree with Juergen's comments. The key question that's still not = clear in my mind here is, what is the real motivation for this work? Is = it to simply support dynamic VLAN provisioning for Ethernet based on = shortest/TE paths using GMPLS instead of using OSS? Or is it to develop = an alternative CO-PS solution (to say ATM or MPLS) using the Ethernet = frame structure, regardless of whether or not the Ethernet forwarding = behaviour (and hardware) needs to be modified? If the scope of the proposal does not prevent the Ethernet forwarding = behaviour being modified resulting in "the switching hardware, the = filter and relay, etc. are totally different new dataplane switch", then = this seems to be contradictory to an earlier comment from Dimitri that = said "the definition of any data plane behaviour is outside of the scope = of the DT charter and the document it produced". This also raises = questions regarding the role of the IEEE and what the = applications/goals/benefits of the new Ethernet forwarding behaviour = will be.=20 If on the other hand the goal is to use the Ethernet data plane as it = stands today based on MAC address learning/switching, but to replace = manual/OSS VLAN provisioning with a GMPLS CP, it seems to me that there = is a fundamental difference between this proposal and all other GMPLS = solutions. In all the GMPLS solutions I'm aware of, the label signalled = via GMPLS (e.g. ATM VPI/VCI, WDM wavelength, etc.) is also used for = forwarding in the data plane. However in this case, data plane = forwarding is not based on a label (VLAN ID) but on an address, i.e. the = destination MAC address. Instead, the label is used to restrict the = broadcast domain (e.g. to a P2P path), which (following MAC = flooding/learning) results in the packets belonging to a particular VLAN = following the same path. Although the result is effectively the same, = the fact that forwarding is not based on the label signalled using GMPLS = may cause complications. For example, when a failure occurs, in addition = to re-routing or switching over to backup paths, the MAC = databases/tables of all effected nodes must be purged on a per VLAN = basis. Looking at section 4, I agree that one obvious benefit (as mentioned in = section 4) would be the ability to dynamically calculate loop-free = shortest/TE paths, which would allow operators to switch off spanning = tree. However as Juergen points out, this is one of the goals for the = TRILL WG (although the solution proposed is different), so I would like = to understand what the common/opposing goals are. Another example from = section 4 is the point about improving scalability by introducing IP = addressing with GMPLS rather relying on MAC addresses, which are not = hierarchical. This point is only valid if the control plane is being = used to carry Ethernet addresses. So, this might be a valid point when = comparing the TRILL solution to the GMPLS solution for example, but is = not valid when comparing a standard 802.1 Ethernet network to a GMPLS = L2-LSP network. The reason being that both solutions are still based on = data plane learning using MAC addresses and are therefore both subject = to the same data plane scaling issues, i.e. for the solution to scale to = large numbers of Ethernet customers, MAC-in-MAC encapsulation will be = required. Therefore, what will the benefits of using L2-LSPs with = MAC-in-MAC be compared to encapsulating Ethernet in MPLS PWs? Regards, Richard > -----Original Message----- > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On > Behalf Of Heiles Juergen > Sent: 25 July 2005 10:45 > To: CHO, JAI HYUNG; Heiles Juergen; per@defero.se > Cc: ccamp@ops.ietf.org > Subject: RE: Frameformat in a l2cs gmpls rnvironment. >=20 >=20 > Dear Jaihyung, >=20 > you mention several options that could be used as a label.=20 > Most of them have strong impacts on the data plane and=20 > introduce new layer networks. Is this in the scope of CCAMP?=20 > What is the goal, to introduce a control plane for a existing=20 > L2 technology or introduce a new L2 technology? What would be=20 > the benefit of such a new L2 co switching technology compared=20 > to Ethernet over MPLS as defined by PWE3 and L2VPN? > You mentioned to use the MAC address or part of it as a=20 > label. The TRILL working group is defining shortest path=20 > routing for Ethernet. How would this fit together? >=20 >=20 > Regards >=20 > Juergen > =20 >=20 >=20 >=20 >=20 > > -----Original Message----- > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr]=20 > > Sent: Saturday, July 23, 2005 6:19 AM > > To: Heiles Juergen; per@defero.se > > Cc: ccamp@ops.ietf.org > > Subject: RE: Frameformat in a l2cs gmpls rnvironment. > >=20 > >=20 > > Dear Juergen, Par and all, > >=20 > > The proposed framework of L2SC was not intended to suggest=20 > > any specific solution. It is still an open question.=20 > > As Adrian noted, it only aims to present requirements=20 > > as in scenarios that may drive people toward next step in many=20 > > viable solutions. > >=20 > > In fact, there are other options we may also consider for L2=20 > > label encoding.=20 > > IEEE 802.1 Ethernet bridges forward frames based on 48bits of=20 > > MAC address,=20 > > and additionally using VLAN tag.=20 > >=20 > > When the purpose of GMPLS control over Ethernet is not to=20 > > create new dataplane=20 > > but to utilize IEEE 802.1 bridge architecture, we may=20 > consider using=20 > > one of the two forwarding methods, MAC forwarding or VLAN=20 > forwarding.=20 > > (and perhaps any other combination of fields in MAC, but I'll=20 > > not discuss it) > > =20 > > Use of VLAN ID for label encoding may automate VLAN configuration=20 > > using IP protocols. However, GMPLS protocol cannot use the=20 > > field exclusively=20 > > because public/private operators already use VLAN for various=20 > > purpose.=20 > > There is a potential conflict with existing use of VLAN and=20 > > GMPLS use of VLAN label. =20 > > Furthermore, scalability of VLAN ID has been frequently noted=20 > > as weakness=20 > > because the size of VLAN ID is at most 4096 (12bit).=20 > >=20 > > The scalability may be improved if the scope of VLAN label is=20 > > confined to=20 > > link-local, and some additional swapping function of VLAN ID=20 > > is introduced in=20 > > Internal Sublayer [802.1D] of GMPLS implemented switch.=20 > > However, this will only be effective when the configuration=20 > > of network=20 > > is mesh structure that multiple LSP paths exist. If the=20 > > configuration of network=20 > > is star or tree shape, as normal configuration of access network,=20 > > LSPs concentrate in root node and total number of=20 > > LSPs that the network can hold still be limited by available=20 > > label space=20 > > at a few root links.=20 > >=20 > > For these reasons, I do not think any form of VLAN ID label is=20 > > an appropriate choice for layer-2 label encoding. > > =20 > > There are some other proposals assuming new assignment of Ethernet=20 > > Length/Type value (e.g. new TPID in VLAN tag) and re-definition of=20 > > information fields placed between 802.3 MAC header and IP packet.=20 > > In this case, only the format of VLAN tag or extended VLAN tag=20 > > is borrowed, however, inside the switching hardware, the=20 > > filter and relay,=20 > > etc. are totally different new dataplane switch. > >=20 > > I do not see such approach is a GMPLS implementation for Ethernet=20 > > because the core switching technique is not 802.1 Ethernet=20 > > bridge at all. > > =20 > > The other option we may consider is using MAC address filed=20 > as below. > > =20 > >=20 > > +-------+-------+-------+-------+-------+-------+=20 > > | 1byte | 2byte | 3byte | 4byte | 5byte | 6byte |=20 > > +-------+-------+-------+-------+-------+-------+ > > +-----------------------+-----------------------+=20 > > | OUI Prefix (=3DGMPLS) | DA-label (24bit) |=20 > > +-----------------------+-----------------------+=20 > > | OUI Prefix (=3DGMPLS) | SA-label (24bit) |=20 > > +---------------+-------+-----------------------+=20 > > | Length/Type |=20 > > +---------------+ > > =20 > >=20 > > IEEE is designated by the ISO Council to act as the=20 > > registration authority=20 > > for the higher three-octet of OUI number in the MAC address=20 > > to be used by manufacturer. Ethernet manufacturer may generate=20 > > global unique MAC address using the OUI prefix and address block of=20 > > lower three-octet (24bit). Taking advantage of the=20 > addressing scheme,=20 > > GMPLS may use the lower three-octet exclusively if a unique=20 > > OUI number=20 > > is reserved for the protocol. With this labeling scheme, GMPLS will=20 > > control MAC forwarding entry, not VLAN table. > > =20 > > All Ethernet frames controlled by GMPLS will have identical=20 > > OUI number=20 > > that they can easily be distinguished from other Ethernet frames.=20 > > In principle, the label lookup hardware is identical to MAC lookup=20 > > hardware in this labeling scheme. Therefore GMPLS implemented=20 > > switch may still function as normal Ethernet bridge to the frames=20 > > that OUI number is not GMPLS. This also facilitates GMPLS=20 > implemented=20 > > switches being deployed in operating Ethernet with minimum=20 > > service disruption. > > =20 > > Note also that above proposed label encoding method is transparent=20 > > to the use of Ethernet Length/Type field. End-user device may use=20 > > the Length/Type field as defined in IEEE 802.3 protocol.=20 > > It also allows network operators configure VLAN for their own=20 > > purpose.=20 > > When IEEE 802.1p is used in conjunction with L2-LSP, the priority=20 > > field of VLAN tag can also be used for imposing consistent TE=20 > > policy in=20 > > legacy switches and GMPLS switches. > > =20 > > Any way, my conclusion is, there are other options we may consider,=20 > > and this issue is still open to discuss. > > =20 > > Thanks, > > =20 > > Sincerely, > > =20 > > Jaihyung > > =20 > > =20 > > =20 > > =20 > >=20 > > -----?? ???-----=20 > > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com>=20 > > ?? ??: 2005-07-22 ?? 10:23:43=20 > > ?? ??: "Loa Andersson" <loa@pi.se>, "richard.spencer@bt.com"=20 > > <richard.spencer@bt.com>=20 > > ??: "per@defero.se" <per@defero.se>, "ccamp@ops.ietf.org"=20 > > <ccamp@ops.ietf.org>=20 > > ??: RE: Frameformat in a l2cs gmpls rnvironment.=20 > >=20 > >=20 > >=20 > >=20 > >=20 > >=20 > > Loa,=20 > >=20 > > I interpret the ID as a proposal to use GMPLS for VLAN setup.=20 > > So GMPLS and VLAN to not compete. The VLAN is at the data=20 > > plane and GMPLS at the control plane. The question is how and=20 > > should different control plane techniques like GMPLS and=20 > > (GVRP and STB) work together?=20 > >=20 > > Regards=20 > >=20 > > Juergen=20 > >=20 > >=20 > > > -----Original Message-----=20 > > > From: owner-ccamp@ops.ietf.org=20 > > > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson=20 > > > Sent: Friday, July 22, 2005 2:53 PM=20 > > > To: richard.spencer@bt.com=20 > > > Cc: per@defero.se; ccamp@ops.ietf.org=20 > > > Subject: Re: Frameformat in a l2cs gmpls rnvironment.=20 > > >=20 > > > Richard,=20 > > >=20 > > > I agree to most of this. Since we have doubts about the viability=20 > > > of taking GMPLS all the way to end-user or enterprise I think it=20 > > > would be good, from a wg perspective, if we agreed to solve the=20 > > > core network problems first.=20 > > >=20 > > > Do you have any comment on the requirement to run both VLANs and=20 > > > GMPLS on the same switch?=20 > > >=20 > > > /Loa=20 > > >=20 > > > richard.spencer@bt.com wrote:=20 > > > > Regardless of whether or not a switch is directly connected=20 > > > > to hosts, it must be able to forward packets using the=20 > > > connectionless=20 > > > > Ethernet data plane. This is due to the fundamental=20 > > > requirement that=20 > > > > for GMPLS switches to be able to exchange control=20 > > information with=20 > > > > each other, a data plane for control traffic must be present.=20 > > > > This is akin to using the IP data plane for MPLS=20 > > signalling in an=20 > > > > IP/MPLS network. An alternative would be to use a static=20 > > reserved=20 > > > > L2-LSP for control traffic in the same way that reserved=20 > > VPI/VCIs=20 > > > > are used for PNNI signalling in ATM.=20 > > > >=20 > > > > Regarding connecting hosts to GMPLS switches, I=20 > personally don't=20 > > > > think extending L2-LSPs into the enterprise/home network is=20 > > > > commercially viable. However, if you do want to use=20 > > GMPLS switches=20 > > > > in the home/enterprise network and for some reason=20 > don't want to=20 > > > > extend L2-LSPs down to the host then you will not be=20 > performing=20 > > > > normal Ethernet Mac address switching anyway. Instead you will=20 > > > > need some kind of policy on the switch that maps=20 > connectionless=20 > > > > Ethernet packets (e.g. based on MAC src/dest, 802.1p,=20 > VLAN) to a=20 > > > > L2-LSP. This is because multiple L2-LSPs to the same=20 > destination=20 > > > > (e.g. a gateway router) may exist for different services/flows=20 > > > > (e.g. video download, VoIP call, etc.).=20 > > > >=20 > > > > Regards,=20 > > > > Richard=20 > > > >=20 > > > >=20 > > > >>-----Original Message-----=20 > > > >>From: owner-ccamp@ops.ietf.org=20 > > [mailto:owner-ccamp@ops.ietf.org]On=20 > > > >>Behalf Of Par Mattsson=20 > > > >>Sent: 22 July 2005 11:42=20 > > > >>To: Loa Andersson=20 > > > >>Cc: ccamp@ops.ietf.org=20 > > > >>Subject: Re: Frameformat in a l2cs gmpls rnvironment.=20 > > > >>=20 > > > >>=20 > > > >>=20 > > > >>>Per and Dimitri,=20 > > > >>>=20 > > > >>>I would like to come down stronger than that, for me it is=20 > > > >>>a very strong requirement that the same switch can handle=20 > > > >>>both VLANs and GMPLs trafic correctly. I can't dsee how that=20 > > > >>>could be done if using the VLAN tpid to indicate GMPLS=20 > > > >>>traffic.=20 > > > >>=20 > > > >>If you ever want that same switch to handle traffic for a=20 > > directly=20 > > > >>connected host (not to uncommen) you would want that to=20 > > use normal=20 > > > >>ethernet macaddress switching. So of course you do not want=20 > > > to have to=20 > > > >>choose between vlan and gmpls, you would want both at the=20 > > same time.=20 > > > >>=20 > > > >>/per=20 > > > >>=20 > > > >>=20 > > > >>=20 > > > >>>/Loa=20 > > > >>>=20 > > > >>>Par Mattsson wrote:=20 > > > >>>=20 > > > >>>>>hi par, one of the possibilities that has been considered=20 > > > >>=20 > > > >>to cope with=20 > > > >>=20 > > > >>>>>this requirement is to use a dedicated TPID for the=20 > > > >>=20 > > > >>Ethernet labeled=20 > > > >>=20 > > > >>>>>frames; this would allow differentiated processing with=20 > > > non-labeled=20 > > > >>>>>framesthanks.=20 > > > >>>>=20 > > > >>>>=20 > > > >>>>That seems to make more sence. If that frame is to be=20 > > sized like a=20 > > > >>>>802.1q=20 > > > >>>>frame. There is not that much space left to a label. Or is=20 > > > >>=20 > > > >>the demand to=20 > > > >>=20 > > > >>>>use jumboframes ?=20 > > > >>>>Has there been any discussion on labelstacking, and=20 > > > mainly where to=20 > > > >>>>place=20 > > > >>>>the information?=20 > > > >>>>=20 > > > >>>>Regards.=20 > > > >>>>Per=20 > > > >>>>=20 > > > >>>>=20 > > > >>>>=20 > > > >>>=20 > > > >>>=20 > > > >>>--=20 > > > >>>Loa Andersson=20 > > > >>>=20 > > > >>>Principal Networking Architect=20 > > > >>>Acreo AB phone: +46 8 632 77 14=20 > > > >>>Isafjordsgatan 22 mobile: +46 739 81 21 64=20 > > > >>>Kista, Sweden email: =20 > > loa.andersson@acreo.se=20 > > > >>> loa@pi.se=20 > > > >>>=20 > > > >>=20 > > > >>=20 > > > >>=20 > > > >=20 > > >=20 > > >=20 > > > --=20 > > > Loa Andersson=20 > > >=20 > > > Principal Networking Architect=20 > > > Acreo AB phone: +46 8 632 77 14=20 > > > Isafjordsgatan 22 mobile: +46 739 81 21 64=20 > > > Kista, Sweden email: loa.andersson@acreo.se=20 > > > loa@pi.se=20 > > >=20 > >=20 > >=20 > >=20 > >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 14:37:03 +0000 Thread-Topic: Frameformat in a l2cs gmpls rnvironment. thread-index: AcWRJlni+Pe0SQFSTtCL57vT0szvAQ== Reply-To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> From: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> To: "Heiles Juergen" <juergen.heiles@siemens.com>, <per@defero.se> Cc: <ccamp@ops.ietf.org> Bcc: Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Mon, 25 Jul 2005 23:37:15 +0900 Comment: GQ19@|@ZEk=E?,18?x, BcN1b<z:P<.F@, 4c4g Message-ID: <024101c59126$59e76990$8310fe81@email1> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Class: urn:content-classes:message hi, Juergen see some in-line below, >> 3. use lower 3 bytes of MAC address for L2 label encoding. >> >> 802.1Q bridge forwards Ethernet frames using two dataplane tables >> - MAC forwarding table and VLAN forwarding table. >> Bridge control protocols, such as GARP, GVRP, GMRP, >> manipulate one of the two dataplane entities. >> >> Similarly, option 1 and 3 are about which one of two >> dataplane entities >> GMPLS protocol should control on behalf of bridge control protocols. >> The two proposals do not intend to modify bridge behavior >> seriously, such as MAC learning, aging, filtering. >> Therefore, the approaches 1 and 3 are in the scope of CCAMP. > >Even so your option 3 may not change the hardware it totaly changes >the forwarding behaviour as it redefines the meaning of a MAC address. >The MAC address is a end station address that identfies the source >and sink of the MAC frame. In your proposal it will be used as link >local label. There's no (in principle) physical change in the MAC forwarding behavior. GMPLS will configure MAC forwarding entry as if static entry is configured in MAC table by operator. Suppose a network that Ethernet bridges interconnect routers. Normally single L2 data path is established between routers by MAC learning. If an operator manually configure another MAC forwarding path in the bridges, two L2 paths are established between routers. Of course such configuration is not possible in 802.1 bridged network. However, GMPLS implementation may automate such configuration of semi-static MAC forwarding entry that otherwise be configured manually by operator. >It would be the same as if MPLS would use a dedicated >IP address range as lable instead of its own label. You may think it different perspective. MPLS forwards data using shim header, and the shim label is, again, encapsulated by Ethernet header. From the view of input and output link, Ethernet header (i.e. MAC addresses) is replaced at every hop together with shim label. What if the router knows which Ethernet header is to be replaced to which Ethernet header in next link? Does it still need to lookup MPLS shim header at L3 level? In this case, it is sufficient for the switch performing swap operation in L2SC level. The whole idea of GMPLS is employing common control over underlying transport method, and in the case of Ethernet, the major transport method of Ethernet frame is MAC forwarding using Ethernet address, and so is GMPLS implementation for Ethernet. >What will happen with the original MAC address when you enter >such a GMPLS Ethernet domain? Will you add a new "MAC header"? You are actually referring three cases, 1. in local network that host forwards Ethernet frame via L2-LSP 2. in core edge that provides L2VPN service 3. in network that legacy switch and GMPLS switches are interconnected. In 1, Ethernet address is swapped to GMPLS prefixed address and forwarded via established L2-LSP. In 2, user frames are encapsulated as you said. In 3, normal Ethernet frames are forwarded via normal STP path, however GMPLS frames may be forwarded via different path established using RSVP-TE. ...... >> >> The benefit of L2SC switch is explained in section 4 of >> <draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt> >> >> Overall objective is improving scalability, traffic engineering (TE) >> characteristics of 802.1 bridge that it can be reliable, manageable >> enough to replace some core routers. > >This is also provided by Ethernet over MPLS even with label stacking. but not by Ethernet bridges. When the price of Ethernet bridge is compared to MPLS router just by port and switching capacity, it is nearly ten times cheaper. What we are trying to do is delivering the capability of MPLS switch at the price of Ethernet switch. >> The switching technique you mentioned above, such as >> Ethernet over MPLS as defined by PWE3 and L2VPN, >> are all actually router based technology, however this >> work is based on simple bridge architecture. >> Cost-effectiveness is the key differentiator. >> >This is what I hear often and what I mean with "we still call >it Ethernet as it will be cahep and simple as todays Ehternet". >In the end the same functionality is needed: lable lookup, >label swapping, signaling protocol, routing protocol, discovery protocol, ... >So from the functionality and architecture point of view no difference exists. I understand well how today's routers become a machine of millions of gates. This is the solution I struggled for some years to simplify it best. That's why I am trying to preserve the original simplicity of hardware architecture of 802.1 Ethernet bridge. >One advantage compared to Ethernet over MPLS I see, is that >MPLS has to be again mapped into Ethernet for transport. >This addtional Ethernet layer is not needed. >In your response to Richard you mention to use a simpler routing >for Ethernet in this enviroment using dynamic learning like in Ethernet. >How about a unified GMPLS control plane and restoration. >Will you introduce another spanning tree? No, I'll just propose another way to achieve TE-LSP without relying on pre-established routing table. Please take a time to see <draft-jaihyung-ccamp-arpsignal-00.txt> and read the section about flood-routing. thanks Jaihyung -----?? ???----- ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com> ?? ??: 2005-07-25 ?? 9:32:40 ?? ??: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles Juergen" <juergen.heiles@siemens.com>, "per@defero.se" <per@defero.se> ??: "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> ??: RE: Frameformat in a l2cs gmpls rnvironment. Dear Jaihyung, the discussion now goes too much into implementations, however I would like to add my comments/questions. Regards Juergen > -----Original Message----- > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > Sent: Monday, July 25, 2005 1:40 PM > To: Heiles Juergen; per@defero.se > Cc: ccamp@ops.ietf.org > Subject: RE: Frameformat in a l2cs gmpls rnvironment. > > .... > 3. use lower 3 bytes of MAC address for L2 label encoding. > > > 802.1Q bridge forwards Ethernet frames using two dataplane tables > - MAC forwarding table and VLAN forwarding table. > Bridge control protocols, such as GARP, GVRP, GMRP, > manipulate one of the two dataplane entities. > > Similarly, option 1 and 3 are about which one of two > dataplane entities > GMPLS protocol should control on behalf of bridge control protocols. > The two proposals do not intend to modify bridge behavior > seriously, such as MAC learning, aging, filtering. > Therefore, the approaches 1 and 3 are in the scope of CCAMP. Even so your option 3 may not change the hardware it totaly changes the forwarding behaviour as it redefines the meaning of a MAC address. The MAC address is a end station address that identfies the source and sink of the MAC frame. In your proposal it will be used as link local label. It would be the same as if MPLS would use a dedicated IP address range as lable instead of its own label. What will happen with the original MAC address when you enter such a GMPLS Ethernet domain? Will you add a new "MAC header"? ...... > > The benefit of L2SC switch is explained in section 4 of > <draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt> > > Overall objective is improving scalability, traffic engineering (TE) > characteristics of 802.1 bridge that it can be reliable, manageable > enough to replace some core routers. This is also provided by Ethernet over MPLS even with label stacking. > The switching technique you mentioned above, such as > Ethernet over MPLS as defined by PWE3 and L2VPN, > are all actually router based technology, however this > work is based on simple bridge architecture. > Cost-effectiveness is the key differentiator. > This is what I hear often and what I mean with "we still call it Ethernet as it will be cahep and simple as todays Ehternet". In the end the same functionality is needed: lable lookup, label swapping, signaling protocol, routing protocol, discovery protocol, ... So from the functionality and architecture point of view no difference exists. One advantage compared to Ethernet over MPLS I see, is that MPLS has to be again mapped into Ethernet for transport. This addtional Ethernet layer is not needed. In your response to Richard you mention to use a simpler routing for Ethernet in this enviroment using dynamic learning like in Ethernet. How about a unified GMPLS control plane and restoration. Will you introduce another spanning tree? > .. Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 13:10:44 +0000 Message-ID: <42E4E3EE.4060408@pi.se> Date: Mon, 25 Jul 2005 15:06:54 +0200 From: Loa Andersson <loa@pi.se> Organization: Acreo AB User-Agent: Mozilla Thunderbird 1.0.5 (Windows/20050711) MIME-Version: 1.0 To: Dimitri.Papadimitriou@alcatel.be CC: =?ISO-8859-1?Q?P=E4r_Mattsson?= <per@defero.se>, ccamp@ops.ietf.org Subject: dimitris question on VLAN vs label (L2SC) Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 8bit Dimitri, the simple answer is that I need my VLANs, they give me a good, understood and reliable service, even when I can set up LSP through an Ethernet network. As I see it VLANsd and LSPs is not the same, if we use the VLAN to encapsulate LSPs we get into a problem of distinguishing between a VLAN and an LSP. VLAN is a scarce resource, 2048 per network (yeah - know I can increase it by managing the VLAN budget carefully, but it is kind of awkward). I think we should aim for something like 2000 labels per link (I wouldn't object if we can do better). I'd also say that the the Ethernet network needs to be able to handle all types of traffic, gmpls, tagged and untagged over the same swtich. Don't see that a interface separation would help since VLAN are per platform. Did I answer your questions? /Loa Dimitri.Papadimitriou@alcatel.be wrote: > loa, > > during the discussion loop one of the possibilities that came out > consisted in using a dedicated TPID value (different from the value > 0x8100 used in .1q), that would precede the encoded value as part of the > TCI > > would you please clarify your specific concern ? i would like also to > point out that there are several cases here > > 1. both type of traffic on the same switch but not necessarily on the > same interface > > 2. both type of traffic on the same interface > > with the following variant > > 2a. interface must accommodate any type of traffic (un-/tagged and labeled) > > 2b. interface must accommodate tagged and labeled type of traffic only > > thanks, > > - dimitri. > > *Loa Andersson <loa@pi.se>* > Sent by: owner-ccamp@ops.ietf.org > 07/22/2005 12:04 ZE2 > > To: Pär Mattsson <per@defero.se> > cc: ccamp@ops.ietf.org > bcc: > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > Per and Dimitri, > > I would like to come down stronger than that, for me it is > a very strong requirement that the same switch can handle > both VLANs and GMPLs trafic correctly. I can't dsee how that > could be done if using the VLAN tpid to indicate GMPLS > traffic. > > /Loa > > Pär Mattsson wrote: > >>hi par, one of the possibilities that has been considered to cope with > >>this requirement is to use a dedicated TPID for the Ethernet labeled > >>frames; this would allow differentiated processing with non-labeled > >>framesthanks. > > > > > > That seems to make more sence. If that frame is to be sized like a 802.1q > > frame. There is not that much space left to a label. Or is the demand to > > use jumboframes ? > > Has there been any discussion on labelstacking, and mainly where to place > > the information? > > > > Regards. > > Per > > > > > > > > > -- > Loa Andersson > > Principal Networking Architect > Acreo AB phone: +46 8 632 77 14 > Isafjordsgatan 22 mobile: +46 739 81 21 64 > Kista, Sweden email: loa.andersson@acreo.se > > loa@pi.se -- Loa Andersson Principal Networking Architect Acreo AB phone: +46 8 632 77 14 Isafjordsgatan 22 mobile: +46 739 81 21 64 Kista, Sweden email: loa.andersson@acreo.se loa@pi.se Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 12:33:22 +0000 Message-ID: <FF8AC5030873D6118BCB0002A58EDA990B9D2C16@mchh2a7e.mchh.siemens.de> From: Heiles Juergen <juergen.heiles@siemens.com> To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, Heiles Juergen <juergen.heiles@siemens.com>, per@defero.se Cc: ccamp@ops.ietf.org Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Mon, 25 Jul 2005 14:32:40 +0200 MIME-Version: 1.0 Content-Type: text/plain Dear Jaihyung, the discussion now goes too much into implementations, however I would like to add my comments/questions. Regards Juergen > -----Original Message----- > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > Sent: Monday, July 25, 2005 1:40 PM > To: Heiles Juergen; per@defero.se > Cc: ccamp@ops.ietf.org > Subject: RE: Frameformat in a l2cs gmpls rnvironment. > > ..... > 3. use lower 3 bytes of MAC address for L2 label encoding. > > > 802.1Q bridge forwards Ethernet frames using two dataplane tables > - MAC forwarding table and VLAN forwarding table. > Bridge control protocols, such as GARP, GVRP, GMRP, > manipulate one of the two dataplane entities. > > Similarly, option 1 and 3 are about which one of two > dataplane entities > GMPLS protocol should control on behalf of bridge control protocols. > The two proposals do not intend to modify bridge behavior > seriously, such as MAC learning, aging, filtering. > Therefore, the approaches 1 and 3 are in the scope of CCAMP. Even so your option 3 may not change the hardware it totaly changes the forwarding behaviour as it redefines the meaning of a MAC address. The MAC address is a end station address that identfies the source and sink of the MAC frame. In your proposal it will be used as link local label. It would be the same as if MPLS would use a dedicated IP address range as lable instead of its own label. What will happen with the original MAC address when you enter such a GMPLS Ethernet domain? Will you add a new "MAC header"? ....... > > The benefit of L2SC switch is explained in section 4 of > <draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt> > > Overall objective is improving scalability, traffic engineering (TE) > characteristics of 802.1 bridge that it can be reliable, manageable > enough to replace some core routers. This is also provided by Ethernet over MPLS even with label stacking. > The switching technique you mentioned above, such as > Ethernet over MPLS as defined by PWE3 and L2VPN, > are all actually router based technology, however this > work is based on simple bridge architecture. > Cost-effectiveness is the key differentiator. > This is what I hear often and what I mean with "we still call it Ethernet as it will be cahep and simple as todays Ehternet". In the end the same functionality is needed: lable lookup, label swapping, signaling protocol, routing protocol, discovery protocol, ... So from the functionality and architecture point of view no difference exists. One advantage compared to Ethernet over MPLS I see, is that MPLS has to be again mapped into Ethernet for transport. This addtional Ethernet layer is not needed. In your response to Richard you mention to use a simpler routing for Ethernet in this enviroment using dynamic learning like in Ethernet. How about a unified GMPLS control plane and restoration. Will you introduce another spanning tree? > ... Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 11:58:06 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: RE: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Date: Mon, 25 Jul 2005 12:56:27 +0100 Message-ID: <0536FC9B908BEC4597EE721BE6A353890A9F1E3F@i2km07-ukbr.domain1.systemhost.net> Thread-Topic: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Thread-Index: AcWPqwVLppO1DKV4Sju5p0rdnFh0zwBYqxMA From: <neil.2.harrison@bt.com> To: <nwnetworks@dial.pipex.com> Cc: <ccamp@ops.ietf.org> Tom, G.805 (co) and G.809 (cl) discuss general functional architectural properties of the data-plane network. Specific technology ones are then generated from these. We've found this stuff invaluable for many reasons....but its esp crucial if you want sound management information models.....and for large operators esp this is a very important topic IMO. I find it most useful however in a day-day sense (at least in terms of the principles its based on) to be able to quickly analyse a technology and spot its strengths/weaknesses. =20 Of course a data-plane network can carry traffic or control/management protocols. When you are dealing with co mode networks its quite sensible (and forced in case of the co-cs mode) to have different data-plane networks working in the same network system. But see my post to Igor a while back wrt to what this means when adapting these to be carried by a lower layer network. There are rather important network owenship issues here. Note also the design (eg for survivability) of the traffic and control/management data-plane networks can be very different. And it is particularly important that the control/management-plane networks only take their design cues from the duct-graph (you must know about real physical separacy here if you want to do a good job) The stuff in SG15 on ASON (various Recs here) discuss control-plane arch. And there is a stuff in the M series Recs on management arch......but I'd look to TMF for this. If you want to know a good text book, I've still to find anything better than Reid/Sexton 'Broadband networking'. Chapters 3 and 4 therein give an excellent introduction to func arch and what this means for the 3 planes.....but there is lots of other good stuff in here on many topics. Bit old-hat wrt technologies discussed, but this is not so importnat as the general networking principles are invariant of technology (provided one understands the basic modal differences). regards, neil > -----Original Message----- > From: Tom Petch [mailto:nwnetworks@dial.pipex.com]=20 > Sent: 23 July 2005 16:58 > To: Harrison,N,Neil,IKM1 R > Cc: ccamp@ops.ietf.org > Subject: Re: Ethernet Control Plane [Was: Re: Frameformat in=20 > a l2cs gmpls rnvironment.] >=20 >=20 > Neil >=20 > Is there an ITU-T Recommendation that gives an architectural=20 > definition of what it means by control plane, data plane,=20 > management plane, user plane, traffic dataplane, control data=20 > plane etc etc? >=20 > I know what I mean when I use the terms but am not sure I=20 > always know what others mean. >=20 > Tom Petch > ----- Original Message ----- > From: <neil.2.harrison@bt.com> > To: <John.E.Drake2@boeing.com>; <adrian@olddog.co.uk>;=20 > <richard.spencer@bt.com>; <juergen.heiles@siemens.com>;=20 > <loa@pi.se>; <per@defero.se> > Cc: <ccamp@ops.ietf.org> > Sent: Saturday, July 23, 2005 2:15 PM > Subject: RE: Ethernet Control Plane [Was: Re: Frameformat in=20 > a l2cs gmpls rnvironment.] >=20 >=20 > > > > IP is cl-ps networking techology period....and its > > > > jolly important, but PLEASE don't try and feed me any=20 > of this 'IP=20 > > > > control plane' nonsense. > > > > JD: Over the years I have enjoyed Neil's rants immensely,=20 > but I have=20 > > to say that this one is a classic. >=20 > NH=3D> Thanks John for the kind words....its good to know=20 > you've had at least something to smile about ;-) >=20 > Seriously though, see my updated response to Adrian. But if=20 > you are still struggling to grasp what I mean by the above=20 > (out of context > quote) try this: >=20 > A single cl-ps IP network carrying traffic and everything=20 > else (eg public Internet) is very different animal to a co-cs=20 > network system where the traffic data-plane network and=20 > control/management data-planes network are logically=20 > disjoint, ie the role of IP in the latter is importantly=20 > different to the role of IP in the former...and let's hope=20 > never the 2 shall meet. >=20 > regards, neil >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 11:39:32 +0000 Thread-Topic: Frameformat in a l2cs gmpls rnvironment. thread-index: AcWRDY8OHVZh5WBvSbOBxtrHtj5d+w== Reply-To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> From: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> To: "Heiles Juergen" <juergen.heiles@siemens.com>, <per@defero.se> Cc: <ccamp@ops.ietf.org> Bcc: Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Mon, 25 Jul 2005 20:39:47 +0900 Comment: GQ19@|@ZEk=E?,18?x, BcN1b<z:P<.F@, 4c4g Message-ID: <4e95c01c5910d$8f111170$8310fe81@email1> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Class: urn:content-classes:message hi, Juergen >Dear Jaihyung, > >you mention several options that could be used as a label. >Most of them have strong impacts on the data plane and introduce >new layer networks. Is this in the scope of CCAMP? What is the goal, In below mail, I discussed about three proposed approaches. 1. using VLAN tag as it is for L2 label encoding. 2. defining new protocol ID (TPID) and borrow VLAN tag format, ?or extended VLAN tag to implement label swapping. 3. use lower 3 bytes of MAC address for L2 label encoding. 802.1Q bridge forwards Ethernet frames using two dataplane tables - MAC forwarding table and VLAN forwarding table. Bridge control protocols, such as GARP, GVRP, GMRP, manipulate one of the two dataplane entities. Similarly, option 1 and 3 are about which one of two dataplane entities GMPLS protocol should control on behalf of bridge control protocols. The two proposals do not intend to modify bridge behavior seriously, such as MAC learning, aging, filtering. Therefore, the approaches 1 and 3 are in the scope of CCAMP. (% note however, I would regard implementing label swapping function is acceptable change considering the Internal Sublayer design of 802.1D bridge.) However, option 2 is about what TPID code we will choose to implement new switching function. The switch of this implementation should have new hardware design in addition to normal bride hardware. I think such proposal as option 2 is out of CCAMP scope. >to introduce a control plane for a existing L2 technology or >introduce a new L2 technology? What would be the benefit >of such a new L2 co switching technology compared to Ethernet >over MPLS as defined by PWE3 and L2VPN? The benefit of L2SC switch is explained in section 4 of <draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt> Overall objective is improving scalability, traffic engineering (TE) characteristics of 802.1 bridge that it can be reliable, manageable enough to replace some core routers. The switching technique you mentioned above, such as Ethernet over MPLS as defined by PWE3 and L2VPN, are all actually router based technology, however this work is based on simple bridge architecture. Cost-effectiveness is the key differentiator. In access or enterprise network, capability of providing end-to-end L2-LSP will enable service providers policing, measuring, charging application flows using Ethernet based network. This will eventually improve income structure and introduce new session based commercial service. I have been discussing this aspect in mail thread of title 'End-to-end L2-LSP'. Please read the mail thread and comment on the discussion. >You mentioned to use the MAC address or part of it as a label. >The TRILL working group is defining shortest path routing for Ethernet. >How would this fit together? TRILL employs new encapsulation layer outside of Ethernet frames. However in my proposal of option-3, I simply use the Ethernet header to encode GMPLS label. The overall format of rbridge is [Ethernet][r-tag][Ethernet][ data..] It is not LSP based technology but a connectionless routing bridge. TTL count is very important in rbridge because frames are routed hop-by-hop and it may loop. Whereas in option-3 above, Ethernet frames that have certain OUI prefix in MAC address are forwarded via a path that GMPLS protocol has configured on MAC forwarding table. Thanks Jaihyung -----?? ???----- ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com> ?? ??: 2005-07-25 ?? 6:44:34 ?? ??: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, "Heiles Juergen" <juergen.heiles@siemens.com>, "per@defero.se" <per@defero.se> ??: "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> ??: RE: Frameformat in a l2cs gmpls rnvironment. Dear Jaihyung, you mention several options that could be used as a label. Most of them have strong impacts on the data plane and introduce new layer networks. Is this in the scope of CCAMP? What is the goal, to introduce a control plane for a existing L2 technology or introduce a new L2 technology? What would be the benefit of such a new L2 co switching technology compared to Ethernet over MPLS as defined by PWE3 and L2VPN? You mentioned to use the MAC address or part of it as a label. The TRILL working group is defining shortest path routing for Ethernet. How would this fit together? Regards Juergen > -----Original Message----- > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > Sent: Saturday, July 23, 2005 6:19 AM > To: Heiles Juergen; per@defero.se > Cc: ccamp@ops.ietf.org > Subject: RE: Frameformat in a l2cs gmpls rnvironment. > > > Dear Juergen, Par and all, > > The proposed framework of L2SC was not intended to suggest > any specific solution. It is still an open question. > As Adrian noted, it only aims to present requirements > as in scenarios that may drive people toward next step in many > viable solutions. > > In fact, there are other options we may also consider for L2 > label encoding. > IEEE 802.1 Ethernet bridges forward frames based on 48bits of > MAC address, > and additionally using VLAN tag. > > When the purpose of GMPLS control over Ethernet is not to > create new dataplane > but to utilize IEEE 802.1 bridge architecture, we may consider using > one of the two forwarding methods, MAC forwarding or VLAN forwarding. > (and perhaps any other combination of fields in MAC, but I'll > not discuss it) > > Use of VLAN ID for label encoding may automate VLAN configuration > using IP protocols. However, GMPLS protocol cannot use the > field exclusively > because public/private operators already use VLAN for various > purpose. > There is a potential conflict with existing use of VLAN and > GMPLS use of VLAN label. > Furthermore, scalability of VLAN ID has been frequently noted > as weakness > because the size of VLAN ID is at most 4096 (12bit). > > The scalability may be improved if the scope of VLAN label is > confined to > link-local, and some additional swapping function of VLAN ID > is introduced in > Internal Sublayer [802.1D] of GMPLS implemented switch. > However, this will only be effective when the configuration > of network > is mesh structure that multiple LSP paths exist. If the > configuration of network > is star or tree shape, as normal configuration of access network, > LSPs concentrate in root node and total number of > LSPs that the network can hold still be limited by available > label space > at a few root links. > > For these reasons, I do not think any form of VLAN ID label is > an appropriate choice for layer-2 label encoding. > > There are some other proposals assuming new assignment of Ethernet > Length/Type value (e.g. new TPID in VLAN tag) and re-definition of > information fields placed between 802.3 MAC header and IP packet. > In this case, only the format of VLAN tag or extended VLAN tag > is borrowed, however, inside the switching hardware, the > filter and relay, > etc. are totally different new dataplane switch. > > I do not see such approach is a GMPLS implementation for Ethernet > because the core switching technique is not 802.1 Ethernet > bridge at all. > > The other option we may consider is using MAC address filed as below. > > > +-------+-------+-------+-------+-------+-------+ > | 1byte | 2byte | 3byte | 4byte | 5byte | 6byte | > +-------+-------+-------+-------+-------+-------+ > +-----------------------+-----------------------+ > | OUI Prefix (=GMPLS) | DA-label (24bit) | > +-----------------------+-----------------------+ > | OUI Prefix (=GMPLS) | SA-label (24bit) | > +---------------+-------+-----------------------+ > | Length/Type | > +---------------+ > > > IEEE is designated by the ISO Council to act as the > registration authority > for the higher three-octet of OUI number in the MAC address > to be used by manufacturer. Ethernet manufacturer may generate > global unique MAC address using the OUI prefix and address block of > lower three-octet (24bit). Taking advantage of the addressing scheme, > GMPLS may use the lower three-octet exclusively if a unique > OUI number > is reserved for the protocol. With this labeling scheme, GMPLS will > control MAC forwarding entry, not VLAN table. > > All Ethernet frames controlled by GMPLS will have identical > OUI number > that they can easily be distinguished from other Ethernet frames. > In principle, the label lookup hardware is identical to MAC lookup > hardware in this labeling scheme. Therefore GMPLS implemented > switch may still function as normal Ethernet bridge to the frames > that OUI number is not GMPLS. This also facilitates GMPLS implemented > switches being deployed in operating Ethernet with minimum > service disruption. > > Note also that above proposed label encoding method is transparent > to the use of Ethernet Length/Type field. End-user device may use > the Length/Type field as defined in IEEE 802.3 protocol. > It also allows network operators configure VLAN for their own > purpose. > When IEEE 802.1p is used in conjunction with L2-LSP, the priority > field of VLAN tag can also be used for imposing consistent TE > policy in > legacy switches and GMPLS switches. > > Any way, my conclusion is, there are other options we may consider, > and this issue is still open to discuss. > > Thanks, > > Sincerely, > > Jaihyung > > > > > > -----?? ???----- > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com> > ?? ??: 2005-07-22 ?? 10:23:43 > ?? ??: "Loa Andersson" <loa@pi.se>, "richard.spencer@bt.com" > <richard.spencer@bt.com> > ??: "per@defero.se" <per@defero.se>, "ccamp@ops.ietf.org" > <ccamp@ops.ietf.org> > ??: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > > Loa, > > I interpret the ID as a proposal to use GMPLS for VLAN setup. > So GMPLS and VLAN to not compete. The VLAN is at the data > plane and GMPLS at the control plane. The question is how and > should different control plane techniques like GMPLS and > (GVRP and STB) work together? > > Regards > > Juergen > > > > -----Original Message----- > > From: owner-ccamp@ops.ietf.org > > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson > > Sent: Friday, July 22, 2005 2:53 PM > > To: richard.spencer@bt.com > > Cc: per@defero.se; ccamp@ops.ietf.org > > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > > Richard, > > > > I agree to most of this. Since we have doubts about the viability > > of taking GMPLS all the way to end-user or enterprise I think it > > would be good, from a wg perspective, if we agreed to solve the > > core network problems first. > > > > Do you have any comment on the requirement to run both VLANs and > > GMPLS on the same switch? > > > > /Loa > > > > richard.spencer@bt.com wrote: > > > Regardless of whether or not a switch is directly connected > > > to hosts, it must be able to forward packets using the > > connectionless > > > Ethernet data plane. This is due to the fundamental > > requirement that > > > for GMPLS switches to be able to exchange control > information with > > > each other, a data plane for control traffic must be present. > > > This is akin to using the IP data plane for MPLS > signalling in an > > > IP/MPLS network. An alternative would be to use a static > reserved > > > L2-LSP for control traffic in the same way that reserved > VPI/VCIs > > > are used for PNNI signalling in ATM. > > > > > > Regarding connecting hosts to GMPLS switches, I personally don't > > > think extending L2-LSPs into the enterprise/home network is > > > commercially viable. However, if you do want to use > GMPLS switches > > > in the home/enterprise network and for some reason don't want to > > > extend L2-LSPs down to the host then you will not be performing > > > normal Ethernet Mac address switching anyway. Instead you will > > > need some kind of policy on the switch that maps connectionless > > > Ethernet packets (e.g. based on MAC src/dest, 802.1p, VLAN) to a > > > L2-LSP. This is because multiple L2-LSPs to the same destination > > > (e.g. a gateway router) may exist for different services/flows > > > (e.g. video download, VoIP call, etc.). > > > > > > Regards, > > > Richard > > > > > > > > >>-----Original Message----- > > >>From: owner-ccamp@ops.ietf.org > [mailto:owner-ccamp@ops.ietf.org]On > > >>Behalf Of Par Mattsson > > >>Sent: 22 July 2005 11:42 > > >>To: Loa Andersson > > >>Cc: ccamp@ops.ietf.org > > >>Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > >> > > >> > > >> > > >>>Per and Dimitri, > > >>> > > >>>I would like to come down stronger than that, for me it is > > >>>a very strong requirement that the same switch can handle > > >>>both VLANs and GMPLs trafic correctly. I can't dsee how that > > >>>could be done if using the VLAN tpid to indicate GMPLS > > >>>traffic. > > >> > > >>If you ever want that same switch to handle traffic for a > directly > > >>connected host (not to uncommen) you would want that to > use normal > > >>ethernet macaddress switching. So of course you do not want > > to have to > > >>choose between vlan and gmpls, you would want both at the > same time. > > >> > > >>/per > > >> > > >> > > >> > > >>>/Loa > > >>> > > >>>Par Mattsson wrote: > > >>> > > >>>>>hi par, one of the possibilities that has been considered > > >> > > >>to cope with > > >> > > >>>>>this requirement is to use a dedicated TPID for the > > >> > > >>Ethernet labeled > > >> > > >>>>>frames; this would allow differentiated processing with > > non-labeled > > >>>>>framesthanks. > > >>>> > > >>>> > > >>>>That seems to make more sence. If that frame is to be > sized like a > > >>>>802.1q > > >>>>frame. There is not that much space left to a label. Or is > > >> > > >>the demand to > > >> > > >>>>use jumboframes ? > > >>>>Has there been any discussion on labelstacking, and > > mainly where to > > >>>>place > > >>>>the information? > > >>>> > > >>>>Regards. > > >>>>Per > > >>>> > > >>>> > > >>>> > > >>> > > >>> > > >>>-- > > >>>Loa Andersson > > >>> > > >>>Principal Networking Architect > > >>>Acreo AB phone: +46 8 632 77 14 > > >>>Isafjordsgatan 22 mobile: +46 739 81 21 64 > > >>>Kista, Sweden email: > loa.andersson@acreo.se > > >>> loa@pi.se > > >>> > > >> > > >> > > >> > > > > > > > > > -- > > Loa Andersson > > > > Principal Networking Architect > > Acreo AB phone: +46 8 632 77 14 > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > Kista, Sweden email: loa.andersson@acreo.se > > loa@pi.se > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 10:13:32 +0000 Thread-Topic: End-to-end L2-LSP thread-index: AcWRAaPmSrrXoDteQTejc8QSKAORQQ== Reply-To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> From: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> To: <richard.spencer@bt.com>, <ccamp@ops.ietf.org> Cc: Bcc: Subject: RE: End-to-end L2-LSP Date: Mon, 25 Jul 2005 19:14:28 +0900 Comment: GQ19@|@ZEk=E?,18?x, BcN1b<z:P<.F@, 4c4g Message-ID: <4e13c01c59101$a3eb7880$8310fe81@email1> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Class: urn:content-classes:message Hi, Richard Here's continued part of reply to your original question. reply to the questions 1~5 are also enclosed below. >6. Enterprise networks are built using not just switches >but routers as well. Firstly, for L2-LSPs to be set up >between switches and routers, they must belong to >the same address space to communicate with each other. >However, this address space must be separate from the >IPv4 forwarding address space. To use an example >of why this is true, routers connected via Ethernet >segments are considered to be directly connected, however, >if the IPv4 and L2-LSP address spaces were the same, >then IP routers separated by GMPLS switches would be >multiple hops away from each other. In addition to address >space separation, routing protocol instances must also be separated. >This need for this separation is obvious when considering >using a link state routing protocol which relies on a topological >view of the network, i.e. the IPv4 and L2-LSP topologies are different. Allow me a bit lengthy writing at this time, because there's some message I want to tell you and people who might read this. I think routing of GMPLS signal should be different in the access/enterprise network. If GMPLS implemented Ethernet switch aims to replace some core routers, it has to offer the same routing property as core routers do. However, if a vendor targets Ethernet market using the GMPLS implemented switch, the switch must be able to offer similar characteristics as Ethernet switches. In Ethernet, such notion of addressing and routing configuration as in IGP doesn't exist. The only routing feature Ethernet has is 'broadcast, MAC learning, and filtering'. (%note: STP, MSTP, etc. must be seen in different perspective in this context) This simplistic mind of routing results in successful acceptance of Ethernet. Now ordinary users may purchase and install Ethernet switch without burden of complex network configuration. Access network engineers are able to install Ethernet switches quickly in many outdoor sites because manual work is relatively simple. This helps reducing labor cost and OPEX a lot. In summary, what made Ethernet proliferating is the lightweightness, i.e. no manual numbering on ports, near-zero configuration, dynamic learning, broadcast & filtering, etc. Such lightweightness can also be achieved in GMPLS when GMPLS signal doesn't rely on IP routing protocols, such as link-status or distance vector algorithm, but itself has controlled broadcast and learning capability as seen in Ethernet. Ethernet learns destination MAC address from ARP broadcast. The switch optimizes data path from dynamic learning of backward path. Similarly, GMPLS signal may also be broadcast in a controlled way. Destination host group or access point to core network can be found from broadcast. L2-LSP can be establishment via the path learned from broadcast if the broadcast path is systematically recorded. You may reference <draft-jaihyung-ccamp-arpsignal-00.txt> for detail of such routing method. Although the draft was written for ARP signaling, it can also be applied to RSVP-TE. I'll soon submit updated flood-routing method for RSVP-TE. The key message I want to tell is, the GMPLS control stack can be simplified if routing and signaling are merged in unified protocol. I'm sure this approach doesn't have the problem of routing hop or address space separation as you raised above. >Secondly, routers break up Ethernet segments and must therefore >must always be the headend or tailend for a L2-LSPs, i.e. they >cannot be midpoints. This means that from a control plane perspective, >on receiving a L2-LSP path message a router must terminate the LSP >and send back a resv message whether it is the intended destination/sink >or not. It must then originate a path message for a new L2-LSP >and send it onto the next hop GMPLS switch. This is not how >RSVP-TE works today and therefore will require modifications. >In addition, traffic will be blackholed if transmitted by the source >before all the separate L2-LSPs have been setup end-to-end to >the sink/destination. This means that some kind of end-to-end >signalling message will need to be sent by the sink/destination >once the final L2-LSP has been established to inform the source >that the end-to-end connectivity has been provided. This functionality >is not supported by RSVP-TE today and will therefore require protocol modifications. I agree in the point that routers break up L2-LSP connectivity. There's no way to provide end-to-end L2-LSP connectivity across L3 routers unless the routers are capable of supporting L2SC GMPLS. Perhaps local routers, gateways and edge routers are the first target to deploy L2SC GMPLS rather than Ethernet switches. It is also no doubt some extension to RSVP-TE for L2SC will be necessary for end-to-end exchange of RSVP-TE. Clearly things that you pointed out need to be done. >Thirdly, routers forward packets based on information in the IP header. >So, when a router receives a packet via a L2-LSP it will strip of the >L2 header and perform an IP lookup. Now, for the packet to be >forwarded using the correct L2-LSP the router must have a policy >that tells it which L2-LSP to use as there may be multiple >L2-LSPs to the same destination, e.g. for different voice calls or >video downloads. However, before a policy can even be configured, >the router needs to know what traffic (e.g. based on src/dest address, >port number, DSCP, etc.) to map to which L2-LSP, and this information >needs to come from the traffic source using signalling, i.e. RSVP-TE. >I do not believe this is supported today and will therefore require modifications. Before you reach this conclusion, I want you look again where the source of problem is originated. If routers simply dispose MAC information, the router is losing important information to determine forwarding policy. The policy should have been known priori during LSP establishment time using RSVP-TE. This information is encoded in L2 label by ingress, thus the egress router should be able to keep the label information and use it for determining next L2-LSP to map. If a router doesn't have such capability of L2SC GMPLS, of course the policy decision will be that complex. Sincerely, Jaihyung -----?? ???----- ?? ??: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> ?? ??: 2005-07-21 ?? 11:59:07 ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com>, "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> ??: ??: RE: End-to-end L2-LSP Hi, Richard, Thank you for your tons of questions. It is a bit hard to answer in single mail. So, I divided it in two pieces. Here I send you the first part of answers to questions 1~5. >I agree that extending signalling down to the end-user is one >possible solution to address charging/billing issues, another possible >solution would be to use session based billing. I also agree that using >L2-LSPs could be a viable solution for providing traffic separation >and bandwidth guarantees in a service providers network, another Indeed, that's one important message I wanted to cast in this WG. Not only this is technically interesting, but the behind goal is about improving income structure of Internet service providers. I hope this can be an answer to the issue of QoS&Pricing that Neil Harrison once noted in last mail of this thread. Customers may basically use cl-ps service in fixed rate base, and when they need real-time service such as IP-TV or video phone, they may request simple CBR-like co-ps SVC-based service (i.e. L2-LSP) over cl-ps network using GMPLS signal. This way, operators may improve their service structure and invest to upgrade their access network. [snip] >1. If you want to eliminate the possibility of congestion in the >enterprise/home network using L2-LSPs for the reasons you >give below, then on every node in the network you will need: > >- A CPU: for control plane protocol processing >- Routing: At least static, an IGP for large networks and >OSPF-TE if TE is required >- Signalling: RSVP-TE >- Policing: Per connection policing, e.g. per VLAN > >I do not know how much it would cost to upgrade the $20 five >port 10/100 switch sitting on my desk to support all this, >but I do not believe this would be a trivial amount, particularly >at the volumes demanded by enterprises. I guess your five port $20 switch will be connected to machines owned only by you. If you can control resource usage of your machines, you don't need GMPLS implemented switch. However, if you need to share a switch with some of your colleague, I will recommend spending more that $20 on switch in order to enjoy nice video communication while your colleague abusing your network with P2P file transfer. I think this is worth while to spend money considering importance of your personal communication. As I noted in previous mail, GMPLS implemented switch will be necessary in nodes where automated resource control is crucial. >In addition, a hierarchical >architecture would be required for the solution to scale and >therefore aggregation switches within the enterprise network >would also need to support hierarchical CAC and policing. This is beyond issue we can discuss at the moment. I'd like further technical discussion, however, I think we need to wait until this idea of L2-LSP is officially chartered in CCAMP. I hope enough number of people support the idea that we can continue work on this issue. >2. Unless you want to reinvent LANE, then you will need to >continue to support CL Ethernet broadcasts and multicasts on >GMPLS enabled switches. To protect the L2-LSP traffic from >CL broadcast traffic you could rate limit it to say 1% of >the bandwidth available. However, how would you capacity >plan for and rate limit (dynamic, receiver driven) multicast >traffic for which usage patterns are difficult to predict? >In my opinion you will need to use over provisioning along with CoS, >which would raise the question, why is over provisioning >and CoS OK for multicast traffic >(which could be critical, e.g. trading floor information), >but not OK for unicast traffic? There will be a technical solution for multicast and broadcast. A member of DT once showed his interest on multicast solution for Ethernet using RSVP instead of IGMP. I also have an idea of extending point-to-point L2-LSP to multipoint L2-LSP as similar way that learning bridges learn backward path from source address. However, I have some doubt on necessity of multicast service. Do we really need multicast for streaming service in local/access ? I think even IP-TV will be better serviced using point-to-point connection in access network. There's very low chance of two subscribers receiving same channel at the same time if the receiver group is not that great. Multicasting is necessary in core network, however I doubt the same needs in local/access network, except some rare case. >3. Routers are relatively stable systems that are switched on >permanently and therefore RSVP-TE sessions are stable >and only go down when there is a failure. End user systems >however such as PCs, set top boxes, games consoles, etc. >are not always switched on and are not as stable. >What will the effect on the network be when large number >of RSVP-TE sessions are going up/down, e.g. as a result >of game console resets or PC crashes/freezes? >These effects will ripple through the enterprise/home >network into the operators network, and therefore mechanisms >must be put in place to protect against them. Control overhead of RSVP was discussed a lot in conjunction with IntServ. There have been proposals to reduce the overhead in core network such as RFC3175 or [HIER]. In particular, proposed architecture for LSP aggregation [HIER] will greatly help to solve the issue. % [HIER] draft-ieft-mpls-lsp-hierarchy-08.txt >4. When an IP device wants to forward a packet it uses ARP >to discover the MAC address of the IP next hop >(unless its already in the MAC table) and adds the Ethernet >header before forwarding the packet on. Now, if you want to >extend connections down to the end user device, then the >device will need to map IP traffic to different L2-LSPs. >A simple 1-to-1 mapping between an IP next hop address >and a MAC address will not be sufficient as there may >be multiple connections (e.g. using different VLAN IDs) >associated with an IP next hop. So, policies will need >to be configured on end user devices to map IP traffic >(e.g. based on src/dest IP address, port numbers, etc.) >to a specific L2-LSP (e.g. based on src/dest MAC address and/or VLAN ID). You pointed out important issue. Currently, Ethernet layer only maps IP and MAC in 1-to-1. Unless Ethernet-IP interface is extended as you suggested, users must be satisfied with at most single L2-LSP established between machine to machine. If they want another connection, they have to disconnect previous connection. This scenario will be OK in most real-life applications. However in future, extension to layer interface will be necessary and the solution should be consulted with IEEE. >5. End user devices such as PCs and VoIP phones are >likely to only need to support a relatively low number >of L2-LSPs. However, devices such as content servers >and call servers/gateways will need to support >hundreds/thousands of L2-LSPs. Connection setup and >state maintenance for large numbers of L2-LSPs will >consume a significant amount of processing resources, >as will policies to map IP traffic to the different L2-LSPs. >Therefore, I would expect that the system software >(and possibly hardware) will need to be designed from >the ground up to be able to support large numbers of L2-LSPs. Some high-performance servers are capable of handling millions of TCP connections. I believe the added complexity for LSP connection will not be far greater that the work-load normally assumed in contemporary servers. It may be a problem to local gateways if the signaling load is not bypassed. However, even in this case, there can be some solutions we can devise. ok, for rest of questions, I'll reply it in next mail. thanks Jaihyung -----?? ???----- ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com> ?? ??: 2005-07-21 ?? 12:09:56 ?? ??: "jaihyung@etri.re.kr" <jaihyung@etri.re.kr>, "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> ??: ??: RE: End-to-end L2-LSP Jaihyung, I agree that extending signalling down to the end-user is one possible solution to address charging/billing issues, another possible solution would be to use session based billing. I also agree that using L2-LSPs could be a viable solution for providing traffic separation and bandwidth guarantees in a service providers network, another possible solution would be to extended MPLS into the access network. However, if you want to extend L2-LSPs (not just signalling) down to the end user, then there are a number of complicated technical and commercial issues that must be overcome. Please see below for some of the key issues: 1. If you want to eliminate the possibility of congestion in the enterprise/home network using L2-LSPs for the reasons you give below, then on every node in the network you will need: - A CPU: for control plane protocol processing - Routing: At least static, an IGP for large networks and OSPF-TE if TE is required - Signalling: RSVP-TE - Policing: Per connection policing, e.g. per VLAN I do not know how much it would cost to upgrade the $20 five port 10/100 switch sitting on my desk to support all this, but I do not believe this would be a trivial amount, particularly at the volumes demanded by enterprises. In addition, a hierarchical architecture would be required for the solution to scale and therefore aggregation switches within the enterprise network would also need to support hierarchical CAC and policing. 2. Unless you want to reinvent LANE, then you will need to continue to support CL Ethernet broadcasts and multicasts on GMPLS enabled switches. To protect the L2-LSP traffic from CL broadcast traffic you could rate limit it to say 1% of the bandwidth available. However, how would you capacity plan for and rate limit (dynamic, receiver driven) multicast traffic for which usage patterns are difficult to predict? In my opinion you will need to use over provisioning along with CoS, which would raise the question, why is over provisioning and CoS OK for multicast traffic (which could be critical, e.g. trading floor information), but not OK for unicast traffic? 3. Routers are relatively stable systems that are switched on permanently and therefore RSVP-TE sessions are stable and only go down when there is a failure. End user systems however such as PCs, set top boxes, games consoles, etc. are not always switched on and are not as stable. What will the effect on the network be when large number of RSVP-TE sessions are going up/down, e.g. as a result of game console resets or PC crashes/freezes? These effects will ripple through the enterprise/home network into the operators network, and therefore mechanisms must be put in place to protect against them. 4. When an IP device wants to forward a packet it uses ARP to discover the MAC address of the IP next hop (unless its already in the MAC table) and adds the Ethernet header before forwarding the packet on. Now, if you want to extend connections down to the end user device, then the device will need to map IP traffic to different L2-LSPs. A simple 1-to-1 mapping between an IP next hop address and a MAC address will not be sufficient as there may be multiple connections (e.g. using different VLAN IDs) associated with an IP next hop. So, policies will need to be configured on end user devices to map IP traffic (e.g. based on src/dest IP address, port numbers, etc.) to a specific L2-LSP (e.g. based on src/dest MAC address and/or VLAN ID). 5. End user devices such as PCs and VoIP phones are likely to only need to support a relatively low number of L2-LSPs. However, devices such as content servers and call servers/gateways will need to support hundreds/thousands of L2-LSPs. Connection setup and state maintenance for large numbers of L2-LSPs will consume a significant amount of processing resources, as will policies to map IP traffic to the different L2-LSPs. Therefore, I would expect that the system software (and possibly hardware) will need to be designed from the ground up to be able to support large numbers of L2-LSPs. 6. Enterprise networks are built using not just switches but routers as well. Firstly, for L2-LSPs to be set up between switches and routers, they must belong to the same address space to communicate with each other. However, this address space must be separate from the IPv4 forwarding address space. To use an example of why this is true, routers connected via Ethernet segments are considered to be directly connected, however, if the IPv4 and L2-LSP address spaces were the same, then IP routers separated by GMPLS switches would be multiple hops away from each other. In addition to address space separation, routing protocol instances must also be separated. This need for this separation is obvious when considering using a link state routing protocol which relies on a topological view of the network, i.e. the IPv4 and L2-LSP topologies are different. Secondly, routers break up Ethernet segments and must therefore must always be the headend or tailend for a L2-LSPs, i.e. they cannot be midpoints. This means that from a control plane perspective, on receiving a L2-LSP path message a router must terminate the LSP and send back a resv message whether it is the intended destination/sink or not. It must then originate a path message for a new L2-LSP and send it onto the next hop GMPLS switch. This is not how RSVP-TE works today and therefore will require modifications. In addition, traffic will be blackholed if transmitted by the source before all the separate L2-LSPs have been setup end-to-end to the sink/destination. This means that some kind of end-to-end signalling message will need to be sent by the sink/destination once the final L2-LSP has been established to inform the source that the end-to-end connectivity has been provided. This functionality is not supported by RSVP-TE today and will therefore require protocol modifications. Thirdly, routers forward packets based on information in the IP header. So, when a router receives a packet via a L2-LSP it will strip of the L2 header and perform an IP lookup. Now, for the packet to be forwarded using the correct L2-LSP the router must have a policy that tells it which L2-LSP to use as there may be multiple L2-LSPs to the same destination, e.g. for different voice calls or video downloads. However, before a policy can even be configured, the router needs to know what traffic (e.g. based on src/dest address, port number, DSCP, etc.) to map to which L2-LSP, and this information needs to come from the traffic source using signalling, i.e. RSVP-TE. I do not believe this is supported today and will therefore require modifications. Regards, Richard > -----Original Message----- > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > Sent: 14 July 2005 04:24 > To: Spencer,R,Richard,XDE73 R; ccamp@ops.ietf.org > Subject: RE: End-to-end L2-LSP > > Dear Richard, > > See in-line please, > > >Thank you for taking the time to respond to my questions. I > anticipated > >that your response would include "over-provisioning is the > best answer" > >to decrease the chance of congestion on non-GMPLS switches. > >This is what is done in the LAN environment today in > conjunction with CoS, > >and this is why I believe there will be two opposing views on how > >far L2-LSP connections should be extended down to the user. One view > >is that over provisioning and CoS work fine in the > enterprise network today, > >and therefore the L2-LSP should only be extended down to the point > >at which the enterprise network connects to the SP network, > >e.g. Ethernet DSLAM. I think the other view will be that > over provisioning > >and CoS in the enterprise network don't work and therefore L2-LSPs > >should be provided end-to-end using only GMPLS capable switches. > >I wouldn't have thought there would be much call for mixed > >GMPLS/non-GMPLS enterprise networks (except for during migration) > >due to the opposing views, e.g. if over provisioning is good > enough for > >one switch, why isn't it good enough for the other? > > Over-provisioned network is non-guaranteed network. It can't prevent > occasional congestion at any rate. Further, network condition > in private > network is very diverse. You can't predict minimum service level > inside every private network. > > Suppose that an ISP want to provide high-quality, video-phone service > for enterprise users. Nothing can be guaranteed inside the > over-provisioned > customer network. If the service level is not satisfactory, > customers will complain > to service provider even if the main cause is in his network > or destination network. > We need some structure that coordinate service quality in > both private sector, > as well as in public sector consistently. > An end-to-end signal is an effective means to communicate between > user terminal, private switches and provider switches, and > concert service > quality at each different operations network. This is a goal > that RSVP (RFC2205) > pursued, however ineffective in local network because > Ethernet industry > didn't support RSVP over Ethernet (RFC2814, 2815, 2816). > Now we have another opportunity to realize the goal using L2-LSP. > Therefore, I believe end-to-end LSP is necessary even in > enterprise network > for the purpose of providing guaranteed commercial service. > > >I think where most people will agree is that using connections > >in the SPs access network does make sense in order to > provide efficient > >link utilisation and per connection policing to ensure one customers > >traffic can not affect another's. Regarding the ability to > set up new > >connections based on a request from the user (rather than > constraining > >RSVP signalling to inside the SP network), I think this will come > >down to enterprise customer demand for improved service performance. > >If a customer wasn't getting the performance they needed from their > >SP and was offered a new service that allowed them to setup > new L2-LSP > >connections on demand that would offer better performance > (e.g. for voice > >calls or video conferences), then upgrading the user devices > to support > >the new service could well be an attractive proposition. If > on the other > > Yes, this will be one of prospective service newly enabled > using L2-LSP. > > >hand a customer buys a best effort service and considers the > performance > >to be 'good enough', then there would be no reason to go to > the expense > >of upgrading the user devices to support the set up of L2-LSPs. > > Ever increasing besteffort bandwidth will eventually not satisfy > both providers and customers for many reasons. Service providers > will have difficulty in collecting expenses enough to invest > on increased > bandwidth. This will in turn gradually downgrade overall > communication quality, > and applications requiring real-time service will never > satisfy customers. > We already see such symptom in some low-price service providers. > I believe new service structure that can provide difference > of service > quality will offer benefit to customers, providers as well as > vendors in the end. > > I also look forward to see you at Paris and continue helpful > discussion. > > Sincerely > > Jaihyung > > > > > Dr. Jaihyung Cho > ETRI, Korea > phone : 042) 860-5514 > oversea: +82-42-860-5514 > fax: +82-42-861-5550 > > > > > -----?? ???----- > ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com> > ?? ??: 2005-07-14 ?? 1:10:30 > ?? ??: "jaihyung@etri.re.kr" <jaihyung@etri.re.kr>, > "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> > ??: > ??: RE: End-to-end L2-LSP > > > > Jaihyung, > > Thank you for taking the time to respond to my questions. I > anticipated that your response would include > "over-provisioning is the best answer" to decrease the chance > of congestion on non-GMPLS switches. This is what is done in > the LAN environment today in conjunction with CoS, and this > is why I believe there will be two opposing views on how far > L2-LSP connections should be extended down to the user. One > view is that over provisioning and CoS work fine in the > enterprise network today, and therefore the L2-LSP should > only be extended down to the point at which the enterprise > network connects to the SP network, e.g. Ethernet DSLAM. I > think the other view will be that over provisioning and CoS > in the enterprise network don't work and therefore L2-LSPs > should be provided end-to-end using only GMPLS capable > switches. I wouldn't have thought there would be much call > for mixed GMPLS/non-GMPLS enterprise networks (except for > during migration) due to the opposing views, e.g. if over > provisioning is good enough for one switch, why isn't it good > enough for the other? > > I think where most people will agree is that using > connections in the SPs access network does make sense in > order to provide efficient link utilisation and per > connection policing to ensure one customers traffic can not > affect another's. Regarding the ability to set up new > connections based on a request from the user (rather than > constraining RSVP signalling to inside the SP network), I > think this will come down to enterprise customer demand for > improved service performance. If a customer wasn't getting > the performance they needed from their SP and was offered a > new service that allowed them to setup new L2-LSP connections > on demand that would offer better performance (e.g. for voice > calls or video conferences), then upgrading the user devices > to support the new service could well be an attractive > proposition. If on the other hand a customer buys a best > effort service and considers the performance to be 'good > enough', then there would be no reason to go to the expense > of upgrading the user devices to support the set up of L2-LSPs. > > I look forward to the discussions in Paris. > > Regards, > Richard > > > -----Original Message----- > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > > Sent: 13 July 2005 15:14 > > To: Spencer,R,Richard,XDE73 R; ccamp@ops.ietf.org > > Subject: RE: End-to-end L2-LSP > > > > > > > > > > Dear Richard Spencer > > > > see in-line > > > > >Jaihyung, > > > > > >Perhaps I'm becoming too much of a pessimist;-) > > > > oh, not at all, > > you are reasonable enough to dig up unclear knowledge > > please go ahead until you finally get satisfactory answer. > > > > > > >Ignoring the commercial aspects, please can you help clarify > > a few technical > > >questions I have in order to help me gain a better > > understanding of the proposal? > > > > > >1. You say that one of the key reasons for using GMPLS in > > the enterprise > > >network is to provide end-to-end QoS control. However, you > > also say that not > > >all the switches will need to support GMPLS, only those at > > bottle necks. > > >Please can you expand on this to explain how you can have > end-to-end > > >QoS control when the L2-LSP is not end-to-end? > > > > > > The contextual meaning of 'End-to-end LSP' is that GMPLS label is > > assigned from source terminal and removed at destination terminal, > > because RSVP-TE message is exchanged from host to host. > > It may not perfectly guarantee QoS when the GMPLS deployment > > is not complete, however it will at least be good enough > > for end-user applications to tell those GMPLS implemented > L2 switches > > what frame must be given priority service when congestion occur. > > > > > > >As being discussed > > >on the PWE3 mailing list, the performance a service receives > > is only as > > >good as the worst performing server layer. Therefore, if > there is a > > >standard 802.1 workgroup switch between the user and the > > L2-LSP that is > > >experiencing congestion (e.g. due to a large amount of > > peer-to-peer traffic), > > >then the QoS that the user receives will only be as good as > > that provided > > >by the congested workgroup switch. > > > > > > of course, ETH-GMPLS does not prevent congestion inevitably > occurring > > in those non-GMPLS deployed part. > > > > > > >The only way I can see a network using a mixture of GMPLS > > and non-GMPLS > > >switches guaranteeing end-to-end QoS would be to design the > > network in > > >such a way that the non-GMPLS switches could never become > congested. > > > > > > Yes, that will be the way wise network designer improve > their network. > > > > > > >To do this would require that RSVP requests be used for ALL > > traffic flows > > >to ensure that there was enough bandwidth available, along > > with strict > > >per L2-LSP policing on the user devices (e.g. PC). > > >However, the obvious flaw in this approach is that someone could > > >just come along and connect > > >an end device to a non-GMPLS switch and start transmitting traffic > > >(without performing CAC/policing) causing the switch to > > become congested. > > >Therefore, please can you explain how end-to-end QoS can be > > guaranteed > > >in a mixed non-GMPLS and GMPLS switching environment? > > > > > > Of course traffic came from non controlled ingress can not be > > prevented > > in non-GMPLS node. There's no magic. Only the > > over-provisioning is the > > best answer. However, once the aggregated traffic pass 'ANY' one of > > GMPLS enabled node, those uncontrolled traffic will suffer > > queuing discipline. > > > > > > >2. Section 5.1 currently states "When the customer initiates > > data transmission, > > >the access switch maps the user flow into the L2 LSP. > > Mapping procedure is > > >subject to implementation choices, and is out of the scope of this > > >document." In order to be able to map any traffic, first of > > all the switch needs > > >to know what type of traffic (e.g. based on source/dest MAC, > > VLAN ID, 802.1p > > >bits, etc) should be mapped to the L2-LSP. In Figure 1, if > > an L2-LSP is set up > > >from the DSLAM to the edge router based on a request from > > the user (e.g. via > > >a PC or VoIP phone), then won't the user also need to signal > > what type of > > >traffic should to be mapped to the L2-LSP? > > > > > > Thank you for pointing out important thing. > > That's the most unfortunate part I reluctantly agreed because the > > work condition in the framework DT was not to span L2-LSP > > beyond provider network. > > As you noticed, mapping user traffic at provider access is awkward. > > A smooth method will be for user terminal aware of the L2 label > > to use by receiving RSVP-TE message. > > However, there was strong argument that RSVP-TE must not > > go beyond provider network because the DT is not chartered > > to work in customer network. > > That's why I requested amendment of CCAMP charter > > to include user terminal explicitly in CCAMP work scope, so that > > avoid such silly argument in future. > > > > > > >3. In section 5.1, the text states that the aggregation > > switch between the > > >DSLAM and edge router may be a 802.1 switch, i.e. it doesn't > > have to support > > >GMPLS. To set up a connection, first of all there needs to > > be a data plane > > >capable of carrying control plane traffic between GMPLS > > peers. Secondly, > > >the GMPLS peers need reachability information in order to > > forward control > > >messages onto the next peer (i.e. a static or IGP route). If > > the GMPLS peers > > >(e.g. the DSLAM and edge router in Figure 1) are not > > directly connected, > > >then the GMPLS control packets will need to be forwarded by > > the 802.1 switch. > > >Would the idea be to have say a "control traffic VLAN" > > configured between the > > >DSLAM and edge router via the transit 802.1 switch so that > > control packets > > >could be forwarded transparently by the 802.1 switch? > > > > > > Although it doesn't preclude use of VLAN, I think such > > special configuration > > for RSVP-TE control message is not necessary in legacy switches. > > Legacy switches only need to pass RSVP-TE messages as they > > pass normal data frames. Only the GMPLS implemented switches > > need to distinguish frames containing RSVP-TE messages and > > snoop them. > > Detail of implementation is, of course, for future discussion. > > > > > > >4. If we assume that control messages can be transparently > forwarded > > >by the intermediate 802.1 aggregation switch then how are > > user frames > > >forwarded in the data plane by the 802.1 switch? > > > > > > Legacy switches do not distinguish control/data frames. > > > > > > >If the DSLAM sets up a L2-LSP > > >to the edge router using say VLAN ID 99, how does the switch > > know that any > > >packets coming in with a VLAN ID of 99 should be forwarded > > out of the port > > >connecting to the edge router? > > > > > > oh, you misunderstood the L2-LSP. > > The question you are asking depends on how L2 label is encoded. > > there was hot debate in the DT about the issue how L2 label > > is encoded. > > My proposal is to use part of MAC address for label and the > > other people > > argued to use VLAN ID for label. The debate was not concluded > > because the goal of the L2SC framework document is not to select > > implementation method but to draw necessary work scope. > > I believe this issue will be dealt in depth after Paris meeting. > > So, please reserve your question until then and ask to those > > VLAN proponents. > > > > > > >5. During connection setup, I take it that the GMPLS peers > > assume that > > >resources are available within the intermediate 802.1 switch > > to support > > >the connection (as its transparent). However, as there is no > > >CAC/policing being performed by the 802.1 switch > > >(the L2-LSPs are transparent), and due to the any-to-any and > > broadcast > > >based nature of Ethernet (i.e. its CL properties), then > how can this > > >assumption be made? This comes back to my original question, how > > >can end-to-end QoS be guaranteed in an environment where there is > > >a mixture of 802.1 and GMPLS enabled switches? > > > > > > yes, you came back to the question I answered in above ;-) > > so, please see above. > > GMPLS enabled nodes must be positioned at points congestion > > is likely to be expected. > > > > Thanks > > > > Jaihyung > > > > > > ============================================ > > > > > > > > -----?? ???----- > > ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com> > > ?? ??: 2005-07-13 ?? 8:53:13 > > ?? ??: "jaihyung@etri.re.kr" <jaihyung@etri.re.kr>, > > "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> > > ??: > > ??: RE: End-to-end L2-LSP > > > > > > > > > > Jaihyung, > > > > Perhaps I'm becoming too much of a pessimist;-) Ignoring the > > commercial aspects, please can you help clarify a few > > technical questions I have in order to help me gain a better > > understanding of the proposal? > > > > 1. You say that one of the key reasons for using GMPLS in the > > enterprise network is to provide end-to-end QoS control. > > However, you also say that not all the switches will need to > > support GMPLS, only those at bottle necks. Please can you > > expand on this to explain how you can have end-to-end QoS > > control when the L2-LSP is not end-to-end? As being discussed > > on the PWE3 mailing list, the performance a service receives > > is only as good as the worst performing server layer. > > Therefore, if there is a standard 802.1 workgroup switch > > between the user and the L2-LSP that is experiencing > > congestion (e.g. due to a large amount of peer-to-peer > > traffic), then the QoS that the user receives will only be as > > good as that provided by the congested workgroup switch. > > > > The only way I can see a network using a mixture of GMPLS and > > non-GMPLS switches guaranteeing end-to-end QoS would be to > > design the network in such a way that the non-GMPLS switches > > could never become congested. To do this would require that > > RSVP requests be used for ALL traffic flows to ensure that > > there was enough bandwidth available, along with strict per > > L2-LSP policing on the user devices (e.g. PC). However, the > > obvious flaw in this approach is that someone could just come > > along and connect an end device to a non-GMPLS switch and > > start transmitting traffic (without performing CAC/policing) > > causing the switch to become congested. Therefore, please can > > you explain how end-to-end QoS can be guaranteed in a mixed > > non-GMPLS and GMPLS switching environment? > > > > 2. Section 5.1 currently states "When the customer initiates > > data transmission, the access switch maps the user flow into > > the L2 LSP. Mapping procedure is subject to implementation > > choices, and is out of the scope of this document." In order > > to be able to map any traffic, first of all the switch needs > > to know what type of traffic (e.g. based on source/dest MAC, > > VLAN ID, 802.1p bits, etc) should be mapped to the L2-LSP. In > > Figure 1, if an L2-LSP is set up from the DSLAM to the edge > > router based on a request from the user (e.g. via a PC or > > VoIP phone), then won't the user also need to signal what > > type of traffic should to be mapped to the L2-LSP? > > > > 3. In section 5.1, the text states that the aggregation > > switch between the DSLAM and edge router may be a 802.1 > > switch, i.e. it doesn't have to support GMPLS. To set up a > > connection, first of all there needs to be a data plane > > capable of carrying control plane traffic between GMPLS > > peers. Secondly, the GMPLS peers need reachability > > information in order to forward control messages onto the > > next peer (i.e. a static or IGP route). If the GMPLS peers > > (e.g. the DSLAM and edge router in Figure 1) are not directly > > connected, then the GMPLS control packets will need to be > > forwarded by the 802.1 switch. Would the idea be to have say > > a "control traffic VLAN" configured between the DSLAM and > > edge router via the transit 802.1 switch so that control > > packets could be forwarded transparently by the 802.1 switch? > > > > 4. If we assume that control messages can be transparently > > forwarded by the intermediate 802.1 aggregation switch then > > how are user frames forwarded in the data plane by the 802.1 > > switch? If the DSLAM sets up a L2-LSP to the edge router > > using say VLAN ID 99, how does the switch know that any > > packets coming in with a VLAN ID of 99 should be forwarded > > out of the port connecting to the edge router? > > > > 5. During connection setup, I take it that the GMPLS peers > > assume that resources are available within the intermediate > > 802.1 switch to support the connection (as its transparent). > > However, as there is no CAC/policing being performed by the > > 802.1 switch (the L2-LSPs are transparent), and due to the > > any-to-any and broadcast based nature of Ethernet (i.e. its > > CL properties), then how can this assumption be made? This > > comes back to my original question, how can end-to-end QoS be > > guaranteed in an environment where there is a mixture of > > 802.1 and GMPLS enabled switches? > > > > Thanks, > > Richard > > > > > -----Original Message----- > > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > > > Sent: 12 July 2005 09:21 > > > To: Spencer,R,Richard,XDE73 R; ccamp@ops.ietf.org > > > Subject: RE: End-to-end L2-LSP > > > > > > > > > > > > > > > Dear Richard, > > > > > > Yes, I believe it is not only commercially profitable but it > > > may also grow total > > > revenue of service operators by introducing new services. > > > > > > For example, think the example of IP phone you referred. > > > How do you think you can provide reliable VoIP service and > > > safely authenticate phone users and charge fees as counted by > > > packet usage > > > using simple Ethernet based xDSL network ? > > > > > > There's hardly new service that providers may offer using > > > Ethernet based > > > architecture, and collect sufficient profit to investment. > > > > > > A sophisticated ATM switch might do the work if it is not > > expensive. > > > However, my intension is modifiying existing design of > > > Ethernet hardware > > > with a little bit additional cost, and provide the level of > > > service that > > > ATM switches might offer. > > > > > > We have initial design of the Ethernet-MPLS hardware. I > don't think > > > GMPLS enabled Ethernet switch will be so complex. The > market price > > > of GMPLS enabled switch will not be so expensive than normal > > > Ethernet switches. > > > > > > Further, it only requires upgrade of switches in some > > > bottleneck points of network. > > > Read the framework documents of L2SC. It doesn't require > > > entire replacement > > > of switches to introduce new service. > > > > > > Every year, ISPs purchase new access switches and replaces > > > old machines. > > > With a bit additional cost, ISPs may use the GMPLS enabled > > switch for > > > normal Ethernet switching in initial deployment. > > > > > > When the GMPLS function is enabled in some upgraded part of > > network, > > > ISPs may provide high quality, reliable VoIP service to > > xDSL customers > > > using RSVP-TE enabled VoIP phone. > > > The high quality video-phone service enabled by GMPLS will be > > > competable to legacy voice only phone service. > > > > > > > > > There are lots of things I can argue about commercial aspects > > > of this technique. > > > However I don't think it is not appropriate to discuss this > > > issue in this mailing list. > > > All I can say is, > > > yes, I believe that it may be commercially viable. > > > > > > > > > Thanks > > > > > > Jaihyung > > > > > > ================================================ > > > > > > Do you really think the benefits of extended GMPLS down to > > > the end user could ever justify the costs involved in > > > upgrading LAN switches and things like IP phones (simple low > > > cost commodity items) to support GMPLS signalling/routing? > > > > > > Your proposal may be technically possible (though it would > > > require new hardware and significant software upgrades), but > > > commercially I would consider this to be a non-starter. > > > > > > Regards, > > > Richard > > > > > > > -----Original Message----- > > > > From: owner-ccamp@ops.ietf.org > > [mailto:owner-ccamp@ops.ietf.org]On > > > > Behalf Of CHO, JAI HYUNG > > > > Sent: 12 July 2005 01:10 > > > > To: ccamp@ops.ietf.org > > > > Subject: End-to-end L2-LSP > > > > > > > > > > > > > > > > Dear all, > > > > > > > > Greetings. > > > > > > > > I am a co-editor of the Ethernet-GMPLS framework document. > > > > (draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt) > > > > > > > > I am writing this to explain the motivation behind the > > > > proposal of scenario-1, and to raise the issue of > > > > end-to-end LSP to be included in CCAMP charter. > > > > > > > > An important goal intended in scenario-1 is > > > > QoS control over end-to-end packet delivery using LSP. > > > > In other words, the purpose of scenario-1 is extending coverage > > > > of GMPLS control including customer network and > end-user device, > > > > such as PC, VoIP phone, IP-TV, or even future 4G mobile phone. > > > > End-user applications will have a means to request resources > > > > to network using GMPLS signal. > > > > > > > > Unfortunately, this view is not included in the > > > > framework because there was strong opinion that > > > > L2-LSP must not span beyond provider network because > > > > CCAMP is not chartered to work on customer > > > > network. I am wondering if this is right argument. > > > > Perhaps the charter can be amended to include this work. > > > > > > > > I believe there is little technical reason to preclude > > > > LSP being used in customer network. It is a reasonable > > > > assumption that private company may also use GMPLS > > > > implemented Ethernet switch, once the L2SC work is successful. > > > > Currently, CCAMP is the only place discussing the matter of > > > > Ethernet and LSP. There is no other WG dealing similar issue. > > > > If CCAMP wants to work on Ethernet, > > > > I think the coverage of common GMPLS control must be > > > > extended to include all area where Ethernet is deployed. > > > > Currently, Ethernet is dominantly deployed in first-mile > > > > network where access network as well as customer network > > > > and user hosts are usually included. Therefore, user host must > > > > be considered as initiation and termination point of L2-LSP. > > > > > > > > Another important reason to include user host in CCAMP > > > > work scope is to achieve consistent end-to-end control of QoS. > > > > End-to-end QoS has long been desired goal since RSVP was > > > > first designed. This goal can be best achieved using > > > > end-to-end GMPLS signal and L2-LSP. > > > > Currently, RSVP (RFC2205) is available in many user hosts. > > > > There is only a little difference between RSVP and RSVP-TE. > > > > If the RSVP-TE signal of provider network can reach user host > > > > across Ethernet based private network, an LSP can be > established > > > > from application to application via the provider backbone. > > > > Service providers will be able to control application flows > > > > in L2-LSP level as well as in aggregated LSP pipe of similar > > > > application type. A variety of new services will be > enabled using > > > > the enhanced capability of distinguishing and controlling each > > > > individual application flow. > > > > > > > > I personally believe that end-to-end L2-LSP may offer > > > > innovative solutions for servicing IP-TV as well as fast > > mobility. > > > > There have been some experimental approaches combining > > > > MPLS and IP mobility. However such potential of the new > > > > application can only be explored when end-user host is > > > > included in the scope of CCAMP work area. > > > > > > > > Therefore, I hereby sincerely request people's attention > > > > and support on this perspective of end-to-end LSP. > > > > If there is enough number of people agree on this view, > > > > the WG can request the necessary update of the charter. > > > > > > > > Thank you. > > > > > > > > Jaihyung Cho > > > > > > > > > > > > > > > > ETRI, Korea > > > > phone : 042) 860-5514 > > > > oversea: +82-42-860-5514 > > > > fax: +82-42-861-5550 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 10:06:30 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: RE: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Date: Mon, 25 Jul 2005 11:06:07 +0100 Message-ID: <0536FC9B908BEC4597EE721BE6A353890A9F1E37@i2km07-ukbr.domain1.systemhost.net> Thread-Topic: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Thread-Index: AcWPjVr8JAiKMqpSRTGCmwFxlvJM3QBcYn1w From: <neil.2.harrison@bt.com> To: <i_bryskin@yahoo.com>, <adrian@olddog.co.uk> Cc: <ccamp@ops.ietf.org> Hi Igor, >=20 > > I am also a little concerned that some folks might > > naively assume that a > > control-plane is going to solve all their > > management-plane > > issues......when we move down the network stack to > > the duct there are a > > whole raft of factors that mean the management-plane > > still dominates > > here. If you talk to the right folks in the > > operators who run these > > types of network you will indeed understand what I > > mean by this. >=20 > This kind of statements you've been making in about > every email. I am curious to know as we move towards > the duct which *exactly* features, provisioning > functions or operations that are implemented via NMS > could not be conceptually realized via control plane?=20 > Or you are not the "right folk to ask" ? :=3D) NH=3D> To properly answer your question will take me a couple of pages. = I can help/answer to a fair degree, but if you are really serious about understanding this stuff better I do think you should seek out the right folks in operators (not sure you will get all the feedback you need if you are solely reliant on IETF lists). Anyway, I'll send you a detailed reponse off list...and I also note Alan McGuire here has sent you some stuff off-list on a similar theme wrt to our experiences in BT (quite compelling IMO). regads, Neil Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 10:00:02 +0000 Sensitivity: Subject: RE: Frameformat in a l2cs gmpls rnvironment. To: "<richard.spencer" <richard.spencer@bt.com> Cc: "<jaihyung" <jaihyung@etri.re.kr>, "<ccamp" <ccamp@ops.ietf.org> Message-ID: <OFDD7C95E0.72074290-ONC1257049.0036BD9F-C1257049.0036E37D@uk.marconicomms.com> From: "Diego Caviglia" <Diego.Caviglia@marconi.com> Date: Mon, 25 Jul 2005 11:59:27 +0200 MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii Hi Richard, actually the usage of a specific VLAN tag for the control plane seems to me a good idea. It is a fast and simple way to distinguish data plane traffic from control plane traffic. Regards Diego <richard.spencer@bt.com>@ops.ietf.org on 25/07/2005 11.54.36 Sent by: owner-ccamp@ops.ietf.org To: <jaihyung@etri.re.kr> cc: <ccamp@ops.ietf.org> Subject: RE: Frameformat in a l2cs gmpls rnvironment. Jaihyung, As both you and Adrian thought that I was talking about an Ethernet Control plane, I mustn't have adequately articulated my point. When I mentioned a "data plane for control traffic" below, I was not talking about an Ethernet control plane. I was referring to the transport of GMPLS IP control protocol traffic over a standard 802.1 data plane, using either untagged frames or a common VLAN ID. Regards, Richard > -----Original Message----- > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > Sent: 23 July 2005 05:39 > To: Spencer,R,Richard,XDE73 R > Cc: ccamp@ops.ietf.org > Subject: RE: Frameformat in a l2cs gmpls rnvironment. > > > > Richard, > > IEEE recommends to use Ethernet Length/Type code for > distinguishing control frames. > All GARP, GVRP, GMPR and other proposed OAM frame follow this rule. > IEEE particulary discourages using some number in data > forwarding address > being used for exchange of control data. An example is using > some multicast address > for exchange of hello frames. This rule is also applied to > the case of VLAN forwarding. > I think this policy is to prevent control chaos with other L2 > protocols. > We need to be careful in taking approach to use some L2-LSP > exclusively > for control plane because it may result in unwanted conflict > with both people and machine. > > Regards, > > Jaihyung > > > -----?? ???----- > ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com> > ?? ??: 2005-07-22 ?? 8:44:08 > ?? ??: "per@defero.se" <per@defero.se>, "loa@pi.se" <loa@pi.se> > ??: "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> > ??: RE: Frameformat in a l2cs gmpls rnvironment. > > > > Regardless of whether or not a switch is directly connected > to hosts, it must be able to forward packets using the > connectionless Ethernet data plane. This is due to the > fundamental requirement that for GMPLS switches to be able to > exchange control information with each other, a data plane > for control traffic must be present. This is akin to using > the IP data plane for MPLS signalling in an IP/MPLS network. > An alternative would be to use a static reserved L2-LSP for > control traffic in the same way that reserved VPI/VCIs are > used for PNNI signalling in ATM. > > Regarding connecting hosts to GMPLS switches, I personally > don't think extending L2-LSPs into the enterprise/home > network is commercially viable. However, if you do want to > use GMPLS switches in the home/enterprise network and for > some reason don't want to extend L2-LSPs down to the host > then you will not be performing normal Ethernet Mac address > switching anyway. Instead you will need some kind of policy > on the switch that maps connectionless Ethernet packets (e.g. > based on MAC src/dest, 802.1p, VLAN) to a L2-LSP. This is > because multiple L2-LSPs to the same destination (e.g. a > gateway router) may exist for different services/flows (e.g. > video download, VoIP call, etc.). > > Regards, > Richard > > > -----Original Message----- > > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On > > Behalf Of Par Mattsson > > Sent: 22 July 2005 11:42 > > To: Loa Andersson > > Cc: ccamp@ops.ietf.org > > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > > > > > Per and Dimitri, > > > > > > I would like to come down stronger than that, for me it is > > > a very strong requirement that the same switch can handle > > > both VLANs and GMPLs trafic correctly. I can't dsee how that > > > could be done if using the VLAN tpid to indicate GMPLS > > > traffic. > > > > If you ever want that same switch to handle traffic for a directly > > connected host (not to uncommen) you would want that to use normal > > ethernet macaddress switching. So of course you do not want > to have to > > choose between vlan and gmpls, you would want both at the > same time. > > > > /per > > > > > > > > > > /Loa > > > > > > Par Mattsson wrote: > > >>>hi par, one of the possibilities that has been considered > > to cope with > > >>>this requirement is to use a dedicated TPID for the > > Ethernet labeled > > >>>frames; this would allow differentiated processing with > non-labeled > > >>>framesthanks. > > >> > > >> > > >> That seems to make more sence. If that frame is to be > sized like a > > >> 802.1q > > >> frame. There is not that much space left to a label. Or is > > the demand to > > >> use jumboframes ? > > >> Has there been any discussion on labelstacking, and > mainly where to > > >> place > > >> the information? > > >> > > >> Regards. > > >> Per > > >> > > >> > > >> > > > > > > > > > -- > > > Loa Andersson > > > > > > Principal Networking Architect > > > Acreo AB phone: +46 8 632 77 14 > > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > > Kista, Sweden email: loa.andersson@acreo.se > > > loa@pi.se > > > > > > > > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 09:55:06 +0000 Content-Class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Mon, 25 Jul 2005 10:54:36 +0100 Message-ID: <B5E87B043D4C514389141E2661D255EC0A835AEA@i2km41-ukdy.domain1.systemhost.net> Thread-Topic: Frameformat in a l2cs gmpls rnvironment. Thread-Index: AcWPQGv1s5h7skVjT2KlrNODX+ABpABrRlWQ From: <richard.spencer@bt.com> To: <jaihyung@etri.re.kr> Cc: <ccamp@ops.ietf.org> Jaihyung,=20 As both you and Adrian thought that I was talking about an Ethernet = Control plane, I mustn't have adequately articulated my point. When I = mentioned a "data plane for control traffic" below, I was not talking = about an Ethernet control plane. I was referring to the transport of = GMPLS IP control protocol traffic over a standard 802.1 data plane, = using either untagged frames or a common VLAN ID. Regards, Richard > -----Original Message----- > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > Sent: 23 July 2005 05:39 > To: Spencer,R,Richard,XDE73 R > Cc: ccamp@ops.ietf.org > Subject: RE: Frameformat in a l2cs gmpls rnvironment. >=20 >=20 > =20 > Richard,=20 > =20 > IEEE recommends to use Ethernet Length/Type code for=20 > distinguishing control frames. > All GARP, GVRP, GMPR and other proposed OAM frame follow this rule. > IEEE particulary discourages using some number in data=20 > forwarding address=20 > being used for exchange of control data. An example is using=20 > some multicast address=20 > for exchange of hello frames. This rule is also applied to=20 > the case of VLAN forwarding. > I think this policy is to prevent control chaos with other L2=20 > protocols. > We need to be careful in taking approach to use some L2-LSP=20 > exclusively=20 > for control plane because it may result in unwanted conflict=20 > with both people and machine. > =20 > Regards, > =20 > Jaihyung > =20 >=20 > -----?? ???-----=20 > ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com>=20 > ?? ??: 2005-07-22 ?? 8:44:08=20 > ?? ??: "per@defero.se" <per@defero.se>, "loa@pi.se" <loa@pi.se>=20 > ??: "ccamp@ops.ietf.org" <ccamp@ops.ietf.org>=20 > ??: RE: Frameformat in a l2cs gmpls rnvironment.=20 >=20 >=20 >=20 > Regardless of whether or not a switch is directly connected=20 > to hosts, it must be able to forward packets using the=20 > connectionless Ethernet data plane. This is due to the=20 > fundamental requirement that for GMPLS switches to be able to=20 > exchange control information with each other, a data plane=20 > for control traffic must be present. This is akin to using=20 > the IP data plane for MPLS signalling in an IP/MPLS network.=20 > An alternative would be to use a static reserved L2-LSP for=20 > control traffic in the same way that reserved VPI/VCIs are=20 > used for PNNI signalling in ATM.=20 >=20 > Regarding connecting hosts to GMPLS switches, I personally=20 > don't think extending L2-LSPs into the enterprise/home=20 > network is commercially viable. However, if you do want to=20 > use GMPLS switches in the home/enterprise network and for=20 > some reason don't want to extend L2-LSPs down to the host=20 > then you will not be performing normal Ethernet Mac address=20 > switching anyway. Instead you will need some kind of policy=20 > on the switch that maps connectionless Ethernet packets (e.g.=20 > based on MAC src/dest, 802.1p, VLAN) to a L2-LSP. This is=20 > because multiple L2-LSPs to the same destination (e.g. a=20 > gateway router) may exist for different services/flows (e.g.=20 > video download, VoIP call, etc.).=20 >=20 > Regards,=20 > Richard=20 >=20 > > -----Original Message-----=20 > > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On=20 > > Behalf Of Par Mattsson=20 > > Sent: 22 July 2005 11:42=20 > > To: Loa Andersson=20 > > Cc: ccamp@ops.ietf.org=20 > > Subject: Re: Frameformat in a l2cs gmpls rnvironment.=20 > >=20 > >=20 > > > Per and Dimitri,=20 > > >=20 > > > I would like to come down stronger than that, for me it is=20 > > > a very strong requirement that the same switch can handle=20 > > > both VLANs and GMPLs trafic correctly. I can't dsee how that=20 > > > could be done if using the VLAN tpid to indicate GMPLS=20 > > > traffic.=20 > >=20 > > If you ever want that same switch to handle traffic for a directly=20 > > connected host (not to uncommen) you would want that to use normal=20 > > ethernet macaddress switching. So of course you do not want=20 > to have to=20 > > choose between vlan and gmpls, you would want both at the=20 > same time.=20 > >=20 > > /per=20 > >=20 > >=20 > > >=20 > > > /Loa=20 > > >=20 > > > Par Mattsson wrote:=20 > > >>>hi par, one of the possibilities that has been considered=20 > > to cope with=20 > > >>>this requirement is to use a dedicated TPID for the=20 > > Ethernet labeled=20 > > >>>frames; this would allow differentiated processing with=20 > non-labeled=20 > > >>>framesthanks.=20 > > >>=20 > > >>=20 > > >> That seems to make more sence. If that frame is to be=20 > sized like a=20 > > >> 802.1q=20 > > >> frame. There is not that much space left to a label. Or is=20 > > the demand to=20 > > >> use jumboframes ?=20 > > >> Has there been any discussion on labelstacking, and=20 > mainly where to=20 > > >> place=20 > > >> the information?=20 > > >>=20 > > >> Regards.=20 > > >> Per=20 > > >>=20 > > >>=20 > > >>=20 > > >=20 > > >=20 > > > --=20 > > > Loa Andersson=20 > > >=20 > > > Principal Networking Architect=20 > > > Acreo AB phone: +46 8 632 77 14=20 > > > Isafjordsgatan 22 mobile: +46 739 81 21 64=20 > > > Kista, Sweden email: loa.andersson@acreo.se=20 > > > loa@pi.se=20 > > >=20 > >=20 > >=20 > >=20 >=20 >=20 >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 09:45:36 +0000 Message-ID: <FF8AC5030873D6118BCB0002A58EDA990B9D2B51@mchh2a7e.mchh.siemens.de> From: Heiles Juergen <juergen.heiles@siemens.com> To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr>, Heiles Juergen <juergen.heiles@siemens.com>, per@defero.se Cc: ccamp@ops.ietf.org Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Mon, 25 Jul 2005 11:44:34 +0200 MIME-Version: 1.0 Content-Type: text/plain Dear Jaihyung, you mention several options that could be used as a label. Most of them have strong impacts on the data plane and introduce new layer networks. Is this in the scope of CCAMP? What is the goal, to introduce a control plane for a existing L2 technology or introduce a new L2 technology? What would be the benefit of such a new L2 co switching technology compared to Ethernet over MPLS as defined by PWE3 and L2VPN? You mentioned to use the MAC address or part of it as a label. The TRILL working group is defining shortest path routing for Ethernet. How would this fit together? Regards Juergen > -----Original Message----- > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > Sent: Saturday, July 23, 2005 6:19 AM > To: Heiles Juergen; per@defero.se > Cc: ccamp@ops.ietf.org > Subject: RE: Frameformat in a l2cs gmpls rnvironment. > > > Dear Juergen, Par and all, > > The proposed framework of L2SC was not intended to suggest > any specific solution. It is still an open question. > As Adrian noted, it only aims to present requirements > as in scenarios that may drive people toward next step in many > viable solutions. > > In fact, there are other options we may also consider for L2 > label encoding. > IEEE 802.1 Ethernet bridges forward frames based on 48bits of > MAC address, > and additionally using VLAN tag. > > When the purpose of GMPLS control over Ethernet is not to > create new dataplane > but to utilize IEEE 802.1 bridge architecture, we may consider using > one of the two forwarding methods, MAC forwarding or VLAN forwarding. > (and perhaps any other combination of fields in MAC, but I'll > not discuss it) > > Use of VLAN ID for label encoding may automate VLAN configuration > using IP protocols. However, GMPLS protocol cannot use the > field exclusively > because public/private operators already use VLAN for various > purpose. > There is a potential conflict with existing use of VLAN and > GMPLS use of VLAN label. > Furthermore, scalability of VLAN ID has been frequently noted > as weakness > because the size of VLAN ID is at most 4096 (12bit). > > The scalability may be improved if the scope of VLAN label is > confined to > link-local, and some additional swapping function of VLAN ID > is introduced in > Internal Sublayer [802.1D] of GMPLS implemented switch. > However, this will only be effective when the configuration > of network > is mesh structure that multiple LSP paths exist. If the > configuration of network > is star or tree shape, as normal configuration of access network, > LSPs concentrate in root node and total number of > LSPs that the network can hold still be limited by available > label space > at a few root links. > > For these reasons, I do not think any form of VLAN ID label is > an appropriate choice for layer-2 label encoding. > > There are some other proposals assuming new assignment of Ethernet > Length/Type value (e.g. new TPID in VLAN tag) and re-definition of > information fields placed between 802.3 MAC header and IP packet. > In this case, only the format of VLAN tag or extended VLAN tag > is borrowed, however, inside the switching hardware, the > filter and relay, > etc. are totally different new dataplane switch. > > I do not see such approach is a GMPLS implementation for Ethernet > because the core switching technique is not 802.1 Ethernet > bridge at all. > > The other option we may consider is using MAC address filed as below. > > > +-------+-------+-------+-------+-------+-------+ > | 1byte | 2byte | 3byte | 4byte | 5byte | 6byte | > +-------+-------+-------+-------+-------+-------+ > +-----------------------+-----------------------+ > | OUI Prefix (=GMPLS) | DA-label (24bit) | > +-----------------------+-----------------------+ > | OUI Prefix (=GMPLS) | SA-label (24bit) | > +---------------+-------+-----------------------+ > | Length/Type | > +---------------+ > > > IEEE is designated by the ISO Council to act as the > registration authority > for the higher three-octet of OUI number in the MAC address > to be used by manufacturer. Ethernet manufacturer may generate > global unique MAC address using the OUI prefix and address block of > lower three-octet (24bit). Taking advantage of the addressing scheme, > GMPLS may use the lower three-octet exclusively if a unique > OUI number > is reserved for the protocol. With this labeling scheme, GMPLS will > control MAC forwarding entry, not VLAN table. > > All Ethernet frames controlled by GMPLS will have identical > OUI number > that they can easily be distinguished from other Ethernet frames. > In principle, the label lookup hardware is identical to MAC lookup > hardware in this labeling scheme. Therefore GMPLS implemented > switch may still function as normal Ethernet bridge to the frames > that OUI number is not GMPLS. This also facilitates GMPLS implemented > switches being deployed in operating Ethernet with minimum > service disruption. > > Note also that above proposed label encoding method is transparent > to the use of Ethernet Length/Type field. End-user device may use > the Length/Type field as defined in IEEE 802.3 protocol. > It also allows network operators configure VLAN for their own > purpose. > When IEEE 802.1p is used in conjunction with L2-LSP, the priority > field of VLAN tag can also be used for imposing consistent TE > policy in > legacy switches and GMPLS switches. > > Any way, my conclusion is, there are other options we may consider, > and this issue is still open to discuss. > > Thanks, > > Sincerely, > > Jaihyung > > > > > > -----?? ???----- > ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com> > ?? ??: 2005-07-22 ?? 10:23:43 > ?? ??: "Loa Andersson" <loa@pi.se>, "richard.spencer@bt.com" > <richard.spencer@bt.com> > ??: "per@defero.se" <per@defero.se>, "ccamp@ops.ietf.org" > <ccamp@ops.ietf.org> > ??: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > > Loa, > > I interpret the ID as a proposal to use GMPLS for VLAN setup. > So GMPLS and VLAN to not compete. The VLAN is at the data > plane and GMPLS at the control plane. The question is how and > should different control plane techniques like GMPLS and > (GVRP and STB) work together? > > Regards > > Juergen > > > > -----Original Message----- > > From: owner-ccamp@ops.ietf.org > > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson > > Sent: Friday, July 22, 2005 2:53 PM > > To: richard.spencer@bt.com > > Cc: per@defero.se; ccamp@ops.ietf.org > > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > > Richard, > > > > I agree to most of this. Since we have doubts about the viability > > of taking GMPLS all the way to end-user or enterprise I think it > > would be good, from a wg perspective, if we agreed to solve the > > core network problems first. > > > > Do you have any comment on the requirement to run both VLANs and > > GMPLS on the same switch? > > > > /Loa > > > > richard.spencer@bt.com wrote: > > > Regardless of whether or not a switch is directly connected > > > to hosts, it must be able to forward packets using the > > connectionless > > > Ethernet data plane. This is due to the fundamental > > requirement that > > > for GMPLS switches to be able to exchange control > information with > > > each other, a data plane for control traffic must be present. > > > This is akin to using the IP data plane for MPLS > signalling in an > > > IP/MPLS network. An alternative would be to use a static > reserved > > > L2-LSP for control traffic in the same way that reserved > VPI/VCIs > > > are used for PNNI signalling in ATM. > > > > > > Regarding connecting hosts to GMPLS switches, I personally don't > > > think extending L2-LSPs into the enterprise/home network is > > > commercially viable. However, if you do want to use > GMPLS switches > > > in the home/enterprise network and for some reason don't want to > > > extend L2-LSPs down to the host then you will not be performing > > > normal Ethernet Mac address switching anyway. Instead you will > > > need some kind of policy on the switch that maps connectionless > > > Ethernet packets (e.g. based on MAC src/dest, 802.1p, VLAN) to a > > > L2-LSP. This is because multiple L2-LSPs to the same destination > > > (e.g. a gateway router) may exist for different services/flows > > > (e.g. video download, VoIP call, etc.). > > > > > > Regards, > > > Richard > > > > > > > > >>-----Original Message----- > > >>From: owner-ccamp@ops.ietf.org > [mailto:owner-ccamp@ops.ietf.org]On > > >>Behalf Of Par Mattsson > > >>Sent: 22 July 2005 11:42 > > >>To: Loa Andersson > > >>Cc: ccamp@ops.ietf.org > > >>Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > >> > > >> > > >> > > >>>Per and Dimitri, > > >>> > > >>>I would like to come down stronger than that, for me it is > > >>>a very strong requirement that the same switch can handle > > >>>both VLANs and GMPLs trafic correctly. I can't dsee how that > > >>>could be done if using the VLAN tpid to indicate GMPLS > > >>>traffic. > > >> > > >>If you ever want that same switch to handle traffic for a > directly > > >>connected host (not to uncommen) you would want that to > use normal > > >>ethernet macaddress switching. So of course you do not want > > to have to > > >>choose between vlan and gmpls, you would want both at the > same time. > > >> > > >>/per > > >> > > >> > > >> > > >>>/Loa > > >>> > > >>>Par Mattsson wrote: > > >>> > > >>>>>hi par, one of the possibilities that has been considered > > >> > > >>to cope with > > >> > > >>>>>this requirement is to use a dedicated TPID for the > > >> > > >>Ethernet labeled > > >> > > >>>>>frames; this would allow differentiated processing with > > non-labeled > > >>>>>framesthanks. > > >>>> > > >>>> > > >>>>That seems to make more sence. If that frame is to be > sized like a > > >>>>802.1q > > >>>>frame. There is not that much space left to a label. Or is > > >> > > >>the demand to > > >> > > >>>>use jumboframes ? > > >>>>Has there been any discussion on labelstacking, and > > mainly where to > > >>>>place > > >>>>the information? > > >>>> > > >>>>Regards. > > >>>>Per > > >>>> > > >>>> > > >>>> > > >>> > > >>> > > >>>-- > > >>>Loa Andersson > > >>> > > >>>Principal Networking Architect > > >>>Acreo AB phone: +46 8 632 77 14 > > >>>Isafjordsgatan 22 mobile: +46 739 81 21 64 > > >>>Kista, Sweden email: > loa.andersson@acreo.se > > >>> loa@pi.se > > >>> > > >> > > >> > > >> > > > > > > > > > -- > > Loa Andersson > > > > Principal Networking Architect > > Acreo AB phone: +46 8 632 77 14 > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > Kista, Sweden email: loa.andersson@acreo.se > > loa@pi.se > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 09:14:53 +0000 Sensitivity: Subject: Re: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] To: ccamp@ops.ietf.org Message-ID: <OF71CBD7C2.475B088C-ONC1257049.00310077-C1257049.0032CC03@uk.marconicomms.com> From: "Diego Caviglia" <Diego.Caviglia@marconi.com> Date: Mon, 25 Jul 2005 11:14:45 +0200 MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii Hi all, this is a very long and very interesting thread but IMHO we're losing the focus. W.r.t. the draft-papadimitriou-ccamp-gmpls-ethernet-framework-00 I think that the ID is useful and I hope that this work progress in the CCAMP. Moreover I'd like to have some words in the ID on the fact that a GMPLS enabled Eth TNE can work and interwork also with Eth TNEs that are without GMPLS control plane. In the LMP section (section 6.4) there is no mention about control channel, I'd like to have some words on this. In general I agree that CP is just a way to automatise and speed-up operation that can be done via NMS and thus seems to me that the VLAN usage as GMPLS label is good. Of course we need to go in deep on this matter but the idea seems good. Best regards Diego Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 08:56:38 +0000 Message-ID: <20050725045531.hor2bwdlblgco0k4@www.isocore.com> Date: Mon, 25 Jul 2005 04:55:31 -0400 From: Bijan Jabbari <bjabbari@isocore.com> To: dpapadimitriou@psg.com, dimitri.papadimitriou@alcatel.be Cc: Igor Bryskin <i_bryskin@yahoo.com>, neil.2.harrison@bt.com, adrian@olddog.co.uk, ccamp@ops.ietf.org Subject: Re: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format="flowed" Content-Disposition: inline Content-Transfer-Encoding: 7bit User-Agent: Internet Messaging Program (IMP) H3 (4.0.3) While we are on the subject of control vs. management plane and the need for GMPLS-based control of Ethernet TE, I thought it might be worthwhile to indicate some prior work done in this area and within the context of an experimental optical and IP network (although a few of you might be familiar with this activity, as my colleagues or myself might have shared with you the goal of the project which also involves establishing Ethernet/L2 LSPs). As part of this project (referred to as DRAGON, initiated in the early 2003 and supported by the NSF), we have implemented the GMPLS capabilities (OSPF-TE & RSVP-TE) within an element (referred to as the Virtual LSR, or VLSR) and have used it within experimental and demo networks to control Ethernet, TDM and Fiber-switched capable switches. The implementation is open source software, available free to public and it runs on FreeBSD or Linux. While VLSR participates in a GMPLS-based control network from one end, it enables the devices (Ethernet switches, cross-connects, etc. without GMPLS capability) to participate within the network, by translating specific protocol events into an appropriate sets of commands for the device to effect an associated state change. In our use of VLSR to control Ethernet switches, the LSP set up requests are processed by issuing the appropriate SNMP commands to the said device to place two ports into a single VLAN effecting a port-to-port cross-connect. For control of Ethernet switches, we have used RFC 2674 to dynamically modify VLAN configurations. Efforts are underway to add multi-region capability to VLSR, where we can have the ability to route light-paths across different networks transparently. A case of interest there is to cater to the PSC LSP when no existing path to the destination might exist, however a lower layer path, say an Ethernet L2SC path, could be established to provide the missing link for the PSC request. Regards, - Bijan Quoting dimitri papadimitriou <dpapadimitriou@psg.com>: > igor > >> What I was trying to say here is that for the User how >> TE Ethernet connection is established - via control or >> management planes is of a little importance. > > -> in the CCAMP context where the second C means "Control" it is of > importance because the purpose of this working group is to define IP > control plane mechanism(s) > >> Before >> explaining how cool/elegant it could be provisioned >> via GMPLS you need to answer IMO the following >> questions: > > -> i suggest you also take a look to section 3 and 4 and if you think > there are additional arguments you would consider as part of these > sections i would suggest that you share them with us > >> 1) What does it mean Traffic Engineered Ethernet >> connection in the data plane and what is it good for? >> For instance, we perfectly understand what does ATM >> connection or TDM connection mean in the data plane >> and can provide references to corresponding data plane >> standards. > > -> i suggest you take a look at scenario 2 for inst. that should in > principle provide an answer to your question > >> 2) How does the label-aware Ethernet forwarder >> co-exist with other Ethernet forwarders on the same >> switch > > -> it is an explicit non-objective of THIS document to detail > (existing) Ethernet forwarding behaviour and their co-existence, in > part. due to point 3) here below > >> 3) What information is necessary to provision the >> forwarding tables? > > -> the current document refers to "label" entries without delving > into details but this has been done purposely as the choice which > will ultimately be made (e.g. VLAN Tags, or new TPIDs or something > else) is orthogonal to be capabilities this working group wants to > achieve with GMPLS control of Ethernet networks > >> And so on. > > -> in practice ? > >> Cheers, >> Igor >> >> --- dimitri papadimitriou <dpapadimitriou@psg.com> >> wrote: >> >> >>> hi igor, >>> >>> [snip - hint: igor i suggest you create a thread on >>> control vs management plane] >>> >>> >>>> My view wrt Ethertnet GMPLS is this. I have no >>> >>> doubts >>> >>>> that we can come up with mechanisms to dynamically >>>> provision L2SC LSPs. >>> >>> ouf ;-) >>> >>> >>>> My problem is the order of events. It does not seem to me wise to >>>> come up with >>>> some control plane framework and solution(s) and >>> >>> after >>> >>>> that think what we need to do in the data plane. >>> >>> actually, if this is the only problem you have, i >>> should underline we do not have to initiate the definition of a specific >>> forwarding paradigm so i am not necessarily sure about what you mean by the >>> last part of the above sentence >>> >>> >>>> It seems to me wiser to learn how to statically >>>> provision such LSPs, see how useful they are, and >>> >>> only >>> >>>> after that design and develop ways to provision >>> >>> them >>> >>>> dynamically. >>> >>> i am not sure to follow how are you going to >>> determine the benefit from dynamic provisioning by statically >>> configure these >>> LSPs ? >>> >>> >>>> In other words, it is wiser to repeat what has >>> >>> been >>> >>>> done with TDM LSPs. >>> >>> wiser ? ... would you clarify ? don't you think it >>> is more appropriate to receive feedback on the appropriateness of the >>> use of GMPLS for the scenarios depicted in the problem statement document >>> rather than digress on non-rational argument such as "wiser", "careful" >>> or so... ? >>> >>> >>>> Cheers, >>>> Igor >>>> >>>> >>>> >>>> >>>>> A control-plane is useful in the major co-cs >>>>> transport networks (mainly >>>>> for S-PVCs and resilience) but its a minor player >>>>> compared to role of >>>>> the management-plane. The converse of course >>> >>> holds >>> >>>>> in a *traffic* >>>>> carrying cl-ps IP network, esp when this is in >>> >>> the >>> >>>>> context of the >>>>> public Internet.....but of course this isn't the >>>>> case here and IP is >>>>> only being used as the transport protocol in the >>> >>> OOB >>> >>>>> data-plane network >>>>> that carries the control/management-plane >>> >>> protocols. >>> >>>>> >>>>> These were points I wanted to make.....hopefully >>>>> I've done a better job >>>>> this time. >>>>> >>>>> regards, Neil >>>>> >>>>> >>>>> >>>>> -----Original Message----- >>>>> From: Adrian Farrel [mailto:adrian@olddog.co.uk] Sent: 23 July 2005 00:08 >>>>> To: Harrison,N,Neil,IKM1 R >>>>> Cc: ccamp@ops.ietf.org >>>>> Subject: Re: Ethernet Control Plane [Was: Re: >>>>> Frameformat in a l2cs >>>>> gmpls rnvironment.] >>>>> >>>>> >>>>> Hi Neil, >>>>> >>>>> I think John beat me off the blocks here, but... >>>>> >>>>> >>>>> >>>>> >>>>>>> GMPLS assumes an IP control plane. >>>>>> >>>>>> An IP control-plane? There is actually no such >>>>> >>>>> animal. Just what >>>>>> the heck does that REALLY mean in GMPLS say? >>>>> >>>>> >>>>> Let me explain. >>>>> Perhaps I should have said "IP-based control >>> >>> plane". >>> >>>>> I mean a control plane which: >>>>> - uses IP as its network protocol >>>>> - uses IP addresses to identify control plane >>>>> resources >>>>> - uses IP addresses to identify data plane >>> >>> resources >>> >>>>> within the control plane >>>>> - uses protocols developed for use in the >>> >>> Internet. >>> >>>>> >>>>>> I am not questioning IP as a cl-ps networking >>>>> >>>>> protocol *carrying* >>>>>> a signalling protocol (RSVP-TE, or dare I mention >>>>> >>>>> PNNI, or any >>>>>> signalling protocol yet to be invented) or a >>>>> >>>>> routing protocol >>>>>> (OSPF or ISO or whatever) >>>>> >>>>> >>>>> I am glad to hear it. >>>>> >>>>> >>>>> >>>>>> or even management protocols >>>>> >>>>> >>>>> Fine, but not in the remit of CCAMP. >>>>> >>>>> >>>>> >>>>>> but an 'IP Control Plane' per se means absolutely >>>>> >>>>> nothing to me.... >>>>> >>>>> Well, I think it should. I think the list of >>>>> attributes that I have >>>>> given above define a control plane based on IP. It is undoubtable >>>>> that attempts have been made to >>>>> use control planes >>>>> based on other protocols. Some have been highly >>>>> successful. Some have >>>>> been less fortunate. >>>>> >>>>> >>>>> >>>>>> ...nor should it to anyone else. >>>>> >>>>> >>>>> I think folks who were around at the beginning of >>>>> CCAMP and who were >>>>> part of the debate with the IESG will be very >>>>> familiar with where the >>>>> IETF draws the line here. >>>>> >>>>> >>>>> >>>>> >>>>>> I think some folks might need a reality check >>>>> >>>>> here....and also >>>>> >>>>> >>>>>> on the self-assumed importance of a >>> >>> control-plane. >>> >>>>> Hint: It ain't >>>>>> that important.....the management-plane (which >>> >>> may >>> >>>>> be using IP!) >>>>>> however is. >>>>> >>>>> >>>>> I am not sure how to interpret this. >>>>> It may be that you think that control plane is bad >>>>> per se, but you have >>>>> said elsewhere that you think it has value - but >>>>> much less than the >>>>> management plane. >>>>> It may be that you believe that CCAMP is willfully >>>>> neglecting the >>>>> management plane. This would, in fact, be true. It >>>>> is not in CCAMP's >>> >> === message truncated === >> >> __________________________________________________ >> Do You Yahoo!? >> Tired of spam? Yahoo! Mail has the best spam protection around >> http://mail.yahoo.com . >> > > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 25 Jul 2005 07:23:06 +0000 Message-ID: <FF8AC5030873D6118BCB0002A58EDA990B9D2AA8@mchh2a7e.mchh.siemens.de> From: Heiles Juergen <juergen.heiles@siemens.com> To: Dimitri.Papadimitriou@alcatel.be, Heiles Juergen <juergen.heiles@siemens.com> Cc: Loa Andersson <loa@pi.se>, Par Mattsson <per@defero.se>, ccamp@ops.ietf.org Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Mon, 25 Jul 2005 09:20:19 +0200 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01C590E9.501C479C" This message is in MIME format. Since your mail reader does not understand this format, some or all of this message may not be legible. ------_=_NextPart_001_01C590E9.501C479C Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Hi Dimitri, =20 what I described is the straight forward approach like it is used for = SONET/SDH or OTN. A pure control plane approach for setup of = connectivity. You could even keep the restriction of a domain local = VLAN ID as defined by IEEE. It is "just" a restriction of matrix = connectivity like some older SONET/SDH equipment also has (e.g. no time = slot interchange in a ring). So it would have no major impact on 802.1d = except for the definition of a control plane communications channel as = mentioned by Richard. In case we make any changes to the data plane behavior like disabling = of MAC address learning and MAC address based forwarding and = introduction of a dedicated label we introduce a new layer network like = MPLS did for IP. This has much wider impact and should be carefully = considered. =20 Regards =20 Juergen =20 _____ =20 From: Dimitri.Papadimitriou@alcatel.be = [mailto:Dimitri.Papadimitriou@alcatel.be]=20 Sent: Friday, July 22, 2005 4:03 PM To: Heiles Juergen Cc: Loa Andersson; Par Mattsson; ccamp@ops.ietf.org Subject: RE: Frameformat in a l2cs gmpls rnvironment. hi juergen, if well understand your point of view, you would be in favor of setting = up path without introducing any specific indication in the data plane = Ethernet frame flow more precisely, once the "logical path" is setup using the control = plane, the forwarding behaviour would remain unchanged from legacy = Ethernet behaviour (i.e MAC address learning and forwarding based on = MAC DA) in this case, the control plane function replaces the VLAN id = configuration performed today manually on a per interface basis ? - = please confirm now, i would interested to know what in your view would be the role of = 802.1d and related in this context (also there are several levels of = interaction possible here) Heiles Juergen <juergen.heiles@siemens.com> Sent by: owner-ccamp@ops.ietf.org 07/22/2005 15:17 ZE2 To: Loa Andersson <loa@pi.se>, P=E4r Mattsson <per@defero.se> cc: ccamp@ops.ietf.org bcc:=20 Subject: RE: Frameformat in a l2cs gmpls rnvironment. GMPLS for SONET/SDH has no impact on the SONET/SDH data plane. It = introduces a control plane for connection setup instead of a network = management based connection setup. GMPLS for Ethernet can work in the = same way. Ethernet as such is cl, but VLANs are setup and GMPLS can = provide a control plane for VLAN setup. The Ethernet data plane is not = changed and MAC address learning and forwarding is still be done within = a VLAN. So no change to the Ethernet data plane. No GMPLS traffic is = introduced, it is still VLAN traffic. Introducing a dedicated GMPLS label for Ethernet (GMPLS traffic) goes = beyond GMPLS as a control plane technique. Regards Juergen > -----Original Message----- > From: owner-ccamp@ops.ietf.org > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson > Sent: Friday, July 22, 2005 12:05 PM > To: P=E4r Mattsson > Cc: ccamp@ops.ietf.org > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > Per and Dimitri, > > I would like to come down stronger than that, for me it is > a very strong requirement that the same switch can handle > both VLANs and GMPLs trafic correctly. I can't dsee how that > could be done if using the VLAN tpid to indicate GMPLS > traffic. > > /Loa > > P=E4r Mattsson wrote: > >>hi par, one of the possibilities that has been considered > to cope with > >>this requirement is to use a dedicated TPID for the Ethernet = labeled > >>frames; this would allow differentiated processing with non-labeled > >>framesthanks. > > > > > > That seems to make more sence. If that frame is to be sized > like a 802.1q > > frame. There is not that much space left to a label. Or is > the demand to > > use jumboframes ? > > Has there been any discussion on labelstacking, and mainly > where to place > > the information? > > > > Regards. > > Per > > > > > > > > > -- > Loa Andersson > > Principal Networking Architect > Acreo AB phone: +46 8 632 77 14 > Isafjordsgatan 22 mobile: +46 739 81 21 64 > Kista, Sweden email: loa.andersson@acreo.se > loa@pi.se > ------_=_NextPart_001_01C590E9.501C479C Content-Type: text/html; charset="iso-8859-1" <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1"> <META content="MSHTML 6.00.2800.1505" name=GENERATOR></HEAD> <BODY> <DIV dir=ltr align=left><SPAN class=690085706-25072005><FONT face=Arial color=#0000ff size=2>Hi Dimitri,</FONT></SPAN></DIV> <DIV dir=ltr align=left><SPAN class=690085706-25072005><FONT face=Arial color=#0000ff size=2></FONT></SPAN> </DIV> <DIV dir=ltr align=left><SPAN class=690085706-25072005><FONT face=Arial color=#0000ff size=2>what I described is the straight forward approach like it is used for SONET/SDH or OTN. A pure control plane approach for setup of connectivity. You could even keep the restriction of a domain local VLAN ID as defined by IEEE. It is "just" a restriction of matrix connectivity like some older SONET/SDH equipment also has (e.g. no time slot interchange in a ring). So it would have no major impact on 802.1d except for the definition of a control plane communications channel as mentioned by Richard.</FONT></SPAN></DIV> <DIV dir=ltr align=left><SPAN class=690085706-25072005><FONT face=Arial color=#0000ff size=2>In case we make any changes to the data plane behavior like disabling of MAC address learning and MAC address based forwarding and introduction of a dedicated label we introduce a new layer network like MPLS did for IP.</FONT> </SPAN><FONT face=Arial><FONT color=#0000ff><FONT size=2>T<SPAN class=690085706-25072005>his has much wider impact and should be carefully considered.</SPAN></FONT></FONT></FONT></DIV> <DIV dir=ltr align=left><FONT face=Arial><FONT color=#0000ff><FONT size=2><SPAN class=690085706-25072005></SPAN></FONT></FONT></FONT> </DIV> <DIV><SPAN class=690085706-25072005></SPAN><FONT face=Arial><FONT color=#0000ff><FONT size=2>Regards</FONT></FONT></FONT></DIV> <DIV><FONT face=Arial><FONT color=#0000ff><FONT size=2></FONT></FONT></FONT> </DIV> <DIV><FONT><FONT><FONT face=Arial><FONT color=#0000ff><FONT size=2>J<SPAN class=690085706-25072005>uergen</SPAN></FONT></FONT></FONT></FONT></FONT></DIV> <DIV><FONT><FONT><FONT face=Arial><FONT color=#0000ff><FONT size=2><SPAN class=690085706-25072005></SPAN></FONT></FONT></FONT></FONT></FONT> </DIV> <DIV><BR></DIV> <BLOCKQUOTE dir=ltr style="PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #0000ff 2px solid; MARGIN-RIGHT: 0px"> <DIV class=OutlookMessageHeader lang=de dir=ltr align=left> <HR tabIndex=-1> <FONT face=Tahoma size=2><B>From:</B> Dimitri.Papadimitriou@alcatel.be [mailto:Dimitri.Papadimitriou@alcatel.be] <BR><B>Sent:</B> Friday, July 22, 2005 4:03 PM<BR><B>To:</B> Heiles Juergen<BR><B>Cc:</B> Loa Andersson; Par Mattsson; ccamp@ops.ietf.org<BR><B>Subject:</B> RE: Frameformat in a l2cs gmpls rnvironment.<BR></FONT><BR></DIV> <DIV></DIV> <P><FONT face=Monospace,Courier>hi juergen,</FONT></P> <P><FONT face=Monospace,Courier>if well understand your point of view, you would be in favor of setting up path without introducing any specific indication in the data plane Ethernet frame flow</FONT></P> <P><FONT face=Monospace,Courier>more precisely, once the "logical path" is setup using the control plane, the forwarding behaviour would remain unchanged from legacy Ethernet behaviour (i.e MAC address learning and forwarding based on MAC DA) in this case, the control plane function replaces the VLAN id configuration performed today manually on a per interface basis ? - please confirm</FONT></P> <P><FONT face=Monospace,Courier>now, i would interested to know what in your view would be the role of 802.1d and related in this context (also there are several levels of interaction possible here)</FONT></P> <P><BR><FONT size=2><B>Heiles Juergen <juergen.heiles@siemens.com></B></FONT><BR><FONT size=2>Sent by: owner-ccamp@ops.ietf.org</FONT><BR><FONT size=2>07/22/2005 15:17 ZE2</FONT><BR><BR><FONT size=2>To:</FONT> <FONT size=2>Loa Andersson <loa@pi.se>, Pär Mattsson <per@defero.se></FONT><BR><FONT size=2>cc:</FONT> <FONT size=2>ccamp@ops.ietf.org</FONT><BR><FONT size=2>bcc:</FONT> <BR><FONT size=2>Subject:</FONT> <FONT size=2>RE: Frameformat in a l2cs gmpls rnvironment.</FONT><BR><BR><BR></P> <P><FONT face=Monospace,Courier>GMPLS for SONET/SDH has no impact on the SONET/SDH data plane. It introduces a control plane for connection setup instead of a network management based connection setup. GMPLS for Ethernet can work in the same way. Ethernet as such is cl, but VLANs are setup and GMPLS can provide a control plane for VLAN setup. The Ethernet data plane is not changed and MAC address learning and forwarding is still be done within a VLAN. So no change to the Ethernet data plane. No GMPLS traffic is introduced, it is still VLAN traffic.<BR>Introducing a dedicated GMPLS label for Ethernet (GMPLS traffic) goes beyond GMPLS as a control plane technique.<BR></FONT><BR><FONT face=Monospace,Courier>Regards<BR></FONT><BR><FONT face=Monospace,Courier>Juergen<BR></FONT><BR><BR><BR><BR><FONT face=Monospace,Courier>> -----Original Message-----<BR>> From: owner-ccamp@ops.ietf.org<BR>> [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson<BR>> Sent: Friday, July 22, 2005 12:05 PM<BR>> To: Pär Mattsson<BR>> Cc: ccamp@ops.ietf.org<BR>> Subject: Re: Frameformat in a l2cs gmpls rnvironment.<BR>><BR>> Per and Dimitri,<BR>><BR>> I would like to come down stronger than that, for me it is<BR>> a very strong requirement that the same switch can handle<BR>> both VLANs and GMPLs trafic correctly. I can't dsee how that<BR>> could be done if using the VLAN tpid to indicate GMPLS<BR>> traffic.<BR>><BR>> /Loa<BR>><BR>> Pär Mattsson wrote:<BR>> >>hi par, one of the possibilities that has been considered<BR>> to cope with<BR>> >>this requirement is to use a dedicated TPID for the Ethernet labeled<BR>> >>frames; this would allow differentiated processing with non-labeled<BR>> >>framesthanks.<BR>> ><BR>> ><BR>> > That seems to make more sence. If that frame is to be sized<BR>> like a 802.1q<BR>> > frame. There is not that much space left to a label. Or is<BR>> the demand to<BR>> > use jumboframes ?<BR>> > Has there been any discussion on labelstacking, and mainly<BR>> where to place<BR>> > the information?<BR>> ><BR>> > Regards.<BR>> > Per<BR>> ><BR>> ><BR>> ><BR>><BR>><BR>> --<BR>> Loa Andersson<BR>><BR>> Principal Networking Architect<BR>> Acreo AB phone: +46 8 632 77 14<BR>> Isafjordsgatan 22 mobile: +46 739 81 21 64<BR>> Kista, Sweden email: loa.andersson@acreo.se<BR>> loa@pi.se<BR>></FONT></P></BLOCKQUOTE></BODY></HTML> ------_=_NextPart_001_01C590E9.501C479C-- Envelope-to: ccamp-data@psg.com Delivery-date: Sun, 24 Jul 2005 23:30:30 +0000 Message-ID: <42E42423.50208@kddilabs.jp> Date: Mon, 25 Jul 2005 08:28:35 +0900 From: Tomohiro Otani <otani@kddilabs.jp> User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; ja-JP; rv:1.7.2) Gecko/20040803 MIME-Version: 1.0 To: Kenichi Nagami <nagami@wide.ad.jp> Cc: ccamp@ops.ietf.org Subject: Re: GMPLS OSPF-TE MIB Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi Nagami-san, Thank you for your interest and comments to our draft. Our mib is intended to mainly support GMPLS OSPF-TE, but we need, in the future, to support OSPF-TE as well, whether it is in this draft or another. Arian, can we focus on GMPLS OSPF-TE first ? And then should we start to investigate OSPF-TE ? Other corrections will be fixed in the future version of the draft and are highly appreciated. With best regards, tomo Kenichi Nagami wrote: >Hi tomo, > > >This work is interesting and important from management point of view. >I have questions and comments. > > >- Does this MIB support for only GMPLS OSPF-TE or both GMPLS OSPF-TE and OSPF-TE (RFC3630) ? > > >- I think INDEX of the ospfTeLsdbEntry need to add Area ID. Because OSPF opaque LSA type 10 is > "area local scope". > > >- There are a log of error for MIB definition. Could you fix them? > For example, > 1) "--" is needed "in xxx" > current > DESCRIPTION > "Initial version. Published as RFC xxxx." -- RFC-editor pls fill > in xxx" > > > changed > DESCRIPTION > "Initial version. Published as RFC xxxx." -- RFC-editor pls fill > -- in xxx" > > > 2) need to "," in ospfTeMainGroup OBJECT-GROUP > ospfTeSwitchingType > ospfTeEncoding > are changed to > ospfTeSwitchingType, > ospfTeEncoding, > > >Regards, > > >Ken Nagami >Intec Netcore > > > >> >>At 08:52 05/07/15, Tomohiro Otani wrote: >>Hi everyone, > > >>We posted a new draft about GMPLS OSPF-TE MIB as follows. >>http://www.ietf.org/internet-drafts/draft-otani-ccamp-gmpls-ospf-mib-00.txt > > >>We welcome your comments and feedbacks, pls. >>And please make a comment whether this is within ccamp WG or not (OSPF-WG?). > > >>We believe that this is a very significant definition of OSPF-TE MIB, >>being complementary to other GMPLS related MIBs (TE/LSR/LMP mibs) in order >>to deploy GMPLS control plane in the actual environment with simple but >>reliable network management. > > >>With best regards, > > >>tomo > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Sun, 24 Jul 2005 16:31:59 +0000 Message-ID: <42E3C238.9040502@psg.com> Date: Sun, 24 Jul 2005 18:30:48 +0200 From: dimitri papadimitriou <dpapadimitriou@psg.com> Reply-To: dpapadimitriou@psg.com, dimitri.papadimitriou@alcatel.be User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.8) Gecko/20050511 MIME-Version: 1.0 To: Igor Bryskin <i_bryskin@yahoo.com> CC: dimitri.papadimitriou@alcatel.be, neil.2.harrison@bt.com, adrian@olddog.co.uk, ccamp@ops.ietf.org Subject: Re: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit igor > What I was trying to say here is that for the User how > TE Ethernet connection is established - via control or > management planes is of a little importance. -> in the CCAMP context where the second C means "Control" it is of importance because the purpose of this working group is to define IP control plane mechanism(s) > Before > explaining how cool/elegant it could be provisioned > via GMPLS you need to answer IMO the following > questions: -> i suggest you also take a look to section 3 and 4 and if you think there are additional arguments you would consider as part of these sections i would suggest that you share them with us > 1) What does it mean Traffic Engineered Ethernet > connection in the data plane and what is it good for? > For instance, we perfectly understand what does ATM > connection or TDM connection mean in the data plane > and can provide references to corresponding data plane > standards. -> i suggest you take a look at scenario 2 for inst. that should in principle provide an answer to your question > 2) How does the label-aware Ethernet forwarder > co-exist with other Ethernet forwarders on the same > switch -> it is an explicit non-objective of THIS document to detail (existing) Ethernet forwarding behaviour and their co-existence, in part. due to point 3) here below > 3) What information is necessary to provision the > forwarding tables? -> the current document refers to "label" entries without delving into details but this has been done purposely as the choice which will ultimately be made (e.g. VLAN Tags, or new TPIDs or something else) is orthogonal to be capabilities this working group wants to achieve with GMPLS control of Ethernet networks > And so on. -> in practice ? > Cheers, > Igor > > > > > --- dimitri papadimitriou <dpapadimitriou@psg.com> > wrote: > > >>hi igor, >> >>[snip - hint: igor i suggest you create a thread on >>control vs >>management plane] >> >> >>>My view wrt Ethertnet GMPLS is this. I have no >> >>doubts >> >>>that we can come up with mechanisms to dynamically >>>provision L2SC LSPs. >> >>ouf ;-) >> >> >>>My problem is the order of events. >>>It does not seem to me wise to come up with >>>some control plane framework and solution(s) and >> >>after >> >>>that think what we need to do in the data plane. >> >>actually, if this is the only problem you have, i >>should underline we do >>not have to initiate the definition of a specific >>forwarding paradigm so >>i am not necessarily sure about what you mean by the >>last part of the >>above sentence >> >> >>>It seems to me wiser to learn how to statically >>>provision such LSPs, see how useful they are, and >> >>only >> >>>after that design and develop ways to provision >> >>them >> >>>dynamically. >> >>i am not sure to follow how are you going to >>determine the benefit from >>dynamic provisioning by statically configure these >>LSPs ? >> >> >>>In other words, it is wiser to repeat what has >> >>been >> >>>done with TDM LSPs. >> >>wiser ? ... would you clarify ? don't you think it >>is more appropriate >>to receive feedback on the appropriateness of the >>use of GMPLS for the >>scenarios depicted in the problem statement document >>rather than digress >>on non-rational argument such as "wiser", "careful" >>or so... ? >> >> >>>Cheers, >>>Igor >>> >>> >>> >>> >>>>A control-plane is useful in the major co-cs >>>>transport networks (mainly >>>>for S-PVCs and resilience) but its a minor player >>>>compared to role of >>>>the management-plane. The converse of course >> >>holds >> >>>>in a *traffic* >>>>carrying cl-ps IP network, esp when this is in >> >>the >> >>>>context of the >>>>public Internet.....but of course this isn't the >>>>case here and IP is >>>>only being used as the transport protocol in the >> >>OOB >> >>>>data-plane network >>>>that carries the control/management-plane >> >>protocols. >> >>>> >>>>These were points I wanted to make.....hopefully >>>>I've done a better job >>>>this time. >>>> >>>>regards, Neil >>>> >>>> >>>> >>>>-----Original Message----- >>>>From: Adrian Farrel [mailto:adrian@olddog.co.uk] >>>>Sent: 23 July 2005 00:08 >>>>To: Harrison,N,Neil,IKM1 R >>>>Cc: ccamp@ops.ietf.org >>>>Subject: Re: Ethernet Control Plane [Was: Re: >>>>Frameformat in a l2cs >>>>gmpls rnvironment.] >>>> >>>> >>>>Hi Neil, >>>> >>>>I think John beat me off the blocks here, but... >>>> >>>> >>>> >>>> >>>>>>GMPLS assumes an IP control plane. >>>>> >>>>>An IP control-plane? There is actually no such >>>> >>>>animal. Just what >>>> >>>> >>>>>the heck does that REALLY mean in GMPLS say? >>>> >>>> >>>>Let me explain. >>>>Perhaps I should have said "IP-based control >> >>plane". >> >>>>I mean a control plane which: >>>>- uses IP as its network protocol >>>>- uses IP addresses to identify control plane >>>>resources >>>>- uses IP addresses to identify data plane >> >>resources >> >>>>within >>>> the control plane >>>>- uses protocols developed for use in the >> >>Internet. >> >>>> >>>>>I am not questioning IP as a cl-ps networking >>>> >>>>protocol *carrying* >>>> >>>> >>>>>a signalling protocol (RSVP-TE, or dare I mention >>>> >>>>PNNI, or any >>>> >>>> >>>>>signalling protocol yet to be invented) or a >>>> >>>>routing protocol >>>> >>>> >>>>>(OSPF or ISO or whatever) >>>> >>>> >>>>I am glad to hear it. >>>> >>>> >>>> >>>>>or even management protocols >>>> >>>> >>>>Fine, but not in the remit of CCAMP. >>>> >>>> >>>> >>>>>but an 'IP Control Plane' per se means absolutely >>>> >>>>nothing to me.... >>>> >>>>Well, I think it should. I think the list of >>>>attributes that I have >>>>given above define a control plane based on IP. >>>> >>>>It is undoubtable that attempts have been made to >>>>use control planes >>>>based on other protocols. Some have been highly >>>>successful. Some have >>>>been less fortunate. >>>> >>>> >>>> >>>>>...nor should it to anyone else. >>>> >>>> >>>>I think folks who were around at the beginning of >>>>CCAMP and who were >>>>part of the debate with the IESG will be very >>>>familiar with where the >>>>IETF draws the line here. >>>> >>>> >>>> >>>> >>>>>I think some folks might need a reality check >>>> >>>>here....and also >>>> >>>> >>>>>on the self-assumed importance of a >> >>control-plane. >> >>>>Hint: It ain't >>>> >>>> >>>>>that important.....the management-plane (which >> >>may >> >>>>be using IP!) >>>> >>>> >>>>>however is. >>>> >>>> >>>>I am not sure how to interpret this. >>>>It may be that you think that control plane is bad >>>>per se, but you have >>>>said elsewhere that you think it has value - but >>>>much less than the >>>>management plane. >>>>It may be that you believe that CCAMP is willfully >>>>neglecting the >>>>management plane. This would, in fact, be true. It >>>>is not in CCAMP's >> > === message truncated === > > __________________________________________________ > Do You Yahoo!? > Tired of spam? Yahoo! Mail has the best spam protection around > http://mail.yahoo.com > > > . > Envelope-to: ccamp-data@psg.com Delivery-date: Sun, 24 Jul 2005 15:07:35 +0000 DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=Message-ID:Received:Date:From:Subject:To:Cc:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding; b=5xnt/2kpwuIPap+8Z5dc7hsTz55H00sc7MHWI/zlh9AL8hZGuj2EhnwS8a5PN7rnzU6gWGnGLEKg04IgB/C0ZexS6bkQeKEsYFxrzRM5xP2HIIbI8BegDFXYZlDBQJ9SZ0poI6b1gt9azwRUcFm28B/mnnYZIMJM6P3cAcSIxPM= ; Message-ID: <20050724150533.54903.qmail@web30802.mail.mud.yahoo.com> Date: Sun, 24 Jul 2005 08:05:32 -0700 (PDT) From: Igor Bryskin <i_bryskin@yahoo.com> Subject: Re: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] To: dpapadimitriou@psg.com, dimitri.papadimitriou@alcatel.be Cc: neil.2.harrison@bt.com, adrian@olddog.co.uk, ccamp@ops.ietf.org MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Dimitri, What I was trying to say here is that for the User how TE Ethernet connection is established - via control or management planes is of a little importance. Before explaining how cool/elegant it could be provisioned via GMPLS you need to answer IMO the following questions: 1) What does it mean Traffic Engineered Ethernet connection in the data plane and what is it good for? For instance, we perfectly understand what does ATM connection or TDM connection mean in the data plane and can provide references to corresponding data plane standards. 2) How does the label-aware Ethernet forwarder co-exist with other Ethernet forwarders on the same switch 3) What information is necessary to provision the forwarding tables? And so on. Cheers, Igor --- dimitri papadimitriou <dpapadimitriou@psg.com> wrote: > hi igor, > > [snip - hint: igor i suggest you create a thread on > control vs > management plane] > > > My view wrt Ethertnet GMPLS is this. I have no > doubts > > that we can come up with mechanisms to dynamically > > provision L2SC LSPs. > > ouf ;-) > > > My problem is the order of events. > > It does not seem to me wise to come up with > > some control plane framework and solution(s) and > after > > that think what we need to do in the data plane. > > actually, if this is the only problem you have, i > should underline we do > not have to initiate the definition of a specific > forwarding paradigm so > i am not necessarily sure about what you mean by the > last part of the > above sentence > > > It seems to me wiser to learn how to statically > > provision such LSPs, see how useful they are, and > only > > after that design and develop ways to provision > them > > dynamically. > > i am not sure to follow how are you going to > determine the benefit from > dynamic provisioning by statically configure these > LSPs ? > > > In other words, it is wiser to repeat what has > been > > done with TDM LSPs. > > wiser ? ... would you clarify ? don't you think it > is more appropriate > to receive feedback on the appropriateness of the > use of GMPLS for the > scenarios depicted in the problem statement document > rather than digress > on non-rational argument such as "wiser", "careful" > or so... ? > > > Cheers, > > Igor > > > > > > > >>A control-plane is useful in the major co-cs > >>transport networks (mainly > >>for S-PVCs and resilience) but its a minor player > >>compared to role of > >>the management-plane. The converse of course > holds > >>in a *traffic* > >>carrying cl-ps IP network, esp when this is in > the > >>context of the > >>public Internet.....but of course this isn't the > >>case here and IP is > >>only being used as the transport protocol in the > OOB > >>data-plane network > >>that carries the control/management-plane > protocols. > >> > >> > >>These were points I wanted to make.....hopefully > >>I've done a better job > >>this time. > >> > >>regards, Neil > >> > >> > >> > >>-----Original Message----- > >>From: Adrian Farrel [mailto:adrian@olddog.co.uk] > >>Sent: 23 July 2005 00:08 > >>To: Harrison,N,Neil,IKM1 R > >>Cc: ccamp@ops.ietf.org > >>Subject: Re: Ethernet Control Plane [Was: Re: > >>Frameformat in a l2cs > >>gmpls rnvironment.] > >> > >> > >>Hi Neil, > >> > >>I think John beat me off the blocks here, but... > >> > >> > >> > >>>>GMPLS assumes an IP control plane. > >>> > >>>An IP control-plane? There is actually no such > >> > >>animal. Just what > >> > >>>the heck does that REALLY mean in GMPLS say? > >> > >> > >>Let me explain. > >>Perhaps I should have said "IP-based control > plane". > >>I mean a control plane which: > >>- uses IP as its network protocol > >>- uses IP addresses to identify control plane > >>resources > >>- uses IP addresses to identify data plane > resources > >>within > >> the control plane > >>- uses protocols developed for use in the > Internet. > >> > >> > >>>I am not questioning IP as a cl-ps networking > >> > >>protocol *carrying* > >> > >>>a signalling protocol (RSVP-TE, or dare I mention > >> > >>PNNI, or any > >> > >>>signalling protocol yet to be invented) or a > >> > >>routing protocol > >> > >>>(OSPF or ISO or whatever) > >> > >> > >>I am glad to hear it. > >> > >> > >>>or even management protocols > >> > >> > >>Fine, but not in the remit of CCAMP. > >> > >> > >>>but an 'IP Control Plane' per se means absolutely > >> > >>nothing to me.... > >> > >>Well, I think it should. I think the list of > >>attributes that I have > >>given above define a control plane based on IP. > >> > >>It is undoubtable that attempts have been made to > >>use control planes > >>based on other protocols. Some have been highly > >>successful. Some have > >>been less fortunate. > >> > >> > >>>...nor should it to anyone else. > >> > >> > >>I think folks who were around at the beginning of > >>CCAMP and who were > >>part of the debate with the IESG will be very > >>familiar with where the > >>IETF draws the line here. > >> > >> > >> > >>>I think some folks might need a reality check > >> > >>here....and also > >> > >>>on the self-assumed importance of a > control-plane. > >> > >>Hint: It ain't > >> > >>>that important.....the management-plane (which > may > >> > >>be using IP!) > >> > >>>however is. > >> > >> > >>I am not sure how to interpret this. > >>It may be that you think that control plane is bad > >>per se, but you have > >>said elsewhere that you think it has value - but > >>much less than the > >>management plane. > >>It may be that you believe that CCAMP is willfully > >>neglecting the > >>management plane. This would, in fact, be true. It > >>is not in CCAMP's > === message truncated === __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com Envelope-to: ccamp-data@psg.com Delivery-date: Sun, 24 Jul 2005 03:36:30 +0000 Message-ID: <42E30C19.1010300@cisco.com> Date: Sat, 23 Jul 2005 20:33:45 -0700 From: Arun Satyanarayana <asatyana@cisco.com> Organization: Cisco Systems User-Agent: Mozilla Thunderbird 1.0.2 (Windows/20050317) MIME-Version: 1.0 To: ccamp@ops.ietf.org, Adrian Farrel <adrian@olddog.co.uk>, Kireeti Kompella <kireeti@juniper.net> CC: Lou Berger <lberger@movaz.com>, Dimitri Papadimitriou <dimitri.papadimitriou@alcatel.be>, Reshad Rahman <rrahman@cisco.com>, Anca Zamfir <ancaz@cisco.com>, Junaid Israr <jisrar@cisco.com>, "Arun Satyanarayana (asatyana)" <asatyana@cisco.com> Subject: Re: I-D ACTION:draft-ietf-ccamp-rsvp-restart-ext-03.txt Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Hi all, Version -03 is the updated version resulting from last-call comments. A scalability concern was raised, where, the restarting node does not support the mechanism in this I-D, but, one or more RSVP neighbors do. This would result in unnecessary generation (and possible retransmission) of RecoveryPath msgs by the neighbors, and, receive processing of the RecoveryPath msg until the restarting node has to decide to drop it. This may be a scalability issue when a large number of LSPs are active on the restarting node. The solution is to indicate the capability to transmit and receive RecoveryPath msgs with 2 new bits in the Capability object. Note: The above solution also addresses a similar concern raised earlier, where, the restarting node support the extensions but one or more of its neighbors do not. The restarting node would have to wait until the end of the Recovery Period to revert to restart processing per RFC3473. With an explicit advertisement of RecoveryPath capability, this requirement no longer exists. We (the co-authors) hope that a short last-call to review the above change only is sufficient to progress the I-D. Please review and comment. Thanks, Arun ============================================================== Internet-Drafts@ietf.org wrote: > A New Internet-Draft is available from the on-line Internet-Drafts directories. > This draft is a work item of the Common Control and Measurement Plane Working Group of the IETF. > > Title : Extensions to GMPLS RSVP Graceful Restart > Author(s) : A. Satyanarayana > Filename : draft-ietf-ccamp-rsvp-restart-ext-03.txt > Pages : 24 > Date : 2005-7-19 > > This document describes extensions to the RSVP Graceful Restart > mechanisms defined in RFC 3473. The extensions enable the recovery > of RSVP signaling state based on the Path message last sent by the > node being restarted. Previously defined Graceful Restart > mechanisms, also called recovery from nodal faults, permit recovery > of signaling state from adjacent nodes when the data plane has > retained the associated forwarding state across a restart. These > mechanisms do not fully support signaling state recovery on ingress > nodes or recovery of all RSVP objects. The presented extensions use > the RSVP Hello extensions defined in RFC 3209, and extensions for > state recovery on nodal faults defined in RFC 3473. With the > presented extensions the restarting node can recover all previously > transmitted Path state including the Explicit Route Object and the > downstream (outgoing) interface identifiers. The extensions can also > be used to recover signaling state after the restart of an ingress > node. The extensions optionally support the use of Summary Refresh, > defined in RFC 2961, to reduce the number of messages exchanged > during the Recovery Phase when the restarting node has recovered > signaling state locally for one or more LSPs. > > A URL for this Internet-Draft is: > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-rsvp-restart-ext-03.txt > > To remove yourself from the I-D Announcement list, send a message to > i-d-announce-request@ietf.org with the word unsubscribe in the body of the message. > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > to change your subscription settings. > > > Internet-Drafts are also available by anonymous FTP. Login with the username > "anonymous" and a password of your e-mail address. After logging in, > type "cd internet-drafts" and then > "get draft-ietf-ccamp-rsvp-restart-ext-03.txt". > > A list of Internet-Drafts directories can be found in > http://www.ietf.org/shadow.html > or ftp://ftp.ietf.org/ietf/1shadow-sites.txt > > > Internet-Drafts can also be obtained by e-mail. > > Send a message to: > mailserv@ietf.org. > In the body type: > "FILE /internet-drafts/draft-ietf-ccamp-rsvp-restart-ext-03.txt". > > NOTE: The mail server at ietf.org can return the document in > MIME-encoded form by using the "mpack" utility. To use this > feature, insert the command "ENCODING mime" before the "FILE" > command. To decode the response(s), you will need "munpack" or > a MIME-compliant mail reader. Different MIME-compliant mail readers > exhibit different behavior, especially when dealing with > "multipart" MIME messages (i.e. documents which have been split > up into multiple messages), so check your local documentation on > how to manipulate these messages. > > > Below is the data which will enable a MIME compliant mail reader > implementation to automatically retrieve the ASCII version of the > Internet-Draft. Envelope-to: ccamp-data@psg.com Delivery-date: Sat, 23 Jul 2005 17:22:33 +0000 Message-ID: <014a01c58fa2$82e2c220$0601a8c0@pc6> Reply-To: "Tom Petch" <nwnetworks@dial.pipex.com> From: "Tom Petch" <nwnetworks@dial.pipex.com> To: <neil.2.harrison@bt.com> Cc: <ccamp@ops.ietf.org> Subject: Re: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Date: Sat, 23 Jul 2005 17:58:27 +0200 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Neil Is there an ITU-T Recommendation that gives an architectural definition of what it means by control plane, data plane, management plane, user plane, traffic dataplane, control data plane etc etc? I know what I mean when I use the terms but am not sure I always know what others mean. Tom Petch ----- Original Message ----- From: <neil.2.harrison@bt.com> To: <John.E.Drake2@boeing.com>; <adrian@olddog.co.uk>; <richard.spencer@bt.com>; <juergen.heiles@siemens.com>; <loa@pi.se>; <per@defero.se> Cc: <ccamp@ops.ietf.org> Sent: Saturday, July 23, 2005 2:15 PM Subject: RE: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] > > > IP is cl-ps networking techology period....and its > > > jolly important, but PLEASE don't try and feed me any of this 'IP > > > control plane' nonsense. > > JD: Over the years I have enjoyed Neil's rants immensely, > but I have to say that this one is a classic. NH=> Thanks John for the kind words....its good to know you've had at least something to smile about ;-) Seriously though, see my updated response to Adrian. But if you are still struggling to grasp what I mean by the above (out of context quote) try this: A single cl-ps IP network carrying traffic and everything else (eg public Internet) is very different animal to a co-cs network system where the traffic data-plane network and control/management data-planes network are logically disjoint, ie the role of IP in the latter is importantly different to the role of IP in the former...and let's hope never the 2 shall meet. regards, neil Envelope-to: ccamp-data@psg.com Delivery-date: Sat, 23 Jul 2005 16:44:59 +0000 Message-ID: <42E273A6.1000003@psg.com> Date: Sat, 23 Jul 2005 18:43:18 +0200 From: dimitri papadimitriou <dpapadimitriou@psg.com> Reply-To: dpapadimitriou@psg.com, dimitri.papadimitriou@alcatel.be User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.8) Gecko/20050511 MIME-Version: 1.0 To: Igor Bryskin <i_bryskin@yahoo.com> CC: neil.2.harrison@bt.com, adrian@olddog.co.uk, ccamp@ops.ietf.org Subject: Re: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit hi igor, [snip - hint: igor i suggest you create a thread on control vs management plane] > My view wrt Ethertnet GMPLS is this. I have no doubts > that we can come up with mechanisms to dynamically > provision L2SC LSPs. ouf ;-) > My problem is the order of events. > It does not seem to me wise to come up with > some control plane framework and solution(s) and after > that think what we need to do in the data plane. actually, if this is the only problem you have, i should underline we do not have to initiate the definition of a specific forwarding paradigm so i am not necessarily sure about what you mean by the last part of the above sentence > It seems to me wiser to learn how to statically > provision such LSPs, see how useful they are, and only > after that design and develop ways to provision them > dynamically. i am not sure to follow how are you going to determine the benefit from dynamic provisioning by statically configure these LSPs ? > In other words, it is wiser to repeat what has been > done with TDM LSPs. wiser ? ... would you clarify ? don't you think it is more appropriate to receive feedback on the appropriateness of the use of GMPLS for the scenarios depicted in the problem statement document rather than digress on non-rational argument such as "wiser", "careful" or so... ? > Cheers, > Igor > > > >>A control-plane is useful in the major co-cs >>transport networks (mainly >>for S-PVCs and resilience) but its a minor player >>compared to role of >>the management-plane. The converse of course holds >>in a *traffic* >>carrying cl-ps IP network, esp when this is in the >>context of the >>public Internet.....but of course this isn't the >>case here and IP is >>only being used as the transport protocol in the OOB >>data-plane network >>that carries the control/management-plane protocols. >> >> >>These were points I wanted to make.....hopefully >>I've done a better job >>this time. >> >>regards, Neil >> >> >> >>-----Original Message----- >>From: Adrian Farrel [mailto:adrian@olddog.co.uk] >>Sent: 23 July 2005 00:08 >>To: Harrison,N,Neil,IKM1 R >>Cc: ccamp@ops.ietf.org >>Subject: Re: Ethernet Control Plane [Was: Re: >>Frameformat in a l2cs >>gmpls rnvironment.] >> >> >>Hi Neil, >> >>I think John beat me off the blocks here, but... >> >> >> >>>>GMPLS assumes an IP control plane. >>> >>>An IP control-plane? There is actually no such >> >>animal. Just what >> >>>the heck does that REALLY mean in GMPLS say? >> >> >>Let me explain. >>Perhaps I should have said "IP-based control plane". >>I mean a control plane which: >>- uses IP as its network protocol >>- uses IP addresses to identify control plane >>resources >>- uses IP addresses to identify data plane resources >>within >> the control plane >>- uses protocols developed for use in the Internet. >> >> >>>I am not questioning IP as a cl-ps networking >> >>protocol *carrying* >> >>>a signalling protocol (RSVP-TE, or dare I mention >> >>PNNI, or any >> >>>signalling protocol yet to be invented) or a >> >>routing protocol >> >>>(OSPF or ISO or whatever) >> >> >>I am glad to hear it. >> >> >>>or even management protocols >> >> >>Fine, but not in the remit of CCAMP. >> >> >>>but an 'IP Control Plane' per se means absolutely >> >>nothing to me.... >> >>Well, I think it should. I think the list of >>attributes that I have >>given above define a control plane based on IP. >> >>It is undoubtable that attempts have been made to >>use control planes >>based on other protocols. Some have been highly >>successful. Some have >>been less fortunate. >> >> >>>...nor should it to anyone else. >> >> >>I think folks who were around at the beginning of >>CCAMP and who were >>part of the debate with the IESG will be very >>familiar with where the >>IETF draws the line here. >> >> >> >>>I think some folks might need a reality check >> >>here....and also >> >>>on the self-assumed importance of a control-plane. >> >>Hint: It ain't >> >>>that important.....the management-plane (which may >> >>be using IP!) >> >>>however is. >> >> >>I am not sure how to interpret this. >>It may be that you think that control plane is bad >>per se, but you have >>said elsewhere that you think it has value - but >>much less than the >>management plane. >>It may be that you believe that CCAMP is willfully >>neglecting the >>management plane. This would, in fact, be true. It >>is not in CCAMP's >>remit to look at the management plane. Other SDOs >>are working to >>establish common standards for management of network >>devices across >>multiple vendors - we can wish them luck. >>Clearly some people (vendors and providers) see >>sufficient value in a >>control plane to invest time and energy. >> >> >>>The (hype) party is over for the OTN start-ups. >> >>IP per se does NOT >> >>>define a *control-plane*...IP is cl-ps networking >> >>techology period >> >>>....and its jolly important, but PLEASE don't try >> >>and feed me any of >> >>>this 'IP control plane' nonsense. >> > === message truncated === > > > > > ____________________________________________________ > Start your day with Yahoo! - make it your home page > http://www.yahoo.com/r/hs > > > > . > Envelope-to: ccamp-data@psg.com Delivery-date: Sat, 23 Jul 2005 13:50:45 +0000 DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=Message-ID:Received:Date:From:Subject:To:Cc:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding; b=1eKRLLtEVAw96DvLBpgepSNsC+ukFMe1zsbMKf9T3+X8BEBOGGhCwNSFwkDrcSScPO+CAw1YhnNJWeai/cJKfH5TNl7l1QYtJ43hTk3jMZU6JkQ4ub/8vRpR2EO7QQKUmOp/olKFRT7xgyDSq5sr8PAi7aEcCJrJUZY62KiunCc= ; Message-ID: <20050723134931.3252.qmail@web30805.mail.mud.yahoo.com> Date: Sat, 23 Jul 2005 06:49:31 -0700 (PDT) From: Igor Bryskin <i_bryskin@yahoo.com> Subject: RE: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] To: neil.2.harrison@bt.com, adrian@olddog.co.uk Cc: ccamp@ops.ietf.org MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Neil, [SNIP] > I am also a little concerned that some folks might > naively assume that a > control-plane is going to solve all their > management-plane > issues......when we move down the network stack to > the duct there are a > whole raft of factors that mean the management-plane > still dominates > here. If you talk to the right folks in the > operators who run these > types of network you will indeed understand what I > mean by this. This kind of statements you've been making in about every email. I am curious to know as we move towards the duct which *exactly* features, provisioning functions or operations that are implemented via NMS could not be conceptually realized via control plane? Or you are not the "right folk to ask" ? :=) My view wrt Ethertnet GMPLS is this. I have no doubts that we can come up with mechanisms to dynamically provision L2SC LSPs. My problem is the order of events. It does not seem to me wise to come up with some control plane framework and solution(s) and after that think what we need to do in the data plane. It seems to me wiser to learn how to statically provision such LSPs, see how useful they are, and only after that design and develop ways to provision them dynamically. In other words, it is wiser to repeat what has been done with TDM LSPs. Cheers, Igor > A control-plane is useful in the major co-cs > transport networks (mainly > for S-PVCs and resilience) but its a minor player > compared to role of > the management-plane. The converse of course holds > in a *traffic* > carrying cl-ps IP network, esp when this is in the > context of the > public Internet.....but of course this isn't the > case here and IP is > only being used as the transport protocol in the OOB > data-plane network > that carries the control/management-plane protocols. > > > These were points I wanted to make.....hopefully > I've done a better job > this time. > > regards, Neil > > > > -----Original Message----- > From: Adrian Farrel [mailto:adrian@olddog.co.uk] > Sent: 23 July 2005 00:08 > To: Harrison,N,Neil,IKM1 R > Cc: ccamp@ops.ietf.org > Subject: Re: Ethernet Control Plane [Was: Re: > Frameformat in a l2cs > gmpls rnvironment.] > > > Hi Neil, > > I think John beat me off the blocks here, but... > > > > > GMPLS assumes an IP control plane. > > > > An IP control-plane? There is actually no such > animal. Just what > > the heck does that REALLY mean in GMPLS say? > > Let me explain. > Perhaps I should have said "IP-based control plane". > I mean a control plane which: > - uses IP as its network protocol > - uses IP addresses to identify control plane > resources > - uses IP addresses to identify data plane resources > within > the control plane > - uses protocols developed for use in the Internet. > > > I am not questioning IP as a cl-ps networking > protocol *carrying* > > a signalling protocol (RSVP-TE, or dare I mention > PNNI, or any > > signalling protocol yet to be invented) or a > routing protocol > > (OSPF or ISO or whatever) > > I am glad to hear it. > > > or even management protocols > > Fine, but not in the remit of CCAMP. > > > but an 'IP Control Plane' per se means absolutely > nothing to me.... > > Well, I think it should. I think the list of > attributes that I have > given above define a control plane based on IP. > > It is undoubtable that attempts have been made to > use control planes > based on other protocols. Some have been highly > successful. Some have > been less fortunate. > > > ...nor should it to anyone else. > > I think folks who were around at the beginning of > CCAMP and who were > part of the debate with the IESG will be very > familiar with where the > IETF draws the line here. > > > > I think some folks might need a reality check > here....and also > > on the self-assumed importance of a control-plane. > Hint: It ain't > > that important.....the management-plane (which may > be using IP!) > > however is. > > I am not sure how to interpret this. > It may be that you think that control plane is bad > per se, but you have > said elsewhere that you think it has value - but > much less than the > management plane. > It may be that you believe that CCAMP is willfully > neglecting the > management plane. This would, in fact, be true. It > is not in CCAMP's > remit to look at the management plane. Other SDOs > are working to > establish common standards for management of network > devices across > multiple vendors - we can wish them luck. > Clearly some people (vendors and providers) see > sufficient value in a > control plane to invest time and energy. > > > The (hype) party is over for the OTN start-ups. > IP per se does NOT > > define a *control-plane*...IP is cl-ps networking > techology period > > ....and its jolly important, but PLEASE don't try > and feed me any of > > this 'IP control plane' nonsense. > === message truncated === ____________________________________________________ Start your day with Yahoo! - make it your home page http://www.yahoo.com/r/hs Envelope-to: ccamp-data@psg.com Delivery-date: Sat, 23 Jul 2005 12:16:58 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01C58F80.310DD9EA" Subject: RE: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Date: Sat, 23 Jul 2005 13:15:19 +0100 Message-ID: <0536FC9B908BEC4597EE721BE6A353890A9F1E31@i2km07-ukbr.domain1.systemhost.net> Thread-Topic: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Thread-Index: AcWPEfcioWKblWAORniGMXaBuv1+cgAX+5xA From: <neil.2.harrison@bt.com> To: <adrian@olddog.co.uk> Cc: <ccamp@ops.ietf.org> This is a multi-part message in MIME format. ------_=_NextPart_001_01C58F80.310DD9EA Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Ok thanks Adrian, =20 Re-reading what I wrote and your response (and John's) I can see I did an awful job of trying to explain what I really meant...and I'm not surprised you responded as you did. So my apologies for letting frustration get the better of me. Trying to express it better: =20 Its simply not true that one can have a 'common' control-plane approach from IP to Optics.......for many reasons such as: the modal differences demand different approaches, who owns what layer network, the notion that one can create topology on the fly, the commercial factors that govern GoS network build-out, the relative importance of the control-plane vs the management plane (which also varies with network mode and where you are in the stack), etc. =20 However, if you go back some time the impression that was originally 'sold' to folks was that there IS this common control-plane from IP to Optics. This view may have mellowed now, but I still see evidence of it of every now and then....and its when I see things that IMO are being somewhat 'economical with the truth' that I tend to react. =20 I also don't like it when folks go break architectural rules and 'pretend its OK/good'. This is not generally an issue in CCAMPs as its jolly hard to do much damage to the co-cs mode....but you can for sure create havoc in the co-ps mode. And at the end of the day arch violations and functional omissions/deficiencies end up as operational problems and opex costs in our networks. If more folks understood functional architecture stuff I suspect this would help control this particular issue better. =20 We therefore have the real situation that there is no forced technical reason to be prescriptive in the choice of control-plane functional components when we are dealing with our major SDH/OTN co-cs transport networks.....which, of course, also have nothing to do with the public Internet That is, they can be chosen on a best-of-breed basis. =20 =20 I am also a little concerned that some folks might naively assume that a control-plane is going to solve all their management-plane issues......when we move down the network stack to the duct there are a whole raft of factors that mean the management-plane still dominates here. If you talk to the right folks in the operators who run these types of network you will indeed understand what I mean by this. =20 A control-plane is useful in the major co-cs transport networks (mainly for S-PVCs and resilience) but its a minor player compared to role of the management-plane. The converse of course holds in a *traffic* carrying cl-ps IP network, esp when this is in the context of the public Internet.....but of course this isn't the case here and IP is only being used as the transport protocol in the OOB data-plane network that carries the control/management-plane protocols. =20 =20 These were points I wanted to make.....hopefully I've done a better job this time. =20 regards, Neil =20 =20 -----Original Message----- From: Adrian Farrel [mailto:adrian@olddog.co.uk]=20 Sent: 23 July 2005 00:08 To: Harrison,N,Neil,IKM1 R Cc: ccamp@ops.ietf.org Subject: Re: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Hi Neil, =20 I think John beat me off the blocks here, but... > > GMPLS assumes an IP control plane. > > An IP control-plane? There is actually no such animal. Just what=20 > the heck does that REALLY mean in GMPLS say? =20 Let me explain. Perhaps I should have said "IP-based control plane". I mean a control plane which: - uses IP as its network protocol - uses IP addresses to identify control plane resources - uses IP addresses to identify data plane resources within=20 the control plane - uses protocols developed for use in the Internet. =20 > I am not questioning IP as a cl-ps networking protocol *carrying*=20 > a signalling protocol (RSVP-TE, or dare I mention PNNI, or any=20 > signalling protocol yet to be invented) or a routing protocol=20 > (OSPF or ISO or whatever) =20 I am glad to hear it. =20 > or even management protocols =20 Fine, but not in the remit of CCAMP. =20 > but an 'IP Control Plane' per se means absolutely nothing to me.... =20 Well, I think it should. I think the list of attributes that I have given above define a control plane based on IP.=20 =20 It is undoubtable that attempts have been made to use control planes based on other protocols. Some have been highly successful. Some have been less fortunate. =20 > ...nor should it to anyone else.=20 =20 I think folks who were around at the beginning of CCAMP and who were part of the debate with the IESG will be very familiar with where the IETF draws the line here. =20 =20 > I think some folks might need a reality check here....and also > on the self-assumed importance of a control-plane. Hint: It ain't=20 > that important.....the management-plane (which may be using IP!)=20 > however is. =20 I am not sure how to interpret this. It may be that you think that control plane is bad per se, but you have said elsewhere that you think it has value - but much less than the management plane. It may be that you believe that CCAMP is willfully neglecting the management plane. This would, in fact, be true. It is not in CCAMP's remit to look at the management plane. Other SDOs are working to establish common standards for management of network devices across multiple vendors - we can wish them luck. Clearly some people (vendors and providers) see sufficient value in a control plane to invest time and energy. > The (hype) party is over for the OTN start-ups. IP per se does NOT > define a *control-plane*...IP is cl-ps networking techology period > ....and its jolly important, but PLEASE don't try and feed me any of > this 'IP control plane' nonsense. =20 Tush. =20 > Adrian....you are far smarter than this IMO and should know better. =20 Thanks for the endorsement. May I quote you? =20 Adrian ------_=_NextPart_001_01C58F80.310DD9EA Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; = charset=3Dus-ascii"> <TITLE>Message</TITLE> <META content=3D"MSHTML 6.00.2800.1106" name=3DGENERATOR> <STYLE></STYLE> </HEAD> <BODY> <DIV><SPAN class=3D924593210-23072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2>Ok thanks Adrian,</FONT></SPAN></DIV> <DIV><SPAN class=3D924593210-23072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2></FONT></SPAN> </DIV> <DIV><SPAN class=3D924593210-23072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2>Re-reading what I wrote and your response (and John's) I can = see I did an=20 awful job of trying to explain what I really meant...and I'm not = surprised you=20 responded as you did. So my apologies for letting frustration get = the=20 better of me. Trying to express it better:</FONT></SPAN></DIV> <DIV><SPAN class=3D924593210-23072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2></FONT></SPAN> </DIV> <DIV><SPAN class=3D924593210-23072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2>Its simply not true that one can have a 'common' control-plane = approach=20 from IP to Optics.......for many reasons such as: the modal differences = demand=20 different approaches, who owns what layer network, the notion that one = can=20 create topology on the fly, the commercial factors that govern GoS = network=20 build-out, the relative importance of the control-plane vs the = management plane=20 (which also varies with network mode and where you are in the stack),=20 etc.</FONT></SPAN></DIV> <DIV><SPAN class=3D924593210-23072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2></FONT></SPAN> </DIV> <DIV><SPAN class=3D924593210-23072005><SPAN = class=3D924593210-23072005><FONT=20 face=3D"Comic Sans MS" color=3D#800000 size=3D2><SPAN = class=3D924593210-23072005><FONT=20 face=3D"Comic Sans MS" color=3D#800000 size=3D2>However, if you go back = some time the=20 impression that was originally 'sold' to folks was that there IS this = common=20 control-plane from IP to Optics. This view may have mellowed now, = but I=20 still see evidence of it of every now and then....and its when I see = things that=20 IMO are being somewhat 'economical with the truth' that I tend to=20 react.</FONT></SPAN></FONT></SPAN></SPAN></DIV> <DIV><SPAN class=3D924593210-23072005><SPAN = class=3D924593210-23072005><FONT=20 face=3D"Comic Sans MS" color=3D#800000 size=3D2><SPAN=20 class=3D924593210-23072005></SPAN></FONT></SPAN></SPAN> </DIV> <DIV><SPAN class=3D924593210-23072005><SPAN = class=3D924593210-23072005><FONT=20 face=3D"Comic Sans MS" color=3D#800000 size=3D2><SPAN = class=3D924593210-23072005>I also=20 don't like it when folks go break architectural rules and 'pretend its=20 OK/good'. This is not generally an issue in CCAMPs as its jolly = hard to do=20 much damage to the co-cs mode....but you can for sure create havoc in = the co-ps=20 mode. And at the end of the day arch violations and = functional=20 omissions/deficiencies end up as operational problems and opex costs in = our=20 networks. If more folks understood functional architecture stuff I = suspect=20 this would help control this particular issue=20 better.</SPAN></FONT></SPAN></SPAN></DIV> <DIV><SPAN class=3D924593210-23072005><SPAN = class=3D924593210-23072005><FONT=20 face=3D"Comic Sans MS" color=3D#800000 size=3D2><SPAN=20 class=3D924593210-23072005></SPAN></FONT></SPAN></SPAN> </DIV> <DIV><SPAN class=3D924593210-23072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2>We therefore have the real situation that there is no forced = technical=20 reason to be prescriptive in the choice of control-plane functional = components=20 when we are dealing with our major SDH/OTN co-cs transport = networks.....which,=20 of course, also have nothing to do with the public=20 Internet </FONT></SPAN><SPAN class=3D924593210-23072005><FONT = face=3D"Comic Sans MS" color=3D#800000 size=3D2><SPAN = class=3D924593210-23072005><FONT=20 face=3D"Comic Sans MS" color=3D#800000 size=3D2> </FONT></SPAN>That = is, they can=20 be chosen on a best-of-breed basis. </FONT></SPAN></DIV> <DIV><SPAN class=3D924593210-23072005></SPAN> </DIV> <DIV><SPAN class=3D924593210-23072005></SPAN><SPAN = class=3D924593210-23072005><FONT=20 face=3D"Comic Sans MS" color=3D#800000 size=3D2>I am also a little = concerned that some=20 folks might naively assume that a control-plane is going to solve all = their=20 management-plane issues......when we move down the network stack to the = duct=20 there are a whole raft of factors that mean the management-plane still = dominates=20 here. If you talk to the right folks in the operators who run = these types=20 of network you will indeed understand what I mean by = this.</FONT></SPAN></DIV> <DIV><SPAN class=3D924593210-23072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2></FONT></SPAN> </DIV> <DIV><SPAN class=3D924593210-23072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2>A control-plane is useful in the major co-cs transport networks = (mainly=20 for S-PVCs and resilience) but its a minor player compared to role of = the=20 management-plane. The converse of course holds in a = *traffic*=20 carrying cl-ps IP network, esp when this is in the context of = the=20 public Internet.....but of course this isn't the case here and IP is = only being=20 used as the transport protocol in the OOB data-plane network that = carries the=20 control/management-plane protocols. </FONT></SPAN></DIV> <DIV><SPAN class=3D924593210-23072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2></FONT></SPAN><SPAN class=3D924593210-23072005><FONT = face=3D"Comic Sans MS"=20 color=3D#800000 size=3D2></FONT></SPAN> </DIV> <DIV><SPAN class=3D924593210-23072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2>These were points I wanted to make.....hopefully I've done a = better job=20 this time.</FONT></SPAN></DIV> <DIV><SPAN class=3D924593210-23072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2></FONT></SPAN> </DIV> <DIV><SPAN class=3D924593210-23072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2>regards, Neil</FONT></SPAN></DIV> <DIV><SPAN class=3D924593210-23072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2></FONT></SPAN> </DIV> <DIV><SPAN class=3D924593210-23072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2> </FONT></SPAN></DIV> <BLOCKQUOTE dir=3Dltr=20 style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #800000 2px = solid; MARGIN-RIGHT: 0px"> <DIV></DIV> <DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr = align=3Dleft><FONT=20 face=3DTahoma size=3D2>-----Original Message-----<BR><B>From:</B> = Adrian Farrel=20 [mailto:adrian@olddog.co.uk] <BR><B>Sent:</B> 23 July 2005 = 00:08<BR><B>To:</B>=20 Harrison,N,Neil,IKM1 R<BR><B>Cc:</B> = ccamp@ops.ietf.org<BR><B>Subject:</B> Re:=20 Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls=20 rnvironment.]<BR><BR></FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Hi Neil,</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>I think John beat me off the blocks = here,=20 but...</FONT></DIV> <DIV><FONT face=3DCourier><BR><FONT size=3D2></FONT></FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> > GMPLS assumes an IP = control=20 plane.<BR>><BR>> An IP control-plane? There is actually no = such=20 animal. Just what </FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> the </FONT><FONT = face=3DCourier=20 size=3D2>heck does that REALLY mean in GMPLS say?</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Let me explain.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Perhaps I should have said = "IP-based control=20 plane".</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>I mean a control plane = which:</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>- uses IP as its network = protocol</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>- uses IP addresses to identify = control plane=20 resources</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>- uses IP addresses to identify = data plane=20 resources within </FONT></DIV> <DIV><FONT face=3DCourier size=3D2> the control = plane</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>- uses protocols developed for use = in the=20 Internet.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>> I am not = questioning </FONT><FONT=20 face=3DCourier size=3D2>IP</FONT><FONT face=3DCourier = size=3D2> as a cl-ps=20 networking protocol *carrying* </FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> a signalling = protocol </FONT><FONT=20 face=3DCourier size=3D2>(</FONT><FONT face=3DCourier size=3D2>RSVP-TE, = or dare I=20 mention PNNI, or any </FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> signalling protocol = yet </FONT><FONT=20 face=3DCourier><FONT size=3D2><FONT = size=3D+0>to</FONT></FONT></FONT><FONT=20 face=3DCourier size=3D2> be invented) or a routing protocol = </FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> (OSPF or ISO or = whatever)</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>I am glad to hear it.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>> or </FONT><FONT = face=3DCourier=20 size=3D2>even management protocols</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Fine, but not in the remit of=20 CCAMP.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>> but an 'IP Control Plane' = per se=20 means </FONT><FONT face=3DCourier size=3D2>absolutely nothing to=20 me....</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Well, I think it should. I think = the list of=20 attributes that I have given above define a control plane based on IP. = </FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>It is undoubtable that attempts = have been made=20 to use control planes based on other protocols. Some have been highly=20 successful. Some have been less fortunate.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>> ...nor should it to anyone=20 else. </FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>I think folks who were around at = the beginning=20 of CCAMP and who were part of the debate with the IESG will be very = familiar=20 with where the IETF draws the line here.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>> I think </FONT><FONT = face=3DCourier=20 size=3D2>some folks might need a reality check here....and = also</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> on the self-</FONT><FONT = face=3DCourier=20 size=3D2>assumed importance of a control-plane. H</FONT><FONT = face=3DCourier=20 size=3D2>int: It ain't </FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> that </FONT><FONT = face=3DCourier=20 size=3D2>important.....the management-plane (which may be using IP!)=20 </FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> however</FONT><FONT = face=3DCourier=20 size=3D2> is.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>I am not sure how to interpret=20 this.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>It may be that you think that = control plane is=20 bad per se, but you have said elsewhere that you think it has value - = but much=20 less than the management plane.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>It may be that you believe that = CCAMP is=20 willfully neglecting the management plane. This would, in fact, be = true. It is=20 not in CCAMP's remit to look at the management plane. Other SDOs are = working=20 to establish common standards for management of network devices across = multiple vendors - we can wish them luck.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Clearly some people (vendors and = providers) see=20 sufficient value in a control plane to invest time and energy.</DIV> <DIV><BR>> The (hype) party is over for the OTN start-ups. IP = per se=20 does NOT</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> define a = *control-plane*...IP is=20 cl-ps networking techology period</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> ....and its jolly important, = but PLEASE=20 don't try and feed me any of</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> this 'IP control plane'=20 nonsense.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Tush.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> </DIV> <DIV>> Adrian....you are far smarter than this IMO and should know=20 better.</DIV> <DIV> </DIV> <DIV>Thanks for the endorsement. May I quote you?</DIV> <DIV> </DIV> <DIV>Adrian<BR></DIV></BLOCKQUOTE></FONT><FONT face=3DCourier=20 size=3D2></FONT></BODY></HTML> ------_=_NextPart_001_01C58F80.310DD9EA-- Envelope-to: ccamp-data@psg.com Delivery-date: Sat, 23 Jul 2005 12:16:51 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: RE: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Date: Sat, 23 Jul 2005 13:15:18 +0100 Message-ID: <0536FC9B908BEC4597EE721BE6A353890A9F1E2F@i2km07-ukbr.domain1.systemhost.net> Thread-Topic: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Thread-Index: AcWO6PXwNV9J9GsyTyywah8dc4SA0QAESRbgAAQtiVAAB5V/cAAVQatg From: <neil.2.harrison@bt.com> To: <John.E.Drake2@boeing.com>, <adrian@olddog.co.uk>, <richard.spencer@bt.com>, <juergen.heiles@siemens.com>, <loa@pi.se>, <per@defero.se> Cc: <ccamp@ops.ietf.org> > > > IP is cl-ps networking techology period....and its > > > jolly important, but PLEASE don't try and feed me any of this 'IP > > > control plane' nonsense. >=20 > JD: Over the years I have enjoyed Neil's rants immensely, > but I have to say that this one is a classic. NH=3D> Thanks John for the kind words....its good to know you've had at least something to smile about ;-) Seriously though, see my updated response to Adrian. But if you are still struggling to grasp what I mean by the above (out of context quote) try this: A single cl-ps IP network carrying traffic and everything else (eg public Internet) is very different animal to a co-cs network system where the traffic data-plane network and control/management data-planes network are logically disjoint, ie the role of IP in the latter is importantly different to the role of IP in the former...and let's hope never the 2 shall meet. regards, neil =20 Envelope-to: ccamp-data@psg.com Delivery-date: Sat, 23 Jul 2005 10:34:23 +0000 To: <neil.2.harrison@bt.com> Cc: <Dimitri.Papadimitriou@alcatel.be>, <adrian@olddog.co.uk>, <ccamp@ops.ietf.org> MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Sat, 23 Jul 2005 12:32:19 +0200 Message-ID: <OFA4BFD7BF.452A350B-ONC1257047.0039E39E-C1257047.0039E436@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPm5laWwgLSB0byBiZSBwZXJmZWN0bHkg Y2xlYXIgLSA8L0ZPTlQ+PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5kdXJp bmcgbGFzdCByb3VuZCBvZiBkaXNjdXNzaW9uIHNvbWUgaW5kaXZpZHVhbHMgc3RhcnRlZCBzcGVh a2luZyB1c2luZyBhdXRob3JpdGF0aXZlIGFuZCBpbXBlcnNvbmFsIGFyZ3VtZW50cyBzdWNoIGFz ICZxdW90O0V0aGVybmV0IGJlbG9uZ3MgdG8gYW5vdGhlciBTRE8gaS5lLiB0aGUgSUVFRSZxdW90 OzwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPnRoZSBEVCBhbmQg bW9zdCBwZW9wbGUgZGlzY3Vzc2luZyB0aGlzIHRocmVhZCBhcmUgcGVyZmVjdGx5IGF3YXJlIG9m IHRoaXMgKHNlZSBzZWN0aW9uIDcuMSkgYW5kIGFsbCB3aGF0IGl0IGltcGxpZXM7IG5vdywgc3Vj aCBlbGVtZW50IG1heSBiZSAocG90ZW50aWFsbHkpIGludGVyZmVyaW5nIGluIHRoZSBwcmVzZW50 IGRpc2N1c3Npb24gYnV0IHRoaXMgaXMgc2ltcGx5IGEgZmFjdCBmb3Igd2hpY2ggdGhlIENDQU1Q IFdHIChpZmYgYSBjb25zZW5zdXMgaXMgZm91bmQpIGhhcyB0byBjb3BlIHdpdGg7IHRoZSBwcm9i bGVtIHN0YXRlbWVudCBkb2N1bWVudCBoYXMgdHJpZWQgdG8gYmUgYXMgZXhwbGljaXQgYXMgcG9z c2libGUgd2l0aCB0aGF0IHJlc3BlY3Q7IHRoaXMgc2VjdGlvbiA3LjEgc2F5cyB0aGF0IElFRUUg d2lsbCBiZSBjb25zdWx0ZWQgYW5kIGl0cyBmZWVkYmFjayB3aWxsIGJlIGNvbnNpZGVyZWQgaWZm IENDQU1QIGZpbmRzIGVub3VnaCBpbnRlcmVzdCBpbiBzdGFydGluZyBhIHdvcmtpbmcgaXRlbSBv biBHTVBMUyAoSVApIGNvbnRyb2wgcGxhbmUgZm9yIEV0aGVybmV0IChiYXNlZCBvbiB0aGUgcHJv cG9zZWQgc2NlbmFyaW9zKSBzbyBhcyB0byBzdGFydCB3b3JraW5nIG9uIHRoZSByZXF1aXJlbWVu dHMgKGxpc3RlZCBpbiB0aGlzIGRvY3VtZW50KSBhcyBwYXJ0IG9mIHRoZSBDQ0FNUCBXRyBlZmZv cnQgKiBidXQgd2UgYXJlIG5vdCB0aGF0IGZhciB5ZXQgKjwvRk9OVD48L1A+PFA+PEZPTlQgRkFD RT0iTW9ub3NwYWNlLENvdXJpZXIiPmhvd2V2ZXIsIHRoaXMgYXJndW1lbnQgY2FuIG5vdCBiZSB1 c2VkIGFzIHRoZSBjb3JuZXJzdG9uZSBpbiB0aGUgcHJlc2VudCBzdGFnZSBvZiBkZWNpc2lvbiBw cm9jZXNzIHdoaWNoIGlzIHNpbXBseSB3aGV0aGVyIENDQU1QIFdHIHdhbnRzIHRvIHdvcmsgb3Ig bm90IG9uIGEgR01QTFMgKElQKSBjb250cm9sIHBsYW5lIGZvciBFdGhlcm5ldDsgc28sIGl0IGlz IGluYXBwcm9wcmlhdGUgdG8gYXJndWUgZnJvbSB0aGUgQ0NBTVAgcGVyc3BlY3RpdmUgYnkgbWFr aW5nIHVzZSBvZiB0aGUgcG90ZW50aWFsIHJlYWN0aW9uKHMpIG9mIHRoZSBJRUVFLCBpbiBwYXJ0 aWN1bGFyLCBrbm93aW5nIHRoZXJlIGlzIG5vIGludGVudGlvbiB0byBpbml0aWF0ZSBhIHdvcmsg aXRlbSBvbiBFdGhlcm5ldCBkYXRhIHBsYW5lICZuYnNwOzwvRk9OVD48QlI+PEJSPjxGT05UIFNJ WkU9Mj48Qj4mbHQ7bmVpbC4yLmhhcnJpc29uQGJ0LmNvbSZndDs8L0I+PC9GT05UPjxCUj48Rk9O VCBTSVpFPTI+U2VudCBieTogb3duZXItY2NhbXBAb3BzLmlldGYub3JnPC9GT05UPjxCUj48Rk9O VCBTSVpFPTI+MDcvMjIvMjAwNSAyMToxMTwvRk9OVD48QlI+PEJSPiA8Rk9OVCBTSVpFPTI+VG86 PC9GT05UPiA8Rk9OVCBTSVpFPTI+RGltaXRyaSBQQVBBRElNSVRSSU9VL0JFL0FMQ0FURUxAQUxD QVRFTCwgJmx0O2FkcmlhbkBvbGRkb2cuY28udWsmZ3Q7PC9GT05UPjxCUj4gPEZPTlQgU0laRT0y PmNjOjwvRk9OVD4gPEZPTlQgU0laRT0yPiZsdDtjY2FtcEBvcHMuaWV0Zi5vcmcmZ3Q7PC9GT05U PjxCUj4gPEZPTlQgU0laRT0yPmJjYzo8L0ZPTlQ+IDxCUj4gPEZPTlQgU0laRT0yPlN1YmplY3Q6 PC9GT05UPiA8Rk9OVCBTSVpFPTI+UkU6IEZyYW1lZm9ybWF0IGluIGEgbDJjcyBnbXBscyBybnZp cm9ubWVudC48L0ZPTlQ+PEJSPiA8QlI+PEJSPjwvUD48UD5EaW1pdHJpLCZuYnNwOyA8QlI+PEJS PldoYXQgZG8geW91IG1lYW4gYnkgJ2ludGVyZmVycmluZyBlbGVtZW50cycuLi4uLnRoaXMgc291 bmRzIGxpa2Ugc29tZSAnYmVoaW5kIGNsb3NlZCBkb29ycyBzZWN0JyBhdCB3b3JrLiZuYnNwOyBJ IHRob3VnaHQgQWRyaWFuJ3MgcG9zdCB3YXMgKGFzIHVzdWFsKSBwbGFpbi9jbGVhciBhcyB0byBp bnRlbnQuPEJSPjxCUj5yZWdhcmRzLCBOZWlsPEJSPi0tLS0tT3JpZ2luYWwgTWVzc2FnZS0tLS0t PEJSPjxCPkZyb206PC9CPiBvd25lci1jY2FtcEBvcHMuaWV0Zi5vcmcgW21haWx0bzpvd25lci1j Y2FtcEBvcHMuaWV0Zi5vcmddIDxCPk9uIEJlaGFsZiBPZiA8L0I+RGltaXRyaS5QYXBhZGltaXRy aW91QGFsY2F0ZWwuYmU8QlI+PEI+U2VudDo8L0I+IDIyIEp1bHkgMjAwNSAxNjo1MTxCUj48Qj5U bzo8L0I+IEFkcmlhbiBGYXJyZWw8QlI+PEI+Q2M6PC9CPiBEaW1pdHJpLlBhcGFkaW1pdHJpb3VA YWxjYXRlbC5iZTsgY2NhbXBAb3BzLmlldGYub3JnPEJSPjxCPlN1YmplY3Q6PC9CPiBSZTogRnJh bWVmb3JtYXQgaW4gYSBsMmNzIGdtcGxzIHJudmlyb25tZW50LjxCUj48QlI+PEJSPjxCUj48Rk9O VCBTSVpFPTQ+YWRyaWFuIC0gb2sgbXVjaCB0aGFua3MgZm9yIHRoZSBjbGFyaWZpY2F0aW9uIC0g PC9GT05UPjxCUj48QlI+PEZPTlQgU0laRT00Pmkgd2FudGVkIHRvIGJlIHN1cmUgdGhhdCB0aGUg d29ya2luZyBncm91cCBjb25zZW5zdXMgaXMgZ29pbmcgdG8gYmUgdGhlIGVsZW1lbnQgb2YgZGVj aXNpb24gZm9yIHB1dHRpbmcgdGhpcyBpdGVtIGluIHRoZSB3b3JraW5nIGxpc3QgLSBhbmQgd2Ug d29uJ3QgaGF2ZSBvdGhlciBpbnRlcmZlcmluZyBlbGVtZW50cyBpbiB0aGUgZGVjaXNpb24gcHJv Y2VzcyAtIDwvRk9OVD48QlI+PEJSPjxCPiZxdW90O0FkcmlhbiBGYXJyZWwmcXVvdDsgJmx0O2Fk cmlhbkBvbGRkb2cuY28udWsmZ3Q7PC9CPjxCUj5TZW50IGJ5OiBvd25lci1jY2FtcEBvcHMuaWV0 Zi5vcmc8QlI+MDcvMjIvMjAwNSAxNjowODxCUj5QbGVhc2UgcmVzcG9uZCB0byAmcXVvdDtBZHJp YW4gRmFycmVsJnF1b3Q7PEJSPjxCUj5Ubzo8Rk9OVCBTSVpFPTQ+IDwvRk9OVD5EaW1pdHJpIFBB UEFESU1JVFJJT1UvQkUvQUxDQVRFTEBBTENBVEVMPEJSPmNjOjxGT05UIFNJWkU9ND4gPC9GT05U PiZsdDtjY2FtcEBvcHMuaWV0Zi5vcmcmZ3Q7PEJSPmJjYzo8Rk9OVCBTSVpFPTQ+IDwvRk9OVD48 QlI+U3ViamVjdDo8Rk9OVCBTSVpFPTQ+IDwvRk9OVD5SZTogRnJhbWVmb3JtYXQgaW4gYSBsMmNz IGdtcGxzIHJudmlyb25tZW50LjxCUj48QlI+PEJSPjxGT05UIFNJWkU9ND5EaW1pdHJpLDwvRk9O VD48QlI+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7IHdoYXQgZG8geW91IEVYQUNUTFkgbWVhbiB3aXRo IHRoZSBmb2xsb3dpbmcgc2VudGVuY2U8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7PC9GT05U PjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmcXVvdDtUaGlzIHF1ZXN0aW9uIGlzIG9uZSB3ZSBtdXN0 IGNvbWUgdG8gYXMgc29vbiBhcyB3ZSBhcmUgY29uZmlkZW50IHRoYXQ8L0ZPTlQ+PEJSPjxGT05U IFNJWkU9ND50aGU8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7IHJlcXVpcmVtZW50cyBuZWVk IHRvIGJlIGFkZHJlc3NlZCBhdCBhbGwuJnF1b3Q7PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0 OzwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgcGxlYXNlIGNsYXJpZnkgLSBidHcgd2hvIGlz ICZxdW90O3dlJnF1b3Q7IGluIHRoaXMgY29udGV4dCA8L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpF PTQ+JnF1b3Q7V2UmcXVvdDsgbWVhbnMgJnF1b3Q7dGhlIENDQU1QIHdvcmtpbmcgZ3JvdXAmcXVv dDs8L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyB0aGlzIHNlbnRlbmNlIHJlYWRzIHRv IG1lIGFzICZxdW90O0NDQU1QIGhhcyBub3QgeWV0IGNvbW1pdHRlZCB0byB3b3JrIG9uPC9GT05U PjxCUj48Rk9OVCBTSVpFPTQ+dGhpczwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgdG9waWMm cXVvdDsgaXMgbXkgcmVhZGluZyBjb3JyZWN0ID8gb3IgZG8geW91IG1lYW4gc29tZXRoaW5nIGVs c2Ugd2l0aCB0aGlzIDwvRk9OVD48QlI+PEJSPjxGT05UIFNJWkU9ND5Zb3VyIHJlYWRpbmcgaXMg ZXhhY3RseSBjb3JyZWN0IGFzIHN0YXRlZC48L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND5JIHJlZmVy IHlvdSB0byB5b3VyIGVtYWlsIHRvIHRoZSBsaXN0IG9uIDE5dGggSnVseS4gWW91IGFza2VkLi4u PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAtIGlmIHllcywgaXMgdGhlIGludGVyZXN0IGlu IEdNUExTIGZvciBFdGhlcm5ldCBpbi0vb3V0LXNpZGUgdGhlPC9GT05UPjxCUj48Rk9OVCBTSVpF PTQ+Jmd0OyAmbmJzcDsgQ0NBTVAgV0cgY29tbXVuaXR5IHN0cm9uZyBlbm91Z2ggdG8gY29uc2lk ZXIgdGhlIGNvcnIuIEdNUExTPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmbmJzcDsgc2ln bmFsaW5nIGFuZCByb3V0aW5nIHJlcXVpcmVtZW50cyAoc2Vjb25kIHBhcnQgb2YgdGhlIGRvY3Vt ZW50KTwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJm5ic3A7IGFzIHBhcnQgb2YgdGhlIEND QU1QIFdHIGVmZm9ydCA/PC9GT05UPjxCUj48QlI+PEZPTlQgU0laRT00PlRodXMsIHRoZSBwdXJw b3NlIG9mIHRoZSBJLUQgd2FzIHRvIGNyeXN0YWxpemUgdGhlIGF1dGhvcnMnIHRob3VnaHRzIG9u PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+R01QTFMgY29udHJvbCBvZiBFdGhlcm5ldCBzd2l0Y2hp bmcgYW5kIHRvIGluaXRpYXRlIGRpc2N1c3Npb24gb24gd2hldGhlcjwvRk9OVD48QlI+PEZPTlQg U0laRT00PnRoZSBwcm9wb3NlZCBkZXBsb3ltZW50IG1vZGVscyBhcmUgcmVhbGlzdGljIGFuZCBk ZXNpcmFibGUuIEZvbGxvd2luZyBhPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+cG9zaXRpdmUgYW5z d2VyIGFuZCBhc3N1bWluZyBzdXBwb3J0IChieSBDQ0FNUCkgZm9yIHRoZSBkb2luZyB0aGUgd29y ayAoaW48L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND5DQ0FNUCkgYWxsIHRoYXQgcmVtYWlucyBpcyB0 byBhZGQgdGhlIHdvcmsgdG8gdGhlIGNoYXJ0ZXIgKHNlZSBwcmV2aW91czwvRk9OVD48QlI+PEZP TlQgU0laRT00PmNoYXJ0ZXIgZGlzY3Vzc2lvbnMpIGFuZCBnZXQgb24gd2l0aCBpdC48L0ZPTlQ+ PEJSPjxCUj48Rk9OVCBTSVpFPTQ+VGhhbmtzLDwvRk9OVD48QlI+PEZPTlQgU0laRT00PkFkcmlh bjwvRk9OVD48QlI+PC9QPg== Envelope-to: ccamp-data@psg.com Delivery-date: Sat, 23 Jul 2005 04:37:47 +0000 Thread-Topic: Frameformat in a l2cs gmpls rnvironment. thread-index: AcWPQGv1s5h7skVjT2KlrNODX+ABpA== Reply-To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> From: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> To: <richard.spencer@bt.com> Cc: <ccamp@ops.ietf.org> Bcc: Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Sat, 23 Jul 2005 13:38:50 +0900 Comment: GQ19@|@ZEk=E?,18?x, BcN1b<z:P<.F@, 4c4g Message-ID: <4280a01c58f40$6bf849c0$8310fe81@email1> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Class: urn:content-classes:message Richard, IEEE recommends to use Ethernet Length/Type code for distinguishing control frames. All GARP, GVRP, GMPR and other proposed OAM frame follow this rule. IEEE particulary discourages using some number in data forwarding address being used for exchange of control data. An example is using some multicast address for exchange of hello frames. This rule is also applied to the case of VLAN forwarding. I think this policy is to prevent control chaos with other L2 protocols. We need to be careful in taking approach to use some L2-LSP exclusively for control plane because it may result in unwanted conflict with both people and machine. Regards, Jaihyung -----?? ???----- ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com> ?? ??: 2005-07-22 ?? 8:44:08 ?? ??: "per@defero.se" <per@defero.se>, "loa@pi.se" <loa@pi.se> ??: "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> ??: RE: Frameformat in a l2cs gmpls rnvironment. Regardless of whether or not a switch is directly connected to hosts, it must be able to forward packets using the connectionless Ethernet data plane. This is due to the fundamental requirement that for GMPLS switches to be able to exchange control information with each other, a data plane for control traffic must be present. This is akin to using the IP data plane for MPLS signalling in an IP/MPLS network. An alternative would be to use a static reserved L2-LSP for control traffic in the same way that reserved VPI/VCIs are used for PNNI signalling in ATM. Regarding connecting hosts to GMPLS switches, I personally don't think extending L2-LSPs into the enterprise/home network is commercially viable. However, if you do want to use GMPLS switches in the home/enterprise network and for some reason don't want to extend L2-LSPs down to the host then you will not be performing normal Ethernet Mac address switching anyway. Instead you will need some kind of policy on the switch that maps connectionless Ethernet packets (e.g. based on MAC src/dest, 802.1p, VLAN) to a L2-LSP. This is because multiple L2-LSPs to the same destination (e.g. a gateway router) may exist for different services/flows (e.g. video download, VoIP call, etc.). Regards, Richard > -----Original Message----- > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On > Behalf Of Par Mattsson > Sent: 22 July 2005 11:42 > To: Loa Andersson > Cc: ccamp@ops.ietf.org > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > > Per and Dimitri, > > > > I would like to come down stronger than that, for me it is > > a very strong requirement that the same switch can handle > > both VLANs and GMPLs trafic correctly. I can't dsee how that > > could be done if using the VLAN tpid to indicate GMPLS > > traffic. > > If you ever want that same switch to handle traffic for a directly > connected host (not to uncommen) you would want that to use normal > ethernet macaddress switching. So of course you do not want to have to > choose between vlan and gmpls, you would want both at the same time. > > /per > > > > > > /Loa > > > > Par Mattsson wrote: > >>>hi par, one of the possibilities that has been considered > to cope with > >>>this requirement is to use a dedicated TPID for the > Ethernet labeled > >>>frames; this would allow differentiated processing with non-labeled > >>>framesthanks. > >> > >> > >> That seems to make more sence. If that frame is to be sized like a > >> 802.1q > >> frame. There is not that much space left to a label. Or is > the demand to > >> use jumboframes ? > >> Has there been any discussion on labelstacking, and mainly where to > >> place > >> the information? > >> > >> Regards. > >> Per > >> > >> > >> > > > > > > -- > > Loa Andersson > > > > Principal Networking Architect > > Acreo AB phone: +46 8 632 77 14 > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > Kista, Sweden email: loa.andersson@acreo.se > > loa@pi.se > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Sat, 23 Jul 2005 04:20:01 +0000 Thread-Topic: Frameformat in a l2cs gmpls rnvironment. thread-index: AcWPPbTjX+0i5VMGS12b9FL9kHYtdQ== Reply-To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> From: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> To: "Heiles Juergen" <juergen.heiles@siemens.com>, <per@defero.se> Cc: <ccamp@ops.ietf.org> Bcc: Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Sat, 23 Jul 2005 13:19:24 +0900 Comment: GQ19@|@ZEk=E?,18?x, BcN1b<z:P<.F@, 4c4g Message-ID: <427a201c58f3d$b4e59910$8310fe81@email1> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Class: urn:content-classes:message Dear Juergen, Par and all, The proposed framework of L2SC was not intended to suggest any specific solution. It is still an open question. As Adrian noted, it only aims to present requirements as in scenarios that may drive people toward next step in many viable solutions. In fact, there are other options we may also consider for L2 label encoding. IEEE 802.1 Ethernet bridges forward frames based on 48bits of MAC address, and additionally using VLAN tag. When the purpose of GMPLS control over Ethernet is not to create new dataplane but to utilize IEEE 802.1 bridge architecture, we may consider using one of the two forwarding methods, MAC forwarding or VLAN forwarding. (and perhaps any other combination of fields in MAC, but I'll not discuss it) Use of VLAN ID for label encoding may automate VLAN configuration using IP protocols. However, GMPLS protocol cannot use the field exclusively because public/private operators already use VLAN for various purpose. There is a potential conflict with existing use of VLAN and GMPLS use of VLAN label. Furthermore, scalability of VLAN ID has been frequently noted as weakness because the size of VLAN ID is at most 4096 (12bit). The scalability may be improved if the scope of VLAN label is confined to link-local, and some additional swapping function of VLAN ID is introduced in Internal Sublayer [802.1D] of GMPLS implemented switch. However, this will only be effective when the configuration of network is mesh structure that multiple LSP paths exist. If the configuration of network is star or tree shape, as normal configuration of access network, LSPs concentrate in root node and total number of LSPs that the network can hold still be limited by available label space at a few root links. For these reasons, I do not think any form of VLAN ID label is an appropriate choice for layer-2 label encoding. There are some other proposals assuming new assignment of Ethernet Length/Type value (e.g. new TPID in VLAN tag) and re-definition of information fields placed between 802.3 MAC header and IP packet. In this case, only the format of VLAN tag or extended VLAN tag is borrowed, however, inside the switching hardware, the filter and relay, etc. are totally different new dataplane switch. I do not see such approach is a GMPLS implementation for Ethernet because the core switching technique is not 802.1 Ethernet bridge at all. The other option we may consider is using MAC address filed as below. +-------+-------+-------+-------+-------+-------+ | 1byte | 2byte | 3byte | 4byte | 5byte | 6byte | +-------+-------+-------+-------+-------+-------+ +-----------------------+-----------------------+ | OUI Prefix (=GMPLS) | DA-label (24bit) | +-----------------------+-----------------------+ | OUI Prefix (=GMPLS) | SA-label (24bit) | +---------------+-------+-----------------------+ | Length/Type | +---------------+ IEEE is designated by the ISO Council to act as the registration authority for the higher three-octet of OUI number in the MAC address to be used by manufacturer. Ethernet manufacturer may generate global unique MAC address using the OUI prefix and address block of lower three-octet (24bit). Taking advantage of the addressing scheme, GMPLS may use the lower three-octet exclusively if a unique OUI number is reserved for the protocol. With this labeling scheme, GMPLS will control MAC forwarding entry, not VLAN table. All Ethernet frames controlled by GMPLS will have identical OUI number that they can easily be distinguished from other Ethernet frames. In principle, the label lookup hardware is identical to MAC lookup hardware in this labeling scheme. Therefore GMPLS implemented switch may still function as normal Ethernet bridge to the frames that OUI number is not GMPLS. This also facilitates GMPLS implemented switches being deployed in operating Ethernet with minimum service disruption. Note also that above proposed label encoding method is transparent to the use of Ethernet Length/Type field. End-user device may use the Length/Type field as defined in IEEE 802.3 protocol. It also allows network operators configure VLAN for their own purpose. When IEEE 802.1p is used in conjunction with L2-LSP, the priority field of VLAN tag can also be used for imposing consistent TE policy in legacy switches and GMPLS switches. Any way, my conclusion is, there are other options we may consider, and this issue is still open to discuss. Thanks, Sincerely, Jaihyung -----?? ???----- ?? ??: "Heiles Juergen" <juergen.heiles@siemens.com> ?? ??: 2005-07-22 ?? 10:23:43 ?? ??: "Loa Andersson" <loa@pi.se>, "richard.spencer@bt.com" <richard.spencer@bt.com> ??: "per@defero.se" <per@defero.se>, "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> ??: RE: Frameformat in a l2cs gmpls rnvironment. Loa, I interpret the ID as a proposal to use GMPLS for VLAN setup. So GMPLS and VLAN to not compete. The VLAN is at the data plane and GMPLS at the control plane. The question is how and should different control plane techniques like GMPLS and (GVRP and STB) work together? Regards Juergen > -----Original Message----- > From: owner-ccamp@ops.ietf.org > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson > Sent: Friday, July 22, 2005 2:53 PM > To: richard.spencer@bt.com > Cc: per@defero.se; ccamp@ops.ietf.org > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > Richard, > > I agree to most of this. Since we have doubts about the viability > of taking GMPLS all the way to end-user or enterprise I think it > would be good, from a wg perspective, if we agreed to solve the > core network problems first. > > Do you have any comment on the requirement to run both VLANs and > GMPLS on the same switch? > > /Loa > > richard.spencer@bt.com wrote: > > Regardless of whether or not a switch is directly connected > > to hosts, it must be able to forward packets using the > connectionless > > Ethernet data plane. This is due to the fundamental > requirement that > > for GMPLS switches to be able to exchange control information with > > each other, a data plane for control traffic must be present. > > This is akin to using the IP data plane for MPLS signalling in an > > IP/MPLS network. An alternative would be to use a static reserved > > L2-LSP for control traffic in the same way that reserved VPI/VCIs > > are used for PNNI signalling in ATM. > > > > Regarding connecting hosts to GMPLS switches, I personally don't > > think extending L2-LSPs into the enterprise/home network is > > commercially viable. However, if you do want to use GMPLS switches > > in the home/enterprise network and for some reason don't want to > > extend L2-LSPs down to the host then you will not be performing > > normal Ethernet Mac address switching anyway. Instead you will > > need some kind of policy on the switch that maps connectionless > > Ethernet packets (e.g. based on MAC src/dest, 802.1p, VLAN) to a > > L2-LSP. This is because multiple L2-LSPs to the same destination > > (e.g. a gateway router) may exist for different services/flows > > (e.g. video download, VoIP call, etc.). > > > > Regards, > > Richard > > > > > >>-----Original Message----- > >>From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On > >>Behalf Of Par Mattsson > >>Sent: 22 July 2005 11:42 > >>To: Loa Andersson > >>Cc: ccamp@ops.ietf.org > >>Subject: Re: Frameformat in a l2cs gmpls rnvironment. > >> > >> > >> > >>>Per and Dimitri, > >>> > >>>I would like to come down stronger than that, for me it is > >>>a very strong requirement that the same switch can handle > >>>both VLANs and GMPLs trafic correctly. I can't dsee how that > >>>could be done if using the VLAN tpid to indicate GMPLS > >>>traffic. > >> > >>If you ever want that same switch to handle traffic for a directly > >>connected host (not to uncommen) you would want that to use normal > >>ethernet macaddress switching. So of course you do not want > to have to > >>choose between vlan and gmpls, you would want both at the same time. > >> > >>/per > >> > >> > >> > >>>/Loa > >>> > >>>Par Mattsson wrote: > >>> > >>>>>hi par, one of the possibilities that has been considered > >> > >>to cope with > >> > >>>>>this requirement is to use a dedicated TPID for the > >> > >>Ethernet labeled > >> > >>>>>frames; this would allow differentiated processing with > non-labeled > >>>>>framesthanks. > >>>> > >>>> > >>>>That seems to make more sence. If that frame is to be sized like a > >>>>802.1q > >>>>frame. There is not that much space left to a label. Or is > >> > >>the demand to > >> > >>>>use jumboframes ? > >>>>Has there been any discussion on labelstacking, and > mainly where to > >>>>place > >>>>the information? > >>>> > >>>>Regards. > >>>>Per > >>>> > >>>> > >>>> > >>> > >>> > >>>-- > >>>Loa Andersson > >>> > >>>Principal Networking Architect > >>>Acreo AB phone: +46 8 632 77 14 > >>>Isafjordsgatan 22 mobile: +46 739 81 21 64 > >>>Kista, Sweden email: loa.andersson@acreo.se > >>> loa@pi.se > >>> > >> > >> > >> > > > > > -- > Loa Andersson > > Principal Networking Architect > Acreo AB phone: +46 8 632 77 14 > Isafjordsgatan 22 mobile: +46 739 81 21 64 > Kista, Sweden email: loa.andersson@acreo.se > loa@pi.se > Envelope-to: ccamp-data@psg.com Delivery-date: Sat, 23 Jul 2005 01:58:17 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: RE: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Date: Fri, 22 Jul 2005 18:55:51 -0700 Message-ID: <626FC7C6A97381468FB872072AB5DDC8369421@XCH-SW-42.sw.nos.boeing.com> Thread-Topic: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Thread-Index: AcWO6PXwNV9J9GsyTyywah8dc4SA0QAESRbgAAQtiVAAB5V/cA== From: "Drake, John E" <John.E.Drake2@boeing.com> To: "Drake, John E" <John.E.Drake2@boeing.com>, <neil.2.harrison@bt.com>, <adrian@olddog.co.uk>, <richard.spencer@bt.com>, <juergen.heiles@siemens.com>, <loa@pi.se>, <per@defero.se> Cc: <ccamp@ops.ietf.org> Snipped ... > > IP is cl-ps networking techology period....and its > > jolly important, but PLEASE don't try and feed me any of this 'IP > > control plane' nonsense. JD: Over the years I have enjoyed Neil's rants immensely, but I have to say that this one is a classic. Envelope-to: ccamp-data@psg.com Delivery-date: Sat, 23 Jul 2005 00:35:00 +0000 Content-class: urn:content-classes:message Subject: RE: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Date: Fri, 22 Jul 2005 20:32:18 -0400 Message-ID: <34B3EAA5B3066A42914D28C5ECF5FEA403315037@zrtphxm2> Thread-Topic: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Thread-Index: AcWPEjVr/Qfq0W46QTOtTRYscPoclQAB4g/gAACM/WA= From: "Don Fedyk" <dwfedyk@nortel.com> To: "Adrian Farrel" <adrian@olddog.co.uk>, <neil.2.harrison@bt.com> Cc: <ccamp@ops.ietf.org> Hi Neil I concur with Adrian's definition of "IP Control Plane" or "IP-based Control Plane". I think your point is more to the question of where the intelligence to select the network paths lives. I believe that in GMPLS, Path selection and path control is not solely the responsibility of "IP-based control plane" but is enabled by IP mechanisms to be where it makes sense. Regards, Don >-----Original Message----- From: owner-ccamp@ops.ietf.org >[mailto:owner-ccamp@ops.ietf.org] On Behalf Of Adrian Farrel >Sent: Friday, July 22, 2005 7:08 PM > > >Hi Neil, > >I think John beat me off the blocks here, but... > > >> > GMPLS assumes an IP control plane. >> >> An IP control-plane? There is actually no such animal. >> Just what the heck does that REALLY mean in GMPLS say? > >Let me explain. Perhaps I should have said "IP-based >control plane". I mean a control plane which: >- uses IP as its network protocol >- uses IP addresses to identify control plane resources >- uses IP addresses to identify data plane resources within > the control plane >- uses protocols developed for use in the Internet. > >> I am not questioning IP as a cl-ps networking protocol >> *carrying* a signalling protocol (RSVP-TE, or dare I >> mention PNNI, or any signalling protocol yet to be >> invented) or a routing protocol (OSPF or ISO or whatever) > >I am glad to hear it. > <snip> Envelope-to: ccamp-data@psg.com Delivery-date: Sat, 23 Jul 2005 00:03:56 +0000 Message-ID: <01f801c58f1a$2adaa120$b7849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <richard.spencer@bt.com> Cc: <ccamp@ops.ietf.org> Subject: Re: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Date: Sat, 23 Jul 2005 01:04:18 +0100 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_01F1_01C58F22.73B35600" This is a multi-part message in MIME format. ------=_NextPart_000_01F1_01C58F22.73B35600 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Hi Richard, I see you're in the same timezone as me :-) > I don't understand why you have renamed the thread "Ethernet Control = Plane" > in response to my email. I haven't even mentioned an Ethernet Control = Plane. Hmmm. I seem to be having some problems with adjectival constructs = today. How about "Control Plane for Ethernet"? I must say, what you wrote sounded like a discussion of establishing a = mechanism to carry control plane traffic over Ethernet for the purpose = of controlling Ethernet. I don't think "Ethernet Control Plane" is such = a bad term for this. But there is a lesson here. If the subject line needs changing (to help = separate the several threads that are running with the same subject = line) and if you don't change it yourself, someone else will. If you = don't like that subject line, change it to something better. Cheers, Adrian > I'm quite aware that GMPLS uses IP labelled control protocol packets = and > I haven't suggested removing IP and just using Ethernet. In fact, I = would > have thought my analogy with how a management VLAN is used today (e.g. = > for ICMP ping and telnet) would have demonstrated that. > > I am simply saying that for two GMPLS peers to send (IP labelled) = control > protocol packets to each other, they must agree to encapsulate the = packets > using either untagged Ethernet frames or using tagged Ethernet frames = with > a common VLAN ID. Again, I'm not saying this is an issue, just = something=20 > that needs to be considered. Switches are not routers and normally = only=20 > send/receive locally addressed IP packets via the management VLAN, = e.g.=20 > to ping between two directly connected switches the interfaces they = are=20 > connected via must both be configured to be in the same management = VLAN. Regards, Richard > -----Original Message----- > From: Adrian Farrel [mailto:adrian@olddog.co.uk] > Sent: 22 July 2005 19:13 > To: Spencer,R,Richard,XDE73 R; juergen.heiles@siemens.com; loa@pi.se; > per@defero.se > Cc: ccamp@ops.ietf.org > Subject: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls > rnvironment.] >=20 >=20 > Whoa there! >=20 > GMPLS assumes an IP control plane. > I am aware that there is work afoot in other SDOs to assign a=20 > dedicated > VLAN tag to specify the control plane traffic. This is not what we are > doing. >=20 > As far as I am aware, everyone in the IETF is comfortable=20 > with the idea of > using IP packets to carry control plane data through an=20 > IP-enabled DCN. In > Ethernet the issue will be whether the control plane is in=20 > band (use IP > address and MAC address) or out of band (use distinct DCN). >=20 > The use of a separate VLAN ID to signify the control plane is=20 > out of scope > for CCAMP. >=20 > Thanks, > Adrian >=20 > ----- Original Message -----=20 > From: <richard.spencer@bt.com> > To: <juergen.heiles@siemens.com>; <loa@pi.se>; <per@defero.se> > Cc: <ccamp@ops.ietf.org> > Sent: Friday, July 22, 2005 4:33 PM > Subject: RE: Frameformat in a l2cs gmpls rnvironment. >=20 >=20 > Juergen, >=20 > Before you can use GMPLS to set up VLANs a data plane must=20 > first exist to > carry the GMPLS control plane traffic. Yes you can use the standard > Ethernet data plane for this purpose (and in fact must do so for the > solution to be viable), but its not as simple as that. You=20 > cannot assume > that just because two GMPLS peers are directly connected that they can > communicate with each other. >=20 > To exchange GMPLS control plane packets between a pair of=20 > GMPLS peers, the > peers will need to agree to use either untagged control=20 > packets or to tag > them with a common VLAN ID. The use of a separate VLAN for=20 > GMPLS control > traffic is analogous to using a separate VLAN for management traffic > today. If two switches belong to the same management VLAN=20 > then you can use > ICMP ping to verify connectivity between them. However, if=20 > one switch has > been put into the wrong management VLAN then the ping will=20 > obviously fail. >=20 > If you want to use a separate VLAN for GMPLS control plane=20 > traffic, then > you either need to (i) define a reserved GMPLS signalling=20 > VLAN ID, or (ii) > leave it up to the operator to choose a VLAN ID. If you leave=20 > it up to the > operator to decide what VLAN ID to use, then you need to=20 > consider what the > implications in the inter-operator case will be, i.e. what=20 > VLAN ID range > should be used for control plane traffic and what VLAN ID=20 > range should be > used for user plane traffic. >=20 > I'm not saying that defining how GMPLS control plane packets=20 > are forwarded > in the Ethernet data plane will be an issue, simply that this=20 > is something > that must be addressed to develop interoperable solutions. >=20 > Regards, > Richard >=20 > > -----Original Message----- > > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On > > Behalf Of Heiles Juergen > > Sent: 22 July 2005 14:17 > > To: Loa Andersson; P=E4r Mattsson > > Cc: ccamp@ops.ietf.org > > Subject: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > GMPLS for SONET/SDH has no impact on the SONET/SDH data > > plane. It introduces a control plane for connection setup > > instead of a network management based connection setup. GMPLS > > for Ethernet can work in the same way. Ethernet as such is > > cl, but VLANs are setup and GMPLS can provide a control plane > > for VLAN setup. The Ethernet data plane is not changed and > > MAC address learning and forwarding is still be done within a > > VLAN. So no change to the Ethernet data plane. No GMPLS > > traffic is introduced, it is still VLAN traffic. > > Introducing a dedicated GMPLS label for Ethernet (GMPLS > > traffic) goes beyond GMPLS as a control plane technique. > > > > Regards > > > > Juergen > > > > > > > > > > > -----Original Message----- > > > From: owner-ccamp@ops.ietf.org > > > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson > > > Sent: Friday, July 22, 2005 12:05 PM > > > To: P=E4r Mattsson > > > Cc: ccamp@ops.ietf.org > > > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > > > > Per and Dimitri, > > > > > > I would like to come down stronger than that, for me it is > > > a very strong requirement that the same switch can handle > > > both VLANs and GMPLs trafic correctly. I can't dsee how that > > > could be done if using the VLAN tpid to indicate GMPLS > > > traffic. > > > > > > /Loa > > > > > > P=E4r Mattsson wrote: > > > >>hi par, one of the possibilities that has been considered > > > to cope with > > > >>this requirement is to use a dedicated TPID for the > > Ethernet labeled > > > >>frames; this would allow differentiated processing with > > non-labeled > > > >>framesthanks. > > > > > > > > > > > > That seems to make more sence. If that frame is to be sized > > > like a 802.1q > > > > frame. There is not that much space left to a label. Or is > > > the demand to > > > > use jumboframes ? > > > > Has there been any discussion on labelstacking, and mainly > > > where to place > > > > the information? > > > > > > > > Regards. > > > > Per > > > > > > > > > > > > > > > > > > > > > --=20 > > > Loa Andersson > > > > > > Principal Networking Architect > > > Acreo AB phone: +46 8 632 77 14 > > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > > Kista, Sweden email: loa.andersson@acreo.se > > > loa@pi.se > > > > > > > >=20 >=20 >=20 >=20 ------=_NextPart_000_01F1_01C58F22.73B35600 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META http-equiv=3DContent-Type content=3D"text/html; = charset=3Diso-8859-1"> <META content=3D"MSHTML 6.00.2800.1106" name=3DGENERATOR> <STYLE></STYLE> </HEAD> <BODY> <DIV><FONT face=3DCourier size=3D2>Hi Richard,</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>I see you're in the same timezone as = me=20 :-)</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>> I don't understand why you have = renamed the=20 thread "Ethernet Control Plane"</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> in response to my email. I = haven't even=20 mentioned an Ethernet Control Plane.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Hmmm. I seem to be having some = problems with=20 adjectival constructs today.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>How about "Control Plane for=20 Ethernet"?</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>I must say, what you wrote sounded = like a=20 discussion of establishing a mechanism to carry control plane traffic = over=20 Ethernet for the purpose of controlling Ethernet. I don't think = "Ethernet=20 Control Plane" is such a bad term for this.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>But there is a lesson here. If the = subject line=20 needs changing (to help separate the several threads that are running = with the=20 same subject line) and if you don't change it yourself, someone else = will. If=20 you don't like that subject line, change it to something = better.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Cheers,</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Adrian</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>> I'm quite aware that GMPLS uses = IP labelled=20 control protocol packets and</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> I haven't suggested = removing IP and=20 just using Ethernet. In fact, I would</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> have thought my analogy = with how a=20 management VLAN is used today (e.g. </FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> for ICMP ping and telnet) would = have=20 demonstrated that.<BR>><BR>> I am simply saying that for two GMPLS = peers=20 to send (IP labelled) control</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> protocol packets to each = other, they=20 must agree to encapsulate the packets</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> using either untagged = Ethernet frames=20 or using tagged Ethernet frames with</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> a common VLAN ID. Again, = I'm not saying=20 this is an issue, just something </FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> that needs to be considered. = Switches are=20 not routers and normally only </FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> send/receive locally addressed = IP packets=20 via the management VLAN, e.g. </FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> to ping between two directly = connected=20 switches the interfaces they are </FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> connected via must both be = configured to be=20 in the same management VLAN.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2> </DIV> <DIV><BR><BR>Regards,<BR>Richard<BR><BR>> -----Original = Message-----<BR>>=20 From: Adrian Farrel [mailto:adrian@olddog.co.uk]<BR>> Sent: 22 July = 2005=20 19:13<BR>> To: Spencer,R,Richard,XDE73 R; </FONT><A=20 href=3D"mailto:juergen.heiles@siemens.com"><FONT face=3DCourier=20 size=3D2>juergen.heiles@siemens.com</FONT></A><FONT face=3DCourier = size=3D2>;=20 </FONT><A href=3D"mailto:loa@pi.se"><FONT face=3DCourier=20 size=3D2>loa@pi.se</FONT></A><FONT face=3DCourier size=3D2>;<BR>> = </FONT><A=20 href=3D"mailto:per@defero.se"><FONT face=3DCourier=20 size=3D2>per@defero.se</FONT></A><BR><FONT face=3DCourier size=3D2>> = Cc: </FONT><A=20 href=3D"mailto:ccamp@ops.ietf.org"><FONT face=3DCourier=20 size=3D2>ccamp@ops.ietf.org</FONT></A><BR><FONT face=3DCourier = size=3D2>> Subject:=20 Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls<BR>>=20 rnvironment.]<BR>> <BR>> <BR>> Whoa there!<BR>> <BR>> = GMPLS=20 assumes an IP control plane.<BR>> I am aware that there is work afoot = in=20 other SDOs to assign a <BR>> dedicated<BR>> VLAN tag to specify = the=20 control plane traffic. This is not what we are<BR>> doing.<BR>> = <BR>>=20 As far as I am aware, everyone in the IETF is comfortable <BR>> with = the idea=20 of<BR>> using IP packets to carry control plane data through an = <BR>>=20 IP-enabled DCN. In<BR>> Ethernet the issue will be whether the = control plane=20 is in <BR>> band (use IP<BR>> address and MAC address) or out of = band (use=20 distinct DCN).<BR>> <BR>> The use of a separate VLAN ID to signify = the=20 control plane is <BR>> out of scope<BR>> for CCAMP.<BR>> = <BR>>=20 Thanks,<BR>> Adrian<BR>> <BR>> ----- Original Message ----- = <BR>>=20 From: <</FONT><A href=3D"mailto:richard.spencer@bt.com"><FONT = face=3DCourier=20 size=3D2>richard.spencer@bt.com</FONT></A><FONT face=3DCourier = size=3D2>><BR>>=20 To: <</FONT><A href=3D"mailto:juergen.heiles@siemens.com"><FONT = face=3DCourier=20 size=3D2>juergen.heiles@siemens.com</FONT></A><FONT face=3DCourier = size=3D2>>;=20 <</FONT><A href=3D"mailto:loa@pi.se"><FONT face=3DCourier=20 size=3D2>loa@pi.se</FONT></A><FONT face=3DCourier size=3D2>>; = <</FONT><A=20 href=3D"mailto:per@defero.se"><FONT face=3DCourier=20 size=3D2>per@defero.se</FONT></A><FONT face=3DCourier = size=3D2>><BR>> Cc:=20 <</FONT><A href=3D"mailto:ccamp@ops.ietf.org"><FONT face=3DCourier=20 size=3D2>ccamp@ops.ietf.org</FONT></A><FONT face=3DCourier = size=3D2>><BR>> Sent:=20 Friday, July 22, 2005 4:33 PM<BR>> Subject: RE: Frameformat in a l2cs = gmpls=20 rnvironment.<BR>> <BR>> <BR>> Juergen,<BR>> <BR>> Before = you can=20 use GMPLS to set up VLANs a data plane must <BR>> first exist = to<BR>>=20 carry the GMPLS control plane traffic. Yes you can use the = standard<BR>>=20 Ethernet data plane for this purpose (and in fact must do so for = the<BR>>=20 solution to be viable), but its not as simple as that. You <BR>> = cannot=20 assume<BR>> that just because two GMPLS peers are directly connected = that=20 they can<BR>> communicate with each other.<BR>> <BR>> To = exchange GMPLS=20 control plane packets between a pair of <BR>> GMPLS peers, = the<BR>> peers=20 will need to agree to use either untagged control <BR>> packets or to = tag<BR>> them with a common VLAN ID. The use of a separate VLAN for = <BR>>=20 GMPLS control<BR>> traffic is analogous to using a separate VLAN for=20 management traffic<BR>> today. If two switches belong to the same = management=20 VLAN <BR>> then you can use<BR>> ICMP ping to verify connectivity = between=20 them. However, if <BR>> one switch has<BR>> been put into the = wrong=20 management VLAN then the ping will <BR>> obviously fail.<BR>> = <BR>> If=20 you want to use a separate VLAN for GMPLS control plane <BR>> = traffic,=20 then<BR>> you either need to (i) define a reserved GMPLS signalling = <BR>>=20 VLAN ID, or (ii)<BR>> leave it up to the operator to choose a VLAN = ID. If you=20 leave <BR>> it up to the<BR>> operator to decide what VLAN ID to = use, then=20 you need to <BR>> consider what the<BR>> implications in the=20 inter-operator case will be, i.e. what <BR>> VLAN ID range<BR>> = should be=20 used for control plane traffic and what VLAN ID <BR>> range should = be<BR>>=20 used for user plane traffic.<BR>> <BR>> I'm not saying that = defining how=20 GMPLS control plane packets <BR>> are forwarded<BR>> in the = Ethernet data=20 plane will be an issue, simply that this <BR>> is something<BR>> = that must=20 be addressed to develop interoperable solutions.<BR>> <BR>>=20 Regards,<BR>> Richard<BR>> <BR>> > -----Original=20 Message-----<BR>> > From: </FONT><A=20 href=3D"mailto:owner-ccamp@ops.ietf.org"><FONT face=3DCourier=20 size=3D2>owner-ccamp@ops.ietf.org</FONT></A><FONT face=3DCourier = size=3D2>=20 [mailto:owner-ccamp@ops.ietf.org]On<BR>> > Behalf Of Heiles=20 Juergen<BR>> > Sent: 22 July 2005 14:17<BR>> > To: Loa = Andersson;=20 P=E4r Mattsson<BR>> > Cc: </FONT><A = href=3D"mailto:ccamp@ops.ietf.org"><FONT=20 face=3DCourier size=3D2>ccamp@ops.ietf.org</FONT></A><BR><FONT = face=3DCourier=20 size=3D2>> > Subject: RE: Frameformat in a l2cs gmpls = rnvironment.<BR>>=20 ><BR>> ><BR>> > GMPLS for SONET/SDH has no impact on the=20 SONET/SDH data<BR>> > plane. It introduces a control plane for = connection=20 setup<BR>> > instead of a network management based connection = setup.=20 GMPLS<BR>> > for Ethernet can work in the same way. Ethernet as = such=20 is<BR>> > cl, but VLANs are setup and GMPLS can provide a control=20 plane<BR>> > for VLAN setup. The Ethernet data plane is not = changed=20 and<BR>> > MAC address learning and forwarding is still be done = within=20 a<BR>> > VLAN. So no change to the Ethernet data plane. No = GMPLS<BR>>=20 > traffic is introduced, it is still VLAN traffic.<BR>> > = Introducing a=20 dedicated GMPLS label for Ethernet (GMPLS<BR>> > traffic) goes = beyond=20 GMPLS as a control plane technique.<BR>> ><BR>> > = Regards<BR>>=20 ><BR>> > Juergen<BR>> ><BR>> ><BR>> ><BR>> = ><BR>> > > -----Original Message-----<BR>> > > = From:=20 </FONT><A href=3D"mailto:owner-ccamp@ops.ietf.org"><FONT face=3DCourier=20 size=3D2>owner-ccamp@ops.ietf.org</FONT></A><BR><FONT face=3DCourier = size=3D2>>=20 > > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa = Andersson<BR>>=20 > > Sent: Friday, July 22, 2005 12:05 PM<BR>> > > To: = P=E4r=20 Mattsson<BR>> > > Cc: </FONT><A = href=3D"mailto:ccamp@ops.ietf.org"><FONT=20 face=3DCourier size=3D2>ccamp@ops.ietf.org</FONT></A><BR><FONT = face=3DCourier=20 size=3D2>> > > Subject: Re: Frameformat in a l2cs gmpls=20 rnvironment.<BR>> > ><BR>> > > Per and = Dimitri,<BR>> >=20 ><BR>> > > I would like to come down stronger than that, for = me it=20 is<BR>> > > a very strong requirement that the same switch can=20 handle<BR>> > > both VLANs and GMPLs trafic correctly. I can't = dsee how=20 that<BR>> > > could be done if using the VLAN tpid to indicate=20 GMPLS<BR>> > > traffic.<BR>> > ><BR>> > >=20 /Loa<BR>> > ><BR>> > > P=E4r Mattsson wrote:<BR>> = > >=20 >>hi par, one of the possibilities that has been = considered<BR>> >=20 > to cope with<BR>> > > >>this requirement is to use a = dedicated TPID for the<BR>> > Ethernet labeled<BR>> > >=20 >>frames; this would allow differentiated processing with<BR>> = >=20 non-labeled<BR>> > > >>framesthanks.<BR>> > >=20 ><BR>> > > ><BR>> > > > That seems to make = more=20 sence. If that frame is to be sized<BR>> > > like a = 802.1q<BR>> >=20 > > frame. There is not that much space left to a label. Or = is<BR>>=20 > > the demand to<BR>> > > > use jumboframes ?<BR>> = >=20 > > Has there been any discussion on labelstacking, and = mainly<BR>>=20 > > where to place<BR>> > > > the information?<BR>> = >=20 > ><BR>> > > > Regards.<BR>> > > > = Per<BR>>=20 > > ><BR>> > > ><BR>> > > ><BR>> = >=20 ><BR>> > ><BR>> > > -- <BR>> > > Loa=20 Andersson<BR>> > ><BR>> > > Principal Networking=20 Architect<BR>> > > Acreo=20 AB  = ; = =20 phone: +46 8 632 77 14<BR>> > > Isafjordsgatan=20 22  = ; =20 mobile: +46 739 81 21 64<BR>> > > Kista,=20 Sweden &= nbsp; =20 email: </FONT><A href=3D"mailto:loa.andersson@acreo.se"><FONT = face=3DCourier=20 size=3D2>loa.andersson@acreo.se</FONT></A><BR><FONT face=3DCourier = size=3D2>> >=20 > &nb= sp; &nbs= p;  = ; =20 </FONT><A href=3D"mailto:loa@pi.se"><FONT face=3DCourier=20 size=3D2>loa@pi.se</FONT></A><BR><FONT face=3DCourier size=3D2>> > = ><BR>>=20 ><BR>> ><BR>> <BR>> <BR>> <BR>>=20 <BR><BR></DIV></FONT></BODY></HTML> ------=_NextPart_000_01F1_01C58F22.73B35600-- Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 23:43:46 +0000 Content-Class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Date: Sat, 23 Jul 2005 00:42:33 +0100 Message-ID: <B5E87B043D4C514389141E2661D255EC0A835AE8@i2km41-ukdy.domain1.systemhost.net> Thread-Topic: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Thread-Index: AcWO6LXPpMoDbQdjRGKKAn7jG/1uNgAKDu4A From: <richard.spencer@bt.com> To: <adrian@olddog.co.uk>, <juergen.heiles@siemens.com>, <loa@pi.se>, <per@defero.se> Cc: <ccamp@ops.ietf.org> Adrian, I don't understand why you have renamed the thread "Ethernet Control = Plane" in response to my email. I haven't even mentioned an Ethernet = Control Plane. I'm quite aware that GMPLS uses IP labelled control protocol packets and = I haven't suggested removing IP and just using Ethernet. In fact, I = would have thought my analogy with how a management VLAN is used today = (e.g. for ICMP ping and telnet) would have demonstrated that. I am simply saying that for two GMPLS peers to send (IP labelled) = control protocol packets to each other, they must agree to encapsulate = the packets using either untagged Ethernet frames or using tagged = Ethernet frames with a common VLAN ID. Again, I'm not saying this is an = issue, just something that needs to be considered. Switches are not = routers and normally only send/receive locally addressed IP packets via = the management VLAN, e.g. to ping between two directly connected = switches the interfaces they are connected via must both be configured = to be in the same management VLAN. Regards, Richard > -----Original Message----- > From: Adrian Farrel [mailto:adrian@olddog.co.uk] > Sent: 22 July 2005 19:13 > To: Spencer,R,Richard,XDE73 R; juergen.heiles@siemens.com; loa@pi.se; > per@defero.se > Cc: ccamp@ops.ietf.org > Subject: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls > rnvironment.] >=20 >=20 > Whoa there! >=20 > GMPLS assumes an IP control plane. > I am aware that there is work afoot in other SDOs to assign a=20 > dedicated > VLAN tag to specify the control plane traffic. This is not what we are > doing. >=20 > As far as I am aware, everyone in the IETF is comfortable=20 > with the idea of > using IP packets to carry control plane data through an=20 > IP-enabled DCN. In > Ethernet the issue will be whether the control plane is in=20 > band (use IP > address and MAC address) or out of band (use distinct DCN). >=20 > The use of a separate VLAN ID to signify the control plane is=20 > out of scope > for CCAMP. >=20 > Thanks, > Adrian >=20 > ----- Original Message -----=20 > From: <richard.spencer@bt.com> > To: <juergen.heiles@siemens.com>; <loa@pi.se>; <per@defero.se> > Cc: <ccamp@ops.ietf.org> > Sent: Friday, July 22, 2005 4:33 PM > Subject: RE: Frameformat in a l2cs gmpls rnvironment. >=20 >=20 > Juergen, >=20 > Before you can use GMPLS to set up VLANs a data plane must=20 > first exist to > carry the GMPLS control plane traffic. Yes you can use the standard > Ethernet data plane for this purpose (and in fact must do so for the > solution to be viable), but its not as simple as that. You=20 > cannot assume > that just because two GMPLS peers are directly connected that they can > communicate with each other. >=20 > To exchange GMPLS control plane packets between a pair of=20 > GMPLS peers, the > peers will need to agree to use either untagged control=20 > packets or to tag > them with a common VLAN ID. The use of a separate VLAN for=20 > GMPLS control > traffic is analogous to using a separate VLAN for management traffic > today. If two switches belong to the same management VLAN=20 > then you can use > ICMP ping to verify connectivity between them. However, if=20 > one switch has > been put into the wrong management VLAN then the ping will=20 > obviously fail. >=20 > If you want to use a separate VLAN for GMPLS control plane=20 > traffic, then > you either need to (i) define a reserved GMPLS signalling=20 > VLAN ID, or (ii) > leave it up to the operator to choose a VLAN ID. If you leave=20 > it up to the > operator to decide what VLAN ID to use, then you need to=20 > consider what the > implications in the inter-operator case will be, i.e. what=20 > VLAN ID range > should be used for control plane traffic and what VLAN ID=20 > range should be > used for user plane traffic. >=20 > I'm not saying that defining how GMPLS control plane packets=20 > are forwarded > in the Ethernet data plane will be an issue, simply that this=20 > is something > that must be addressed to develop interoperable solutions. >=20 > Regards, > Richard >=20 > > -----Original Message----- > > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On > > Behalf Of Heiles Juergen > > Sent: 22 July 2005 14:17 > > To: Loa Andersson; P=E4r Mattsson > > Cc: ccamp@ops.ietf.org > > Subject: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > GMPLS for SONET/SDH has no impact on the SONET/SDH data > > plane. It introduces a control plane for connection setup > > instead of a network management based connection setup. GMPLS > > for Ethernet can work in the same way. Ethernet as such is > > cl, but VLANs are setup and GMPLS can provide a control plane > > for VLAN setup. The Ethernet data plane is not changed and > > MAC address learning and forwarding is still be done within a > > VLAN. So no change to the Ethernet data plane. No GMPLS > > traffic is introduced, it is still VLAN traffic. > > Introducing a dedicated GMPLS label for Ethernet (GMPLS > > traffic) goes beyond GMPLS as a control plane technique. > > > > Regards > > > > Juergen > > > > > > > > > > > -----Original Message----- > > > From: owner-ccamp@ops.ietf.org > > > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson > > > Sent: Friday, July 22, 2005 12:05 PM > > > To: P=E4r Mattsson > > > Cc: ccamp@ops.ietf.org > > > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > > > > Per and Dimitri, > > > > > > I would like to come down stronger than that, for me it is > > > a very strong requirement that the same switch can handle > > > both VLANs and GMPLs trafic correctly. I can't dsee how that > > > could be done if using the VLAN tpid to indicate GMPLS > > > traffic. > > > > > > /Loa > > > > > > P=E4r Mattsson wrote: > > > >>hi par, one of the possibilities that has been considered > > > to cope with > > > >>this requirement is to use a dedicated TPID for the > > Ethernet labeled > > > >>frames; this would allow differentiated processing with > > non-labeled > > > >>framesthanks. > > > > > > > > > > > > That seems to make more sence. If that frame is to be sized > > > like a 802.1q > > > > frame. There is not that much space left to a label. Or is > > > the demand to > > > > use jumboframes ? > > > > Has there been any discussion on labelstacking, and mainly > > > where to place > > > > the information? > > > > > > > > Regards. > > > > Per > > > > > > > > > > > > > > > > > > > > > --=20 > > > Loa Andersson > > > > > > Principal Networking Architect > > > Acreo AB phone: +46 8 632 77 14 > > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > > Kista, Sweden email: loa.andersson@acreo.se > > > loa@pi.se > > > > > > > >=20 >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 23:23:08 +0000 To: <neil.2.harrison@bt.com> Cc: <ibryskin@movaz.com>, <Ben.Mack-Crane@tellabs.com>, <adrian@olddog.co.uk>, <per@defero.se>, <loa@pi.se>, <ccamp@ops.ietf.org> From: Dimitri.Papadimitriou@alcatel.be Subject: RE: Frameformat in a l2cs gmpls rnvironment. MIME-Version: 1.0 Date: Sat, 23 Jul 2005 01:21:55 +0200 Message-ID: <OF495D5339.8DF644B1-ONC1257046.0080594D-C1257046.00805993@netfr.alcatel.fr> Content-type: text/plain; charset=us-ascii [snip] > Can I make a plea NOT to screw-up ethernet by attempting to replicate MPLS by definition of "Ethernet", this is an explicit non-goal, also the document does not intend to replicate RFC 3031 > ...in any case the ethernet data-plane belongs to IEEE. i don't think the document under discussion states the opposite - all the contrary section 7.1 states IEEE "should be consulted on the scenarios and framework proposed in this document and any solutions developed in the IETF context should be reviewed by the appropriate IEEE working groups to ensure the solution is not harmful to 802.1 networks." Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 23:08:50 +0000 To: <neil.2.harrison@bt.com> Cc: <Ben.Mack-Crane@tellabs.com>, <ibryskin@movaz.com>, <Dimitri.Papadimitriou@alcatel.be>, <adrian@olddog.co.uk>, <per@defero.se>, <loa@pi.se>, <ccamp@ops.ietf.org> MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Sat, 23 Jul 2005 01:03:11 +0200 Message-ID: <OF49D2E5D0.C4DE42CE-ONC1257046.007EA136-C1257046.007EA280@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPm5laWwgLSA8L0ZPTlQ+PC9QPjxQPjxG T05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj50aGUgaW50cm9kdWN0aW9uIG9mIHRoZSBkb2N1 bWVudCB1bmRlciBkaXNjdXNzaW9uIHByb3ZpZGVzIGEgY2xlYXIgZGVzY3JpcHRpb24gb24gdGhl IHBvc2l0aW9uaW5nIG9mIHRoaXMgd29yayB3cnQgdG8gUFcvTVBMUyA8L0ZPTlQ+PC9QPjxQPjxG T05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5zbyBpIGRvIG5vIG5lY2Vzc2FyaWx5IGZvbGxv dyB5b3VyIGFyZ3VtZW50cyAoaWYgYW55KSB3aGVuIHlvdSBhcmUgZGlncmVzc2luZyBvbiBQVy9N UExTIDwvRk9OVD48QlI+PEJSPjxGT05UIFNJWkU9Mj48Qj4mbHQ7bmVpbC4yLmhhcnJpc29uQGJ0 LmNvbSZndDs8L0I+PC9GT05UPjxCUj48Rk9OVCBTSVpFPTI+MDcvMjIvMjAwNSAyMjoyMjwvRk9O VD48QlI+PEJSPiA8Rk9OVCBTSVpFPTI+VG86PC9GT05UPiA8Rk9OVCBTSVpFPTI+Jmx0O0Jlbi5N YWNrLUNyYW5lQHRlbGxhYnMuY29tJmd0OywgJmx0O2licnlza2luQG1vdmF6LmNvbSZndDssIERp bWl0cmkgUEFQQURJTUlUUklPVS9CRS9BTENBVEVMQEFMQ0FURUw8L0ZPTlQ+PEJSPiA8Rk9OVCBT SVpFPTI+Y2M6PC9GT05UPiA8Rk9OVCBTSVpFPTI+Jmx0O2FkcmlhbkBvbGRkb2cuY28udWsmZ3Q7 LCAmbHQ7cGVyQGRlZmVyby5zZSZndDssICZsdDtsb2FAcGkuc2UmZ3Q7LCAmbHQ7Y2NhbXBAb3Bz LmlldGYub3JnJmd0OzwvRk9OVD48QlI+IDxGT05UIFNJWkU9Mj5iY2M6PC9GT05UPiA8QlI+IDxG T05UIFNJWkU9Mj5TdWJqZWN0OjwvRk9OVD4gPEZPTlQgU0laRT0yPlJFOiBGcmFtZWZvcm1hdCBp biBhIGwyY3MgZ21wbHMgcm52aXJvbm1lbnQuPC9GT05UPjxCUj4gPEJSPjxCUj48L1A+PFA+QW5k IG1lIHRvbyBCZW4vSWdvci4mbmJzcDsgQnV0IGdvIGxvb2sgYXQgTVBMUyB3aGVuIGZvbGtzIERJ RCBkZWZpbmUgYSBkYXRhLXBsYW5lIGhlcmUuLi4uZG9uJ3QgYW55b25lIGRhcmUgdGVsbCBtZSAn dGhpcyBpcyBnb29kJyAob3IgdGhlIFBXIHN0dWZmIGl0IHNwYXduZWQpIGJlY2F1c2UgSSBrbm93 IG90aGVyd2lzZS48QlI+PEJSPnJlcWFyZHMsIE5laWw8QlI+LS0tLS1PcmlnaW5hbCBNZXNzYWdl LS0tLS08QlI+PEI+RnJvbTo8L0I+IG93bmVyLWNjYW1wQG9wcy5pZXRmLm9yZyBbbWFpbHRvOm93 bmVyLWNjYW1wQG9wcy5pZXRmLm9yZ10gPEI+T24gQmVoYWxmIE9mIDwvQj5NYWNrLUNyYW5lLCBU LiBCZW5qYW1pbjxCUj48Qj5TZW50OjwvQj4gMjIgSnVseSAyMDA1IDE5OjI1PEJSPjxCPlRvOjwv Qj4gSWdvciBCcnlza2luOyBEaW1pdHJpLlBhcGFkaW1pdHJpb3VAYWxjYXRlbC5iZTxCUj48Qj5D Yzo8L0I+IEFkcmlhbiBGYXJyZWw7IFAmYXVtbDtyIE1hdHRzc29uOyBMb2EgQW5kZXJzc29uOyBj Y2FtcEBvcHMuaWV0Zi5vcmc8QlI+PEI+U3ViamVjdDo8L0I+IFJFOiBGcmFtZWZvcm1hdCBpbiBh IGwyY3MgZ21wbHMgcm52aXJvbm1lbnQuPEJSPjxCUj48QlI+SWdvciw8QlI+PEJSPkksIHRvbywg YW0gYmFmZmxlZCBieSB0aGUgbm90aW9uIG9mIHJlcXVpcmVtZW50cyBmb3IgY29udHJvbCBvZiBh biB1bnNwZWNpZmllZCBkYXRhIHBsYW5lLjxCUj48QlI+UmVnYXJkcyw8QlI+QmVuPEJSPjxIUiBX SURUSD0iOTklIiBTSVpFPSI0IiBBTElHTj0ibGVmdCI+PEZPTlQgQ09MT1I9QkxBQ0s+PEJSPjxC PkZyb206PC9CPiBJZ29yIEJyeXNraW4gW21haWx0bzppYnJ5c2tpbkBtb3Zhei5jb21dIDxCUj48 Qj5TZW50OjwvQj4gRnJpZGF5LCBKdWx5IDIyLCAyMDA1IDEyOjIwIFBNPEJSPjxCPlRvOjwvQj4g RGltaXRyaS5QYXBhZGltaXRyaW91QGFsY2F0ZWwuYmU8QlI+PEI+Q2M6PC9CPiBNYWNrLUNyYW5l LCBULiBCZW5qYW1pbjsgQWRyaWFuIEZhcnJlbDsgUCZhdW1sO3IgTWF0dHNzb247IExvYSBBbmRl cnNzb247IGNjYW1wQG9wcy5pZXRmLm9yZzxCUj48Qj5TdWJqZWN0OjwvQj4gUmU6IEZyYW1lZm9y bWF0IGluIGEgbDJjcyBnbXBscyBybnZpcm9ubWVudC48QlI+PEJSPkRpbWl0cmksPEJSPjxCUj5U aGUgcXVlc3Rpb24gaXMgc2ltcGxlOiBpcyBpdCBwb3NzaWJsZSB0b2RheSB0byBzdGF0aWNhbGx5 IHByb3Zpc2lvbiBMMkxTUHMgdGhhdCBjb3VsZCwgc2F5LCZuYnNwO3N1cHBvcnQgZTJlIFFvUz8g SWYgbm90LCBtYXkgYmUgaXQgaXMgdG8gZWFybHkgdG8gZGlzY3VzcyBhc3BlY3RzIG9mIGR5bmFt aWMgcHJvdmlzaW9uaW5nIG9mIHN1Y2ggTFNQcz8gSXMgaXQgcG9zc2libGUvcmVhc29uYWJsZSwg aW4geW91ciBvcGluaW9uLCB0byAmcXVvdDs8L0ZPTlQ+PEZPTlQgU0laRT00PmRldGFpbCBob3cg Zm9yd2FyZGluZyBpbmZvcm1hdGlvbiBjYW4gYmUgZXhjaGFuZ2VkIHZpYSB0aGUgY29udHJvbCBw bGFuZSAoYW5kIHRoZW4gaW5zdGFsbGVkKSZxdW90OyB3aXRob3V0Jm5ic3A7aGF2aW5nIGluIG1p bmQgJnF1b3Q7YSBzcGVjaWZpYyBmb3J3YXJkaW5nIGJlaGF2aW9yJnF1b3Q7PzwvRk9OVD48QlI+ PEZPTlQgU0laRT00PkhvdyBkbyB5b3Uga25vdyB3aGljaCBmb3J3YXJkaW5nIGluZm9ybWF0aW9u IGlzIG5lZWRlZCBmb3IgdGhlIGZvcndhcmRpbmcgbm9ib2R5IGhhcyBkZWZpbmVkIHlldD88L0ZP TlQ+PEJSPjxGT05UIFNJWkU9ND5UaGUgb3RoZXIgcXVlc3Rpb24gaXMmbmJzcDtkbyB3ZSBuZWVk IGF0IGFsbCBlMmUgUW9TLCByb3V0ZSBjb250cm9sLCBmYXN0IHJlY292ZXJ5PyBTb3VuZHMgbGlr ZSBleGNpdGluZyBpZGVhLCBidXQgZG9lcyBhbGwgdGhhdCBjb21tZXJjaWFsbHkgbWFrZSBzZW5z ZT8gU28sIEkgZ3Vlc3MsIHdlIG5lZWQgdHdvIHRoaW5ncyZuYnNwO2JlZm9yZSB3ZSBjYW4mbmJz cDttb3ZlIGZvcndhcmQ6PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+YSkgdmFsaWRhdGlvbiBvZiB0 aGUgaWRlYSBieSB0aGUgcHJvdmlkZXJzOzwvRk9OVD48QlI+PEZPTlQgU0laRT00PmIpIGRlZmlu aXRpb24gb2YgdGhlIGRhdGEgcGxhbmUgYmVoYXZpb3I7PC9GT05UPjxCUj48QlI+PEZPTlQgU0la RT00Pklnb3I8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4tLS0tLSBPcmlnaW5hbCBNZXNzYWdlIC0t LS0tIDwvRk9OVD48QlI+PEZPTlQgU0laRT00PjxCPkZyb206PC9CPiA8L0ZPTlQ+PEZPTlQgU0la RT00IENPTE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86RGltaXRyaS5QYXBhZGltaXRyaW91QGFs Y2F0ZWwuYmU+RGltaXRyaS5QYXBhZGltaXRyaW91QGFsY2F0ZWwuYmU8L0E+PC9VPjwvRk9OVD48 Rk9OVCBTSVpFPTQgQ09MT1I9QkxBQ0s+IDwvRk9OVD48QlI+PEZPTlQgU0laRT00PjxCPlRvOjwv Qj4gPC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOmlicnlz a2luQG1vdmF6LmNvbT5JZ29yIEJyeXNraW48L0E+PC9VPjwvRk9OVD48Rk9OVCBTSVpFPTQgQ09M T1I9QkxBQ0s+IDwvRk9OVD48QlI+PEZPTlQgU0laRT00PjxCPkNjOjwvQj4gPC9GT05UPjxGT05U IFNJWkU9NCBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOkJlbi5NYWNrLUNyYW5lQHRlbGxh YnMuY29tPk1hY2stQ3JhbmUsIFQuIEJlbmphbWluPC9BPjwvVT48L0ZPTlQ+PEZPTlQgU0laRT00 IENPTE9SPUJMQUNLPiA7IDwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxVRT48VT48QSBIUkVG PW1haWx0bzphZHJpYW5Ab2xkZG9nLmNvLnVrPkFkcmlhbiBGYXJyZWw8L0E+PC9VPjwvRk9OVD48 Rk9OVCBTSVpFPTQgQ09MT1I9QkxBQ0s+IDsgPC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTFVF PjxVPjxBIEhSRUY9bWFpbHRvOnBlckBkZWZlcm8uc2U+UCZhdW1sO3IgTWF0dHNzb248L0E+PC9V PjwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxBQ0s+IDsgPC9GT05UPjxGT05UIFNJWkU9NCBD T0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOmxvYUBwaS5zZT5Mb2EgQW5kZXJzc29uPC9BPjwv VT48L0ZPTlQ+PEZPTlQgU0laRT00IENPTE9SPUJMQUNLPiA7IDwvRk9OVD48Rk9OVCBTSVpFPTQg Q09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0bzpjY2FtcEBvcHMuaWV0Zi5vcmc+Y2NhbXBAb3Bz LmlldGYub3JnPC9BPjwvVT48L0ZPTlQ+PEZPTlQgU0laRT00IENPTE9SPUJMQUNLPiA8L0ZPTlQ+ PEJSPjxGT05UIFNJWkU9ND48Qj5TZW50OjwvQj4gRnJpZGF5LCBKdWx5IDIyLCAyMDA1IDEyOjIx IFBNPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+PEI+U3ViamVjdDo8L0I+IFJlOiBGcmFtZWZvcm1h dCBpbiBhIGwyY3MgZ21wbHMgcm52aXJvbm1lbnQuPC9GT05UPjxCUj48QlI+PEJSPjxGT05UIFNJ WkU9ND5pZ29yIC0gaSBhbSBub3Qgc3VyZSB3aGljaCBwb2ludCB0aGV5IGV4YWN0bHkgaGF2ZSAt IHRoZSBwdXJwb3NlIG9mIHRoaXMgZG9jdW1lbnQgaXMgdG8gZGV0YWlsIGhvdyBmb3J3YXJkaW5n IGluZm9ybWF0aW9uIGNhbiBiZSBleGNoYW5nZWQgdmlhIHRoZSBjb250cm9sIHBsYW5lIChhbmQg dGhlbiBpbnN0YWxsZWQpIG5vdCB0byBkZWZpbmUgYSBzcGVjaWZpYyBmb3J3YXJkaW5nIGJlaGF2 aW91cjwvRk9OVD48QlI+PEJSPjxGT05UIFNJWkU9ND50aGFua3MsPC9GT05UPjxCUj48QlI+PEZP TlQgU0laRT00Pi0gZGltaXRyaS48L0ZPTlQ+PEJSPjxCUj48Qj4mcXVvdDtJZ29yIEJyeXNraW4m cXVvdDsgJmx0OzwvQj48Rk9OVCBDT0xPUj1CTFVFPjxCPjxVPjxBIEhSRUY9bWFpbHRvOmlicnlz a2luQG1vdmF6LmNvbT5pYnJ5c2tpbkBtb3Zhei5jb208L0E+PC9VPjwvQj48L0ZPTlQ+PEZPTlQg Q09MT1I9QkxBQ0s+PEI+Jmd0OzwvQj48QlI+U2VudCBieTogPC9GT05UPjxGT05UIENPTE9SPUJM VUU+PFU+PEEgSFJFRj1tYWlsdG86b3duZXItY2NhbXBAb3BzLmlldGYub3JnPm93bmVyLWNjYW1w QG9wcy5pZXRmLm9yZzwvQT48L1U+PC9GT05UPjxGT05UIENPTE9SPUJMQUNLPjxCUj4wNy8yMi8y MDA1IDExOjEwIEFTVDxCUj48QlI+VG86PC9GT05UPjxGT05UIFNJWkU9ND4gPC9GT05UPiZxdW90 O01hY2stQ3JhbmUsIFQuIEJlbmphbWluJnF1b3Q7ICZsdDs8Rk9OVCBDT0xPUj1CTFVFPjxVPjxB IEhSRUY9bWFpbHRvOkJlbi5NYWNrLUNyYW5lQHRlbGxhYnMuY29tPkJlbi5NYWNrLUNyYW5lQHRl bGxhYnMuY29tPC9BPjwvVT48L0ZPTlQ+PEZPTlQgQ09MT1I9QkxBQ0s+Jmd0OywgJnF1b3Q7QWRy aWFuIEZhcnJlbCZxdW90OyAmbHQ7PC9GT05UPjxGT05UIENPTE9SPUJMVUU+PFU+PEEgSFJFRj1t YWlsdG86YWRyaWFuQG9sZGRvZy5jby51az5hZHJpYW5Ab2xkZG9nLmNvLnVrPC9BPjwvVT48L0ZP TlQ+PEZPTlQgQ09MT1I9QkxBQ0s+Jmd0OywgUCZhdW1sO3IgTWF0dHNzb24gJmx0OzwvRk9OVD48 Rk9OVCBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOnBlckBkZWZlcm8uc2U+cGVyQGRlZmVy by5zZTwvQT48L1U+PC9GT05UPjxGT05UIENPTE9SPUJMQUNLPiZndDssICZxdW90O0xvYSBBbmRl cnNzb24mcXVvdDsgJmx0OzwvRk9OVD48Rk9OVCBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRv OmxvYUBwaS5zZT5sb2FAcGkuc2U8L0E+PC9VPjwvRk9OVD48Rk9OVCBDT0xPUj1CTEFDSz4mZ3Q7 PEJSPmNjOjwvRk9OVD48Rk9OVCBTSVpFPTQ+IDwvRk9OVD4mbHQ7PEZPTlQgQ09MT1I9QkxVRT48 VT48QSBIUkVGPW1haWx0bzpjY2FtcEBvcHMuaWV0Zi5vcmc+Y2NhbXBAb3BzLmlldGYub3JnPC9B PjwvVT48L0ZPTlQ+PEZPTlQgQ09MT1I9QkxBQ0s+Jmd0OzxCUj5iY2M6PC9GT05UPjxGT05UIFNJ WkU9ND4gPC9GT05UPjxCUj5TdWJqZWN0OjxGT05UIFNJWkU9ND4gPC9GT05UPlJlOiBGcmFtZWZv cm1hdCBpbiBhIGwyY3MgZ21wbHMgcm52aXJvbm1lbnQuPEJSPjxCUj48QlI+PEZPTlQgU0laRT00 PkhpLDwvRk9OVD48QlI+PEJSPjxGT05UIFNJWkU9ND5JIHRoaW5rIEJlbiBhbmQgSnVlcmdlbiBo YXZlIGEgcG9pbnQgaGVyZS4gVGhlcmUgaXMgbm90aGluZyB0aGF0IGNvdWxkIGJlIGR5bmFtaWNh bGx5IHByb3Zpc2lvbmVkIHRoYXQgY291bGQgbm90IGJlIGFsc28gcHJvdmlzaW9uZWQgYnkgbWFu YWdlbWVudC4gSW4gb3RoZXIgd29yZHMgZGF0YSBwbGFuZSBpcyB0aGUgS2luZyBoZXJlLiBXZSBu ZWVkIGRhdGEgcGxhbmUgc3RhbmRhcmQocykgb24gaG93IHdlIGVuY29kZSBhIGxhYmVsLCB3aGV0 aGVyIHdlIG5lZWQgYSBsYWJlbCBzdGFjayBvciBub3QsIGhvdyB0aGUgbGFiZWxlZCB0cmFmZmlj IGlzIHN1cHBvc2VkIHRvIGJlJm5ic3A7dHJlYXRlZCwgaG93IGxhYmVsZWQgdHJhZmZpYyBjby1l eGlzdCB3aXRoIG5vbi1sYWJlbCB0cmFmZmljLCBldGMuIFRoaXMgaXMgc29tZXRoaW5nIHRoYXQg bm90IGZvciBDQ0FNUCB0byBkZWZpbmUuIFRha2UgZm9yIGV4YW1wbGUgVERNIG5ldHdvcmtzLiBH TVBMUyBvbmx5IHByb3ZpZGVzIGEgd2F5IHRvIGR5bmFtaWNhbGx5IHByb3Zpc2lvbiBHLjcwNyBu ZXR3b3Jrcy4gSGVuY2UgdGhlcmUgaXMgYSBuZWVkIGluIHBhcmFsbGVsIHN0YW5kYXJkaXphdGlv biBhY3Rpdml0aWVzIGluIElUVS4gT25seSBhZnRlciB0aGF0IHdlIGNhbiBkaXNjdXNzIGhvdyBh bGwgdGhhdCBjb3VsZCBiZSBkeW5hbWljYWxseSBwcm92aXNpb25lZCwgdGhhdCBpcyB0aGUgYXNw ZWN0cyBvZiBjb250cm9sIHBsYW5lLiA8L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTQ+SWdvcjwv Rk9OVD48QlI+PEZPTlQgU0laRT01Pi0tLS0tIE9yaWdpbmFsIE1lc3NhZ2UgLS0tLS0gPC9GT05U PjxCUj48Rk9OVCBTSVpFPTU+PEI+RnJvbTo8L0I+IDwvRk9OVD48Rk9OVCBTSVpFPTUgQ09MT1I9 QkxVRT48VT48QSBIUkVGPW1haWx0bzpCZW4uTWFjay1DcmFuZUB0ZWxsYWJzLmNvbT5NYWNrLUNy YW5lLCBULiBCZW5qYW1pbjwvQT48L1U+PC9GT05UPjxGT05UIFNJWkU9NSBDT0xPUj1CTEFDSz4g PC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+PEI+VG86PC9CPiA8L0ZPTlQ+PEZPTlQgU0laRT01IENP TE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86YWRyaWFuQG9sZGRvZy5jby51az5BZHJpYW4gRmFy cmVsPC9BPjwvVT48L0ZPTlQ+PEZPTlQgU0laRT01IENPTE9SPUJMQUNLPiA7IDwvRk9OVD48Rk9O VCBTSVpFPTUgQ09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0bzpwZXJAZGVmZXJvLnNlPlAmYXVt bDtyIE1hdHRzc29uPC9BPjwvVT48L0ZPTlQ+PEZPTlQgU0laRT01IENPTE9SPUJMQUNLPiA7IDwv Rk9OVD48Rk9OVCBTSVpFPTUgQ09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0bzpsb2FAcGkuc2U+ TG9hIEFuZGVyc3NvbjwvQT48L1U+PC9GT05UPjxGT05UIFNJWkU9NSBDT0xPUj1CTEFDSz4gPC9G T05UPjxCUj48Rk9OVCBTSVpFPTU+PEI+Q2M6PC9CPiA8L0ZPTlQ+PEZPTlQgU0laRT01IENPTE9S PUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86Y2NhbXBAb3BzLmlldGYub3JnPmNjYW1wQG9wcy5pZXRm Lm9yZzwvQT48L1U+PC9GT05UPjxGT05UIFNJWkU9NSBDT0xPUj1CTEFDSz4gPC9GT05UPjxCUj48 Rk9OVCBTSVpFPTU+PEI+U2VudDo8L0I+IEZyaWRheSwgSnVseSAyMiwgMjAwNSAxMDoyMyBBTTwv Rk9OVD48QlI+PEZPTlQgU0laRT01PjxCPlN1YmplY3Q6PC9CPiBSRTogRnJhbWVmb3JtYXQgaW4g YSBsMmNzIGdtcGxzIHJudmlyb25tZW50LjwvRk9OVD48QlI+PEJSPjxGT05UIFNJWkU9ND5IaSw8 L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTQ+SSB0aGluayBKdWVyZ2VuIGhhcyByYWlzZWQgYW4g aW1wb3J0YW50IHF1ZXN0aW9uLiZuYnNwOyBIb3cgZnJhbWVzIGFyZSBsYWJlbGVkIChhbmQgdGhl IHJlbGF0ZWQgZGF0YSBwbGFuZSBmb3J3YXJkaW5nIGJlaGF2aW9yKSBpcyBub3QgZGVmaW5lZCBi eSB0aGUgY29udHJvbCBwbGFuZS4mbmJzcDsgVGhlIGNvbnRyb2wgcGxhbmUgc2VydmVzIHRvIHBy b3Zpc2lvbiB0aGUgZGF0YSBwbGFuZSwgbm90IGRlZmluZSBpdC4mbmJzcDsgSW4gdGhlIGZyYW1l d29yayBkcmFmdCBpdCBpcyBub3QgY2xlYXIgd2hhdCBkYXRhIHBsYW5lIHN0YW5kYXJkcyBhcmUg Y292ZXJlZCBieSB0aGUgc3RhdGVkIGNvbnRyb2wgcGxhbmUgcmVxdWlyZW1lbnRzLiZuYnNwOyBT b21lIHJlZmVyZW5jZXMgc2hvdWxkIGJlIHN1cHBsaWVkLiZuYnNwOyBJbiBhbnkgY2FzZSwgdGhl IGxhYmVsaW5nIGFuZCBmb3J3YXJkaW5nIGJlaGF2aW9yIHNob3VsZCBiZSBkZWZpbmVkJm5ic3A7 YnkgdGhlc2UgcmVmZXJlbmNlZCBzdGFuZGFyZHMsIG5vdCZuYnNwO2J5IEdNUExTLjwvRk9OVD48 QlI+PEJSPjxGT05UIFNJWkU9ND4oSSBhbSBhc3N1bWluZyBkZWZpbml0aW9uIG9mIG5ldyBkYXRh IHBsYW5lIHN0YW5kYXJkcyBpcyBvdXQgb2Ygc2NvcGUgZm9yIENDQU1QLik8L0ZPTlQ+PEJSPjxC Uj48Rk9OVCBTSVpFPTQ+UmVnYXJkcyw8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND5CZW48L0ZPTlQ+ PEJSPjxIUiBXSURUSD0iOTklIiBTSVpFPSI4IiBBTElHTj0ibGVmdCI+PEZPTlQgQ09MT1I9QkxB Q0s+PEJSPjwvRk9OVD48Rk9OVCBTSVpFPTQ+PEI+RnJvbTo8L0I+IG93bmVyLWNjYW1wQG9wcy5p ZXRmLm9yZyBbbWFpbHRvOm93bmVyLWNjYW1wQG9wcy5pZXRmLm9yZ10gPEI+T24gQmVoYWxmIE9m IDwvQj5BZHJpYW4gRmFycmVsPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+PEI+U2VudDo8L0I+IEZy aWRheSwgSnVseSAyMiwgMjAwNSA4OjM3IEFNPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+PEI+VG86 PC9CPiBQJmF1bWw7ciBNYXR0c3NvbjsgTG9hIEFuZGVyc3NvbjwvRk9OVD48QlI+PEZPTlQgU0la RT00PjxCPkNjOjwvQj4gY2NhbXBAb3BzLmlldGYub3JnPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+ PEI+U3ViamVjdDo8L0I+IFJlOiBGcmFtZWZvcm1hdCBpbiBhIGwyY3MgZ21wbHMgcm52aXJvbm1l bnQuPC9GT05UPjxCUj48QlI+PEZPTlQgU0laRT00PkhpLDwvRk9OVD48QlI+PEJSPjxGT05UIFNJ WkU9ND5JdCBpcyBhbHdheXMgcmVmcmVzaGluZyBob3cgZW5naW5lZXJzIGp1bXAgc3RyYWlnaHQg dG8gYSBkaXNjdXNzaW9uIG9mIHRoZSBzb2x1dGlvbiA6LSk8L0ZPTlQ+PEJSPjxCUj48Rk9OVCBT SVpFPTQ+UGVyaGFwcyB3ZSBjYW4gYXNzdW1lIHRoYXQgdGhlIGZyYW1ld29yayBkcmFmdCBpcyBh IGdvb2QgcmVwcmVzZW50YXRpb24gb2YgdGhlIHByb2JsZW0gc3BhY2UsIGFuZCB0aGF0IHdlIGFy ZSByZWFkeSB0byBzdGFydCBkaXNjdXNzaW5nIHRoZSBzb2x1dGlvbnMuPC9GT05UPjxCUj48QlI+ PEZPTlQgU0laRT00Pk9uZSAocGVyaGFwcyB0aGUgb25seT8pIHNpZ25pZmljYW50IHF1ZXN0aW9u IHRvIGFuc3dlciBpcyBob3cgdGhlIGZyYW1lcyB3aWxsIGJlIGxhYmVsbGVkLiBUaGlzIHF1ZXN0 aW9uIGlzIG9uZSB3ZSBtdXN0IGNvbWUgdG8gYXMgc29vbiBhcyB3ZSBhcmUgY29uZmlkZW50IHRo YXQgdGhlIHJlcXVpcmVtZW50cyBuZWVkIHRvIGJlIGFkZHJlc3NlZCBhdCBhbGwuPC9GT05UPjxC Uj48QlI+PEZPTlQgU0laRT00PkFzIGhhcyBiZWVuIHBvaW50ZWQgb3V0LCB0aGVyZSBhcmUgc2V2 ZXJhbCBwb3NzaWJpbGl0aWVzIGFuZCB0byBwaWNrIGZyb20gdGhlc2Ugd2UgbmVlZCB0byB1bmRl cnN0YW5kOjwvRk9OVD48QlI+PEZPTlQgU0laRT00PjEuIERvIHdlIG5lZWQgdG8gc3VwcG9ydCBl eHBsaWNpdCBsYWJlbCBzdGFja2luZz88L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mbmJzcDsmbmJz cDsgTm90ZSB0aGF0IHRoaXMgaXMgbm90IHN1cHBvcnRlZCBpbiBURE0sIExTQyBvciBGU0MuPC9G T05UPjxCUj48Rk9OVCBTSVpFPTQ+Mi4gRG8gd2UgbmVlZCB0byBiZSBhYmxlIHRvIGNvbnRyb2wg KHBlcmhhcHMgdGhyb3VnaCBhbjwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZuYnNwOyZuYnNwOyZu YnNwO2V4dGVybmFsIHNpZ25hbGluZyBjb250cm9sbGVyKSBleGlzdGluZyBoYXJkd2FyZSBhbmQg PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jm5ic3A7Jm5ic3A7IGluc3RhbGwgTFNQcyB0aHJvdWdo IGV4aXN0aW5nIG5ldHdvcmtzPzwvRk9OVD48QlI+PEZPTlQgU0laRT00PjMuIERvIHdlIG5lZWQg dG8gc3VwcG9ydCBleGlzdGluZyBmdW5jdGlvbiBzaW11bHRhbmVvdXMgdG88L0ZPTlQ+PEJSPjxG T05UIFNJWkU9ND4mbmJzcDsmbmJzcDsgdGhlIHN1cHBvcnQgb2YgR01QTFMgTDIgTFNQcz88L0ZP TlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTQ+SSB0aGluayB0aGF0IGZyb20gYSBjaGFpcidzIHBlcnNw ZWN0aXZlIEkgY2FuIGdpdmUgc29tZSBsaW1pdGVkIGd1aWRhbmNlLjwvRk9OVD48QlI+PEJSPjxG T05UIFNJWkU9ND5BLiBUaGVzZSBxdWVzdGlvbnMgbXVzdCBiZSByYWlzZWQgYW5kIGFuc3dlcmVk IGluIHRoZSBmcmFtZXdvcmsgSS1EPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Qi4gVGhlIGFuc3dl ciB0byBxdWVzdGlvbiAzIGlzICZxdW90O3llcyZxdW90Oy4gVGhpcyBtZWFucyB0aGF0IHRoZSB1 c2Ugb2YgbGFiZWxzPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jm5ic3A7Jm5ic3A7IG11c3Qgbm90 IHNpZ25pZmljYW50bHkgZGVwbGV0ZSBhbnkgbmFtZXNwYWNlIHVzZWQgdG8gc3VwcG9ydCBvdGhl cjwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZuYnNwOyZuYnNwOyBmdW5jdGlvbi48L0ZPTlQ+PEJS PjxGT05UIFNJWkU9ND5DLiBDQ0FNUCBpcyBjaGFydGVyZWQgdG8gbG9vayBhdCB0aGUgY29udHJv bCBvZiB0cmFuc3BvcnQgbmV0d29ya3MuIFRoaXM8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mbmJz cDsmbmJzcDsmbmJzcDtpbmNsdWRlcyBNZXRybywgYnV0IEkgYW0gdW5zdXJlIGFib3V0IENhbXB1 cy4gSXQgZG9lcyBub3QmbmJzcDtjb3ZlciA8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mbmJzcDsm bmJzcDsgc2lnbmFsaW5nIHRvIHRoZSBkZXNrdG9wLjwvRk9OVD48QlI+PEJSPjxGT05UIFNJWkU9 ND5UaGFua3MsPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+QWRyaWFuPC9GT05UPjxCUj48QlI+PEZP TlQgU0laRT00Pi0tLS0tIE9yaWdpbmFsIE1lc3NhZ2UgLS0tLS0gPC9GT05UPjxCUj48Rk9OVCBT SVpFPTQ+RnJvbTogJnF1b3Q7UCZhdW1sO3IgTWF0dHNzb24mcXVvdDsgJmx0OzwvRk9OVD48Rk9O VCBTSVpFPTQgQ09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0bzpwZXJAZGVmZXJvLnNlPnBlckBk ZWZlcm8uc2U8L0E+PC9VPjwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxBQ0s+Jmd0OzwvRk9O VD48QlI+PEZPTlQgU0laRT00PlRvOiAmcXVvdDtMb2EgQW5kZXJzc29uJnF1b3Q7ICZsdDs8L0ZP TlQ+PEZPTlQgU0laRT00IENPTE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86bG9hQHBpLnNlPmxv YUBwaS5zZTwvQT48L1U+PC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTEFDSz4mZ3Q7PC9GT05U PjxCUj48Rk9OVCBTSVpFPTQ+Q2M6ICZsdDs8L0ZPTlQ+PEZPTlQgU0laRT00IENPTE9SPUJMVUU+ PFU+PEEgSFJFRj1tYWlsdG86Y2NhbXBAb3BzLmlldGYub3JnPmNjYW1wQG9wcy5pZXRmLm9yZzwv QT48L1U+PC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTEFDSz4mZ3Q7PC9GT05UPjxCUj48Rk9O VCBTSVpFPTQ+U2VudDogRnJpZGF5LCBKdWx5IDIyLCAyMDA1IDExOjQxIEFNPC9GT05UPjxCUj48 Rk9OVCBTSVpFPTQ+U3ViamVjdDogUmU6IEZyYW1lZm9ybWF0IGluIGEgbDJjcyBnbXBscyBybnZp cm9ubWVudC48L0ZPTlQ+PEJSPjxCUj48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyBQZXIgYW5k IERpbWl0cmksPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7PC9GT05UPjxCUj48Rk9O VCBTSVpFPTQ+Jmd0OyAmZ3Q7IEkgd291bGQgbGlrZSB0byBjb21lIGRvd24gc3Ryb25nZXIgdGhh biB0aGF0LCBmb3IgbWUgaXQgaXM8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsgYSB2 ZXJ5IHN0cm9uZyByZXF1aXJlbWVudCB0aGF0IHRoZSBzYW1lIHN3aXRjaCBjYW4gaGFuZGxlPC9G T05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7IGJvdGggVkxBTnMgYW5kIEdNUExzIHRyYWZp YyBjb3JyZWN0bHkuIEkgY2FuJ3QgZHNlZSBob3cgdGhhdDwvRk9OVD48QlI+PEZPTlQgU0laRT00 PiZndDsgJmd0OyBjb3VsZCBiZSBkb25lIGlmIHVzaW5nIHRoZSBWTEFOIHRwaWQgdG8gaW5kaWNh dGUgR01QTFM8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsgdHJhZmZpYy48L0ZPTlQ+ PEJSPjxGT05UIFNJWkU9ND4mZ3Q7IDwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgSWYgeW91 IGV2ZXIgd2FudCB0aGF0IHNhbWUgc3dpdGNoIHRvIGhhbmRsZSB0cmFmZmljIGZvciBhIGRpcmVj dGx5PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyBjb25uZWN0ZWQmbmJzcDsgaG9zdCAobm90 IHRvIHVuY29tbWVuKSB5b3Ugd291bGQgd2FudCB0aGF0IHRvIHVzZSBub3JtYWw8L0ZPTlQ+PEJS PjxGT05UIFNJWkU9ND4mZ3Q7IGV0aGVybmV0IG1hY2FkZHJlc3Mgc3dpdGNoaW5nLiBTbyBvZiBj b3Vyc2UgeW91IGRvIG5vdCB3YW50IHRvIGhhdmUgdG88L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4m Z3Q7IGNob29zZSBiZXR3ZWVuIHZsYW4gYW5kIGdtcGxzLCB5b3Ugd291bGQgd2FudCBib3RoIGF0 IHRoZSBzYW1lIHRpbWUuPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyA8L0ZPTlQ+PEJSPjxG T05UIFNJWkU9ND4mZ3Q7IC9wZXI8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7IDwvRk9OVD48 QlI+PEZPTlQgU0laRT00PiZndDsgPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7PC9G T05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7IC9Mb2E8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9 ND4mZ3Q7ICZndDs8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsgUCZhdW1sO3IgTWF0 dHNzb24gd3JvdGU6PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7Jmd0OyZndDtoaSBw YXIsIG9uZSBvZiB0aGUgcG9zc2liaWxpdGllcyB0aGF0IGhhcyBiZWVuIGNvbnNpZGVyZWQgdG8g Y29wZSB3aXRoPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7Jmd0OyZndDt0aGlzIHJl cXVpcmVtZW50IGlzIHRvIHVzZSBhIGRlZGljYXRlZCBUUElEIGZvciB0aGUgRXRoZXJuZXQgbGFi ZWxlZDwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyZndDsmZ3Q7ZnJhbWVzOyB0aGlz IHdvdWxkIGFsbG93IGRpZmZlcmVudGlhdGVkIHByb2Nlc3Npbmcgd2l0aCBub24tbGFiZWxlZDwv Rk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyZndDsmZ3Q7ZnJhbWVzdGhhbmtzLjwvRk9O VD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyZndDs8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4m Z3Q7ICZndDsmZ3Q7PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7Jmd0OyBUaGF0IHNl ZW1zIHRvIG1ha2UgbW9yZSBzZW5jZS4gSWYgdGhhdCBmcmFtZSBpcyB0byBiZSBzaXplZCBsaWtl IDwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyZndDsgODAyLjFxPC9GT05UPjxCUj48 Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7Jmd0OyBmcmFtZS4gVGhlcmUgaXMgbm90IHRoYXQgbXVjaCBz cGFjZSBsZWZ0IHRvIGEgbGFiZWwuIE9yIGlzIHRoZSBkZW1hbmQgdG88L0ZPTlQ+PEJSPjxGT05U IFNJWkU9ND4mZ3Q7ICZndDsmZ3Q7IHVzZSBqdW1ib2ZyYW1lcyA/PC9GT05UPjxCUj48Rk9OVCBT SVpFPTQ+Jmd0OyAmZ3Q7Jmd0OyBIYXMgdGhlcmUgYmVlbiBhbnkgZGlzY3Vzc2lvbiBvbiBsYWJl bHN0YWNraW5nLCBhbmQgbWFpbmx5IHdoZXJlIHRvPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0 OyAmZ3Q7Jmd0OyBwbGFjZTwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyZndDsgdGhl IGluZm9ybWF0aW9uPzwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyZndDs8L0ZPTlQ+ PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsmZ3Q7IFJlZ2FyZHMuPC9GT05UPjxCUj48Rk9OVCBT SVpFPTQ+Jmd0OyAmZ3Q7Jmd0OyBQZXI8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsm Z3Q7PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7Jmd0OzwvRk9OVD48QlI+PEZPTlQg U0laRT00PiZndDsgJmd0OyZndDs8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDs8L0ZP TlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDs8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7 ICZndDsgLS08L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsgTG9hIEFuZGVyc3Nvbjwv Rk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OzwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZn dDsgJmd0OyBQcmluY2lwYWwgTmV0d29ya2luZyBBcmNoaXRlY3Q8L0ZPTlQ+PEJSPjxGT05UIFNJ WkU9ND4mZ3Q7ICZndDsgQWNyZW8gQUImbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJz cDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsm bmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJz cDsmbmJzcDsgcGhvbmU6Jm5ic3A7ICs0NiA4IDYzMiA3NyAxNDwvRk9OVD48QlI+PEZPTlQgU0la RT00PiZndDsgJmd0OyBJc2Fmam9yZHNnYXRhbiAyMiZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZu YnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNw OyZuYnNwOyZuYnNwOyZuYnNwOyBtb2JpbGU6ICs0NiA3MzkgODEgMjEgNjQ8L0ZPTlQ+PEJSPjxG T05UIFNJWkU9ND4mZ3Q7ICZndDsgS2lzdGEsIFN3ZWRlbiZuYnNwOyZuYnNwOyZuYnNwOyZuYnNw OyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZu YnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyBlbWFpbDombmJz cDsgPC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOmxvYS5h bmRlcnNzb25AYWNyZW8uc2U+bG9hLmFuZGVyc3NvbkBhY3Jlby5zZTwvQT48L1U+PC9GT05UPjxG T05UIENPTE9SPUJMQUNLPjxCUj48L0ZPTlQ+PEZPTlQgU0laRT00PiZndDsgJmd0OyZuYnNwOyZu YnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNw OyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZu YnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNw OyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZu YnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyA8L0ZPTlQ+PEZPTlQgU0laRT00IENPTE9SPUJM VUU+PFU+PEEgSFJFRj1tYWlsdG86bG9hQHBpLnNlPmxvYUBwaS5zZTwvQT48L1U+PC9GT05UPjxG T05UIENPTE9SPUJMQUNLPjxCUj48L0ZPTlQ+PEZPTlQgU0laRT00PiZndDsgJmd0OzwvRk9OVD48 QlI+PEZPTlQgU0laRT00PiZndDsgPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyA8L0ZPTlQ+ PEJSPjxGT05UIFNJWkU9ND4mZ3Q7IDwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgPC9GT05U PjxCUj48QlI+PEZPTlQgU0laRT01Pj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PTwvRk9OVD48QlI+PEZPTlQgU0laRT01PlRoZSBpbmZv cm1hdGlvbiBjb250YWluZWQgaW4gdGhpcyBtZXNzYWdlIG1heSBiZSBwcml2aWxlZ2VkPC9GT05U PjxCUj48Rk9OVCBTSVpFPTU+YW5kIGNvbmZpZGVudGlhbCBhbmQgcHJvdGVjdGVkIGZyb20gZGlz Y2xvc3VyZS4gSWYgdGhlIHJlYWRlcjwvRk9OVD48QlI+PEZPTlQgU0laRT01Pm9mIHRoaXMgbWVz c2FnZSBpcyBub3QgdGhlIGludGVuZGVkIHJlY2lwaWVudCwgb3IgYW4gZW1wbG95ZWU8L0ZPTlQ+ PEJSPjxGT05UIFNJWkU9NT5vciBhZ2VudCByZXNwb25zaWJsZSBmb3IgZGVsaXZlcmluZyB0aGlz IG1lc3NhZ2UgdG8gdGhlPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+aW50ZW5kZWQgcmVjaXBpZW50 LCB5b3UgYXJlIGhlcmVieSBub3RpZmllZCB0aGF0IGFueSByZXByb2R1Y3Rpb24sPC9GT05UPjxC Uj48Rk9OVCBTSVpFPTU+ZGlzc2VtaW5hdGlvbiBvciBkaXN0cmlidXRpb24gb2YgdGhpcyBjb21t dW5pY2F0aW9uIGlzIHN0cmljdGx5PC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+cHJvaGliaXRlZC4g SWYgeW91IGhhdmUgcmVjZWl2ZWQgdGhpcyBjb21tdW5pY2F0aW9uIGluIGVycm9yLDwvRk9OVD48 QlI+PEZPTlQgU0laRT01PnBsZWFzZSBub3RpZnkgdXMgaW1tZWRpYXRlbHkgYnkgcmVwbHlpbmcg dG8gdGhlIG1lc3NhZ2UgYW5kPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+ZGVsZXRpbmcgaXQgZnJv bSB5b3VyIGNvbXB1dGVyLiBUaGFuayB5b3UuIFRlbGxhYnM8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9 NT49PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT08L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIi Pj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PTwvRk9OVD48QlI+PEZPTlQgU0laRT00IEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5UaGUg aW5mb3JtYXRpb24gY29udGFpbmVkIGluIHRoaXMgbWVzc2FnZSBtYXkgYmUgcHJpdmlsZWdlZDwv Rk9OVD48QlI+PEZPTlQgU0laRT00IEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5hbmQgY29uZmlk ZW50aWFsIGFuZCBwcm90ZWN0ZWQgZnJvbSBkaXNjbG9zdXJlLiBJZiB0aGUgcmVhZGVyPC9GT05U PjxCUj48Rk9OVCBTSVpFPTQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPm9mIHRoaXMgbWVzc2Fn ZSBpcyBub3QgdGhlIGludGVuZGVkIHJlY2lwaWVudCwgb3IgYW4gZW1wbG95ZWU8L0ZPTlQ+PEJS PjxGT05UIFNJWkU9NCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+b3IgYWdlbnQgcmVzcG9uc2li bGUgZm9yIGRlbGl2ZXJpbmcgdGhpcyBtZXNzYWdlIHRvIHRoZTwvRk9OVD48QlI+PEZPTlQgU0la RT00IEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5pbnRlbmRlZCByZWNpcGllbnQsIHlvdSBhcmUg aGVyZWJ5IG5vdGlmaWVkIHRoYXQgYW55IHJlcHJvZHVjdGlvbiw8L0ZPTlQ+PEJSPjxGT05UIFNJ WkU9NCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+ZGlzc2VtaW5hdGlvbiBvciBkaXN0cmlidXRp b24gb2YgdGhpcyBjb21tdW5pY2F0aW9uIGlzIHN0cmljdGx5PC9GT05UPjxCUj48Rk9OVCBTSVpF PTQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPnByb2hpYml0ZWQuIElmIHlvdSBoYXZlIHJlY2Vp dmVkIHRoaXMgY29tbXVuaWNhdGlvbiBpbiBlcnJvciw8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9NCBG QUNFPSJNb25vc3BhY2UsQ291cmllciI+cGxlYXNlIG5vdGlmeSB1cyBpbW1lZGlhdGVseSBieSBy ZXBseWluZyB0byB0aGUgbWVzc2FnZSBhbmQ8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9NCBGQUNFPSJN b25vc3BhY2UsQ291cmllciI+ZGVsZXRpbmcgaXQgZnJvbSB5b3VyIGNvbXB1dGVyLiBUaGFuayB5 b3UuIFRlbGxhYnM8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9NCBGQUNFPSJNb25vc3BhY2UsQ291cmll ciI+PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PC9GT05UPjxCUj48L1A+ Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 23:07:53 +0000 Message-ID: <01c401c58f12$4c80cd20$b7849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <neil.2.harrison@bt.com> Cc: <ccamp@ops.ietf.org> Subject: Re: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Date: Sat, 23 Jul 2005 00:07:39 +0100 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_01BD_01C58F1A.8A0CE630" This is a multi-part message in MIME format. ------=_NextPart_000_01BD_01C58F1A.8A0CE630 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Hi Neil, I think John beat me off the blocks here, but... > > GMPLS assumes an IP control plane. > > An IP control-plane? There is actually no such animal. Just what=20 > the heck does that REALLY mean in GMPLS say? Let me explain. Perhaps I should have said "IP-based control plane". I mean a control plane which: - uses IP as its network protocol - uses IP addresses to identify control plane resources - uses IP addresses to identify data plane resources within=20 the control plane - uses protocols developed for use in the Internet. > I am not questioning IP as a cl-ps networking protocol *carrying*=20 > a signalling protocol (RSVP-TE, or dare I mention PNNI, or any=20 > signalling protocol yet to be invented) or a routing protocol=20 > (OSPF or ISO or whatever) I am glad to hear it. > or even management protocols Fine, but not in the remit of CCAMP. > but an 'IP Control Plane' per se means absolutely nothing to me.... Well, I think it should. I think the list of attributes that I have = given above define a control plane based on IP.=20 It is undoubtable that attempts have been made to use control planes = based on other protocols. Some have been highly successful. Some have = been less fortunate. > ...nor should it to anyone else.=20 I think folks who were around at the beginning of CCAMP and who were = part of the debate with the IESG will be very familiar with where the = IETF draws the line here. > I think some folks might need a reality check here....and also > on the self-assumed importance of a control-plane. Hint: It ain't=20 > that important.....the management-plane (which may be using IP!)=20 > however is. I am not sure how to interpret this. It may be that you think that control plane is bad per se, but you have = said elsewhere that you think it has value - but much less than the = management plane. It may be that you believe that CCAMP is willfully neglecting the = management plane. This would, in fact, be true. It is not in CCAMP's = remit to look at the management plane. Other SDOs are working to = establish common standards for management of network devices across = multiple vendors - we can wish them luck. Clearly some people (vendors and providers) see sufficient value in a = control plane to invest time and energy. > The (hype) party is over for the OTN start-ups. IP per se does NOT > define a *control-plane*...IP is cl-ps networking techology period > ....and its jolly important, but PLEASE don't try and feed me any of > this 'IP control plane' nonsense. Tush. > Adrian....you are far smarter than this IMO and should know better. Thanks for the endorsement. May I quote you? Adrian ------=_NextPart_000_01BD_01C58F1A.8A0CE630 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META http-equiv=3DContent-Type content=3D"text/html; = charset=3Diso-8859-1"> <META content=3D"MSHTML 6.00.2800.1106" name=3DGENERATOR> <STYLE></STYLE> </HEAD> <BODY> <DIV><FONT face=3DCourier size=3D2>Hi Neil,</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>I think John beat me off the blocks = here,=20 but...</FONT></DIV> <DIV><FONT face=3DCourier><BR><FONT size=3D2></FONT></FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> > GMPLS assumes an IP control = plane.<BR>><BR>> An IP control-plane? There is actually no = such=20 animal. Just what </FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> the </FONT><FONT = face=3DCourier=20 size=3D2>heck does that REALLY mean in GMPLS say?</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Let me explain.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Perhaps I should have said "IP-based = control=20 plane".</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>I mean a control plane = which:</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>- uses IP as its network = protocol</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>- uses IP addresses to identify = control plane=20 resources</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>- uses IP addresses to identify data = plane=20 resources within </FONT></DIV> <DIV><FONT face=3DCourier size=3D2> the control plane</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>- uses protocols developed for use in = the=20 Internet.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>> I am not = questioning </FONT><FONT=20 face=3DCourier size=3D2>IP</FONT><FONT face=3DCourier size=3D2> as = a cl-ps=20 networking protocol *carrying* </FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> a signalling = protocol </FONT><FONT=20 face=3DCourier size=3D2>(</FONT><FONT face=3DCourier size=3D2>RSVP-TE, = or dare I mention=20 PNNI, or any </FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> signalling protocol = yet </FONT><FONT=20 face=3DCourier><FONT size=3D2><FONT>to</FONT></FONT></FONT><FONT = face=3DCourier=20 size=3D2> be invented) or a routing protocol </FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> (OSPF or ISO or = whatever)</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>I am glad to hear it.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>> or </FONT><FONT = face=3DCourier=20 size=3D2>even management protocols</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Fine, but not in the remit of = CCAMP.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>> but an 'IP Control Plane' = per se=20 means </FONT><FONT face=3DCourier size=3D2>absolutely nothing to=20 me....</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Well, I think it should. I think the = list of=20 attributes that I have given above define a control plane based on IP.=20 </FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>It is undoubtable that attempts have = been made to=20 use control planes based on other protocols. Some have been highly = successful.=20 Some have been less fortunate.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>> ...nor should it to anyone=20 else. </FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>I think folks who were around at the = beginning of=20 CCAMP and who were part of the debate with the IESG will be very = familiar with=20 where the IETF draws the line here.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>> I think </FONT><FONT = face=3DCourier=20 size=3D2>some folks might need a reality check here....and = also</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> on the self-</FONT><FONT = face=3DCourier=20 size=3D2>assumed importance of a control-plane. H</FONT><FONT = face=3DCourier=20 size=3D2>int: It ain't </FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> that </FONT><FONT = face=3DCourier=20 size=3D2>important.....the management-plane (which may be using IP!) = </FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> however</FONT><FONT = face=3DCourier=20 size=3D2> is.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>I am not sure how to interpret = this.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>It may be that you think that control = plane is=20 bad per se, but you have said elsewhere that you think it has value - = but much=20 less than the management plane.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>It may be that you believe that CCAMP = is=20 willfully neglecting the management plane. This would, in fact, be true. = It is=20 not in CCAMP's remit to look at the management plane. Other SDOs are = working to=20 establish common standards for management of network devices across = multiple=20 vendors - we can wish them luck.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Clearly some people (vendors and = providers) see=20 sufficient value in a control plane to invest time and energy.</DIV> <DIV><BR>> The (hype) party is over for the OTN start-ups. IP = per se=20 does NOT</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> define a = *control-plane*...IP is cl-ps=20 networking techology period</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> ....and its jolly important, but = PLEASE=20 don't try and feed me any of</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>> this 'IP control plane'=20 nonsense.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Tush.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> </DIV> <DIV>> Adrian....you are far smarter than this IMO and should know=20 better.</DIV> <DIV> </DIV> <DIV>Thanks for the endorsement. May I quote you?</DIV> <DIV> </DIV> <DIV>Adrian<BR></DIV></FONT><FONT face=3DCourier = size=3D2></FONT></BODY></HTML> ------=_NextPart_000_01BD_01C58F1A.8A0CE630-- Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 22:20:48 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Date: Fri, 22 Jul 2005 15:18:35 -0700 Message-ID: <626FC7C6A97381468FB872072AB5DDC836941D@XCH-SW-42.sw.nos.boeing.com> Thread-Topic: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Thread-Index: AcWO6PXwNV9J9GsyTyywah8dc4SA0QAESRbgAAQtiVA= From: "Drake, John E" <John.E.Drake2@boeing.com> To: <neil.2.harrison@bt.com>, <adrian@olddog.co.uk>, <richard.spencer@bt.com>, <juergen.heiles@siemens.com>, <loa@pi.se>, <per@defero.se> Cc: <ccamp@ops.ietf.org> Neil, Adrian is correct. RSVP, OSPF/IS-IS, and BGP are the control plane = protocols that are used to control an IP network. GMPLS assumes the = same set of control plane protocols. Thanks, John > -----Original Message----- > From: neil.2.harrison@bt.com [mailto:neil.2.harrison@bt.com] > Sent: Friday, July 22, 2005 2:22 PM > To: adrian@olddog.co.uk; richard.spencer@bt.com; > juergen.heiles@siemens.com; loa@pi.se; per@defero.se > Cc: ccamp@ops.ietf.org > Subject: RE: Ethernet Control Plane [Was: Re: Frameformat in a l2cs = gmpls > rnvironment.] >=20 > > GMPLS assumes an IP control plane. >=20 > An IP control-plane? There is actually no such animal. Just what the > heck does that REALLY mean in GMPLS say? I am not questioning IP as a = cl- > ps networking protocol *carrying* a signalling protocol (RSVP-TE, or = dare > I mention PNNI, or any signalling protocol yet to be invented) or a > routing protocol (OSPF or ISO or whatever) or even management = protocols > but an 'IP Control Plane' per se means absolutely nothing to me....nor > should it to anyone else. I think some folks might need a reality = check > here....and also on the self-assumed importance of a control-plane. = Hint: > It ain't that important.....the management-plane (which may be using = IP!) > however is. >=20 > The (hype) party is over for the OTN start-ups. IP per se does NOT = define > a *control-plane*...IP is cl-ps networking techology period....and its > jolly important, but PLEASE don't try and feed me any of this 'IP = control > plane' nonsense. >=20 > Adrian....you are far smarter than this IMO and should know better. >=20 > regards, Neil >=20 > > -----Original Message----- > > From: owner-ccamp@ops.ietf.org > > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Adrian Farrel > > Sent: 22 July 2005 19:13 > > To: Spencer,R,Richard,XDE73 R; juergen.heiles@siemens.com; > > loa@pi.se; per@defero.se > > Cc: ccamp@ops.ietf.org > > Subject: Ethernet Control Plane [Was: Re: Frameformat in a > > l2cs gmpls rnvironment.] > > > > > > Whoa there! > > > > GMPLS assumes an IP control plane. > > I am aware that there is work afoot in other SDOs to assign a > > dedicated VLAN tag to specify the control plane traffic. This > > is not what we are doing. > > > > As far as I am aware, everyone in the IETF is comfortable > > with the idea of using IP packets to carry control plane data > > through an IP-enabled DCN. In Ethernet the issue will be > > whether the control plane is in band (use IP address and MAC > > address) or out of band (use distinct DCN). > > > > The use of a separate VLAN ID to signify the control plane is > > out of scope for CCAMP. > > > > Thanks, > > Adrian > > > > ----- Original Message ----- > > From: <richard.spencer@bt.com> > > To: <juergen.heiles@siemens.com>; <loa@pi.se>; <per@defero.se> > > Cc: <ccamp@ops.ietf.org> > > Sent: Friday, July 22, 2005 4:33 PM > > Subject: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > Juergen, > > > > Before you can use GMPLS to set up VLANs a data plane must > > first exist to carry the GMPLS control plane traffic. Yes you > > can use the standard Ethernet data plane for this purpose > > (and in fact must do so for the solution to be viable), but > > its not as simple as that. You cannot assume that just > > because two GMPLS peers are directly connected that they can > > communicate with each other. > > > > To exchange GMPLS control plane packets between a pair of > > GMPLS peers, the peers will need to agree to use either > > untagged control packets or to tag them with a common VLAN > > ID. The use of a separate VLAN for GMPLS control traffic is > > analogous to using a separate VLAN for management traffic > > today. If two switches belong to the same management VLAN > > then you can use ICMP ping to verify connectivity between > > them. However, if one switch has been put into the wrong > > management VLAN then the ping will obviously fail. > > > > If you want to use a separate VLAN for GMPLS control plane > > traffic, then you either need to (i) define a reserved GMPLS > > signalling VLAN ID, or (ii) leave it up to the operator to > > choose a VLAN ID. If you leave it up to the operator to > > decide what VLAN ID to use, then you need to consider what > > the implications in the inter-operator case will be, i.e. > > what VLAN ID range should be used for control plane traffic > > and what VLAN ID range should be used for user plane traffic. > > > > I'm not saying that defining how GMPLS control plane packets > > are forwarded in the Ethernet data plane will be an issue, > > simply that this is something that must be addressed to > > develop interoperable solutions. > > > > Regards, > > Richard > > > > > -----Original Message----- > > > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On > > > Behalf Of Heiles Juergen > > > Sent: 22 July 2005 14:17 > > > To: Loa Andersson; P=E4r Mattsson > > > Cc: ccamp@ops.ietf.org > > > Subject: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > > > > GMPLS for SONET/SDH has no impact on the SONET/SDH data plane. It > > > introduces a control plane for connection setup instead of > > a network > > > management based connection setup. GMPLS for Ethernet can > > work in the > > > same way. Ethernet as such is cl, but VLANs are setup and GMPLS = can > > > provide a control plane for VLAN setup. The Ethernet data > > plane is not > > > changed and MAC address learning and forwarding is still be done > > > within a VLAN. So no change to the Ethernet data plane. No GMPLS > > > traffic is introduced, it is still VLAN traffic. > > > Introducing a dedicated GMPLS label for Ethernet (GMPLS > > > traffic) goes beyond GMPLS as a control plane technique. > > > > > > Regards > > > > > > Juergen > > > > > > > > > > > > > > > > -----Original Message----- > > > > From: owner-ccamp@ops.ietf.org > > [mailto:owner-ccamp@ops.ietf.org] On > > > > Behalf Of Loa Andersson > > > > Sent: Friday, July 22, 2005 12:05 PM > > > > To: P=E4r Mattsson > > > > Cc: ccamp@ops.ietf.org > > > > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > > > > > > Per and Dimitri, > > > > > > > > I would like to come down stronger than that, for me it is a = very > > > > strong requirement that the same switch can handle both VLANs = and > > > > GMPLs trafic correctly. I can't dsee how that could be > > done if using > > > > the VLAN tpid to indicate GMPLS traffic. > > > > > > > > /Loa > > > > > > > > P=E4r Mattsson wrote: > > > > >>hi par, one of the possibilities that has been considered > > > > to cope with > > > > >>this requirement is to use a dedicated TPID for the > > > Ethernet labeled > > > > >>frames; this would allow differentiated processing with > > > non-labeled > > > > >>framesthanks. > > > > > > > > > > > > > > > That seems to make more sence. If that frame is to be sized > > > > like a 802.1q > > > > > frame. There is not that much space left to a label. Or is > > > > the demand to > > > > > use jumboframes ? > > > > > Has there been any discussion on labelstacking, and mainly > > > > where to place > > > > > the information? > > > > > > > > > > Regards. > > > > > Per > > > > > > > > > > > > > > > > > > > > > > > > > > > -- > > > > Loa Andersson > > > > > > > > Principal Networking Architect > > > > Acreo AB phone: +46 8 632 77 14 > > > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > > > Kista, Sweden email: = loa.andersson@acreo.se > > > > loa@pi.se > > > > > > > > > > > > > > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 21:23:58 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Date: Fri, 22 Jul 2005 22:22:13 +0100 Message-ID: <0536FC9B908BEC4597EE721BE6A353890A9F1E2E@i2km07-ukbr.domain1.systemhost.net> Thread-Topic: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Thread-Index: AcWO6PXwNV9J9GsyTyywah8dc4SA0QAESRbg From: <neil.2.harrison@bt.com> To: <adrian@olddog.co.uk>, <richard.spencer@bt.com>, <juergen.heiles@siemens.com>, <loa@pi.se>, <per@defero.se> Cc: <ccamp@ops.ietf.org> > GMPLS assumes an IP control plane. An IP control-plane? There is actually no such animal. Just what the = heck does that REALLY mean in GMPLS say? I am not questioning IP as a = cl-ps networking protocol *carrying* a signalling protocol (RSVP-TE, or = dare I mention PNNI, or any signalling protocol yet to be invented) or a = routing protocol (OSPF or ISO or whatever) or even management protocols = but an 'IP Control Plane' per se means absolutely nothing to me....nor = should it to anyone else. I think some folks might need a reality = check here....and also on the self-assumed importance of a = control-plane. Hint: It ain't that important.....the management-plane = (which may be using IP!) however is. The (hype) party is over for the OTN start-ups. IP per se does NOT = define a *control-plane*...IP is cl-ps networking techology = period....and its jolly important, but PLEASE don't try and feed me any = of this 'IP control plane' nonsense. Adrian....you are far smarter than this IMO and should know better. regards, Neil > -----Original Message----- > From: owner-ccamp@ops.ietf.org > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Adrian Farrel > Sent: 22 July 2005 19:13 > To: Spencer,R,Richard,XDE73 R; juergen.heiles@siemens.com;=20 > loa@pi.se; per@defero.se > Cc: ccamp@ops.ietf.org > Subject: Ethernet Control Plane [Was: Re: Frameformat in a=20 > l2cs gmpls rnvironment.] >=20 >=20 > Whoa there! >=20 > GMPLS assumes an IP control plane. > I am aware that there is work afoot in other SDOs to assign a > dedicated VLAN tag to specify the control plane traffic. This=20 > is not what we are doing. >=20 > As far as I am aware, everyone in the IETF is comfortable > with the idea of using IP packets to carry control plane data=20 > through an IP-enabled DCN. In Ethernet the issue will be=20 > whether the control plane is in band (use IP address and MAC=20 > address) or out of band (use distinct DCN). >=20 > The use of a separate VLAN ID to signify the control plane is > out of scope for CCAMP. >=20 > Thanks, > Adrian >=20 > ----- Original Message ----- > From: <richard.spencer@bt.com> > To: <juergen.heiles@siemens.com>; <loa@pi.se>; <per@defero.se> > Cc: <ccamp@ops.ietf.org> > Sent: Friday, July 22, 2005 4:33 PM > Subject: RE: Frameformat in a l2cs gmpls rnvironment. >=20 >=20 > Juergen, >=20 > Before you can use GMPLS to set up VLANs a data plane must > first exist to carry the GMPLS control plane traffic. Yes you=20 > can use the standard Ethernet data plane for this purpose=20 > (and in fact must do so for the solution to be viable), but=20 > its not as simple as that. You cannot assume that just=20 > because two GMPLS peers are directly connected that they can=20 > communicate with each other. >=20 > To exchange GMPLS control plane packets between a pair of > GMPLS peers, the peers will need to agree to use either=20 > untagged control packets or to tag them with a common VLAN=20 > ID. The use of a separate VLAN for GMPLS control traffic is=20 > analogous to using a separate VLAN for management traffic=20 > today. If two switches belong to the same management VLAN=20 > then you can use ICMP ping to verify connectivity between=20 > them. However, if one switch has been put into the wrong=20 > management VLAN then the ping will obviously fail. >=20 > If you want to use a separate VLAN for GMPLS control plane > traffic, then you either need to (i) define a reserved GMPLS=20 > signalling VLAN ID, or (ii) leave it up to the operator to=20 > choose a VLAN ID. If you leave it up to the operator to=20 > decide what VLAN ID to use, then you need to consider what=20 > the implications in the inter-operator case will be, i.e.=20 > what VLAN ID range should be used for control plane traffic=20 > and what VLAN ID range should be used for user plane traffic. >=20 > I'm not saying that defining how GMPLS control plane packets > are forwarded in the Ethernet data plane will be an issue,=20 > simply that this is something that must be addressed to=20 > develop interoperable solutions. >=20 > Regards, > Richard >=20 > > -----Original Message----- > > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On > > Behalf Of Heiles Juergen > > Sent: 22 July 2005 14:17 > > To: Loa Andersson; P=E4r Mattsson > > Cc: ccamp@ops.ietf.org > > Subject: RE: Frameformat in a l2cs gmpls rnvironment. > > > > > > GMPLS for SONET/SDH has no impact on the SONET/SDH data plane. It > > introduces a control plane for connection setup instead of=20 > a network > > management based connection setup. GMPLS for Ethernet can > work in the > > same way. Ethernet as such is cl, but VLANs are setup and GMPLS can > > provide a control plane for VLAN setup. The Ethernet data=20 > plane is not > > changed and MAC address learning and forwarding is still be done > > within a VLAN. So no change to the Ethernet data plane. No GMPLS > > traffic is introduced, it is still VLAN traffic. > > Introducing a dedicated GMPLS label for Ethernet (GMPLS > > traffic) goes beyond GMPLS as a control plane technique. > > > > Regards > > > > Juergen > > > > > > > > > > > -----Original Message----- > > > From: owner-ccamp@ops.ietf.org > [mailto:owner-ccamp@ops.ietf.org] On > > > Behalf Of Loa Andersson > > > Sent: Friday, July 22, 2005 12:05 PM > > > To: P=E4r Mattsson > > > Cc: ccamp@ops.ietf.org > > > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > > > > Per and Dimitri, > > > > > > I would like to come down stronger than that, for me it is a very > > > strong requirement that the same switch can handle both VLANs and=20 > > > GMPLs trafic correctly. I can't dsee how that could be=20 > done if using > > > the VLAN tpid to indicate GMPLS traffic. > > > > > > /Loa > > > > > > P=E4r Mattsson wrote: > > > >>hi par, one of the possibilities that has been considered > > > to cope with > > > >>this requirement is to use a dedicated TPID for the > > Ethernet labeled > > > >>frames; this would allow differentiated processing with > > non-labeled > > > >>framesthanks. > > > > > > > > > > > > That seems to make more sence. If that frame is to be sized > > > like a 802.1q > > > > frame. There is not that much space left to a label. Or is > > > the demand to > > > > use jumboframes ? > > > > Has there been any discussion on labelstacking, and mainly > > > where to place > > > > the information? > > > > > > > > Regards. > > > > Per > > > > > > > > > > > > > > > > > > > > > -- > > > Loa Andersson > > > > > > Principal Networking Architect > > > Acreo AB phone: +46 8 632 77 14 > > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > > Kista, Sweden email: loa.andersson@acreo.se > > > loa@pi.se > > > > > > > >=20 >=20 >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 21:23:52 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01C58F03.6D2AEF23" Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 22:22:12 +0100 Message-ID: <0536FC9B908BEC4597EE721BE6A353890A9F1E2D@i2km07-ukbr.domain1.systemhost.net> Thread-Topic: Frameformat in a l2cs gmpls rnvironment. Thread-Index: AcWO4Z/sn5bUTdc8RiiIwKwSk2UxnAACGn2QAAWRjkA= From: <neil.2.harrison@bt.com> To: <Ben.Mack-Crane@tellabs.com>, <ibryskin@movaz.com>, <Dimitri.Papadimitriou@alcatel.be> Cc: <adrian@olddog.co.uk>, <per@defero.se>, <loa@pi.se>, <ccamp@ops.ietf.org> This is a multi-part message in MIME format. ------_=_NextPart_001_01C58F03.6D2AEF23 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable And me too Ben/Igor. But go look at MPLS when folks DID define a = data-plane here....don't anyone dare tell me 'this is good' (or the PW = stuff it spawned) because I know otherwise. =20 reqards, Neil -----Original Message----- From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On = Behalf Of Mack-Crane, T. Benjamin Sent: 22 July 2005 19:25 To: Igor Bryskin; Dimitri.Papadimitriou@alcatel.be Cc: Adrian Farrel; P=E4r Mattsson; Loa Andersson; ccamp@ops.ietf.org Subject: RE: Frameformat in a l2cs gmpls rnvironment. Igor, =20 I, too, am baffled by the notion of requirements for control of an = unspecified data plane. =20 Regards, Ben _____ =20 From: Igor Bryskin [mailto:ibryskin@movaz.com]=20 Sent: Friday, July 22, 2005 12:20 PM To: Dimitri.Papadimitriou@alcatel.be Cc: Mack-Crane, T. Benjamin; Adrian Farrel; P=E4r Mattsson; Loa = Andersson; ccamp@ops.ietf.org Subject: Re: Frameformat in a l2cs gmpls rnvironment. Dimitri, =20 The question is simple: is it possible today to statically provision = L2LSPs that could, say, support e2e QoS? If not, may be it is to early = to discuss aspects of dynamic provisioning of such LSPs? Is it = possible/reasonable, in your opinion, to "detail how forwarding = information can be exchanged via the control plane (and then installed)" = without having in mind "a specific forwarding behavior"? How do you know which forwarding information is needed for the = forwarding nobody has defined yet? The other question is do we need at all e2e QoS, route control, fast = recovery? Sounds like exciting idea, but does all that commercially make = sense? So, I guess, we need two things before we can move forward: a) validation of the idea by the providers; b) definition of the data plane behavior; =20 Igor ----- Original Message -----=20 From: Dimitri.Papadimitriou@alcatel.be=20 To: Igor Bryskin <mailto:ibryskin@movaz.com> =20 Cc: Mack-Crane, T. Benjamin <mailto:Ben.Mack-Crane@tellabs.com> ; = Adrian <mailto:adrian@olddog.co.uk> Farrel ; P=E4r = <mailto:per@defero.se> Mattsson ; Loa <mailto:loa@pi.se> Andersson ; = ccamp@ops.ietf.org=20 Sent: Friday, July 22, 2005 12:21 PM Subject: Re: Frameformat in a l2cs gmpls rnvironment. igor - i am not sure which point they exactly have - the purpose of this = document is to detail how forwarding information can be exchanged via = the control plane (and then installed) not to define a specific = forwarding behaviour thanks, - dimitri. "Igor Bryskin" <ibryskin@movaz.com> Sent by: owner-ccamp@ops.ietf.org 07/22/2005 11:10 AST To: "Mack-Crane, T. Benjamin" <Ben.Mack-Crane@tellabs.com>, "Adrian = Farrel" <adrian@olddog.co.uk>, P=E4r Mattsson <per@defero.se>, "Loa = Andersson" <loa@pi.se> cc: <ccamp@ops.ietf.org> bcc:=20 Subject: Re: Frameformat in a l2cs gmpls rnvironment. Hi, I think Ben and Juergen have a point here. There is nothing that could = be dynamically provisioned that could not be also provisioned by = management. In other words data plane is the King here. We need data = plane standard(s) on how we encode a label, whether we need a label = stack or not, how the labeled traffic is supposed to be treated, how = labeled traffic co-exist with non-label traffic, etc. This is something = that not for CCAMP to define. Take for example TDM networks. GMPLS only = provides a way to dynamically provision G.707 networks. Hence there is a = need in parallel standardization activities in ITU. Only after that we = can discuss how all that could be dynamically provisioned, that is the = aspects of control plane.=20 Igor ----- Original Message -----=20 From: Mack-Crane, T. <mailto:Ben.Mack-Crane@tellabs.com> Benjamin=20 To: Adrian Farrel <mailto:adrian@olddog.co.uk> ; P=E4r Mattsson = <mailto:per@defero.se> ; Loa <mailto:loa@pi.se> Andersson=20 Cc: ccamp@ops.ietf.org=20 Sent: Friday, July 22, 2005 10:23 AM Subject: RE: Frameformat in a l2cs gmpls rnvironment. Hi, I think Juergen has raised an important question. How frames are = labeled (and the related data plane forwarding behavior) is not defined = by the control plane. The control plane serves to provision the data = plane, not define it. In the framework draft it is not clear what data = plane standards are covered by the stated control plane requirements. = Some references should be supplied. In any case, the labeling and = forwarding behavior should be defined by these referenced standards, not = by GMPLS. (I am assuming definition of new data plane standards is out of scope = for CCAMP.) Regards, Ben _____ =20 From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On = Behalf Of Adrian Farrel Sent: Friday, July 22, 2005 8:37 AM To: P=E4r Mattsson; Loa Andersson Cc: ccamp@ops.ietf.org Subject: Re: Frameformat in a l2cs gmpls rnvironment. Hi, It is always refreshing how engineers jump straight to a discussion of = the solution :-) Perhaps we can assume that the framework draft is a good representation = of the problem space, and that we are ready to start discussing the = solutions. One (perhaps the only?) significant question to answer is how the frames = will be labelled. This question is one we must come to as soon as we are = confident that the requirements need to be addressed at all. As has been pointed out, there are several possibilities and to pick = from these we need to understand: 1. Do we need to support explicit label stacking? Note that this is not supported in TDM, LSC or FSC. 2. Do we need to be able to control (perhaps through an external signaling controller) existing hardware and=20 install LSPs through existing networks? 3. Do we need to support existing function simultaneous to the support of GMPLS L2 LSPs? I think that from a chair's perspective I can give some limited = guidance. A. These questions must be raised and answered in the framework I-D B. The answer to question 3 is "yes". This means that the use of labels must not significantly deplete any namespace used to support other function. C. CCAMP is chartered to look at the control of transport networks. This includes Metro, but I am unsure about Campus. It does not cover=20 signaling to the desktop. Thanks, Adrian ----- Original Message -----=20 From: "P=E4r Mattsson" <per@defero.se> To: "Loa Andersson" <loa@pi.se> Cc: <ccamp@ops.ietf.org> Sent: Friday, July 22, 2005 11:41 AM Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > Per and Dimitri, > > > > I would like to come down stronger than that, for me it is > > a very strong requirement that the same switch can handle > > both VLANs and GMPLs trafic correctly. I can't dsee how that > > could be done if using the VLAN tpid to indicate GMPLS > > traffic. >=20 > If you ever want that same switch to handle traffic for a directly > connected host (not to uncommen) you would want that to use normal > ethernet macaddress switching. So of course you do not want to have to > choose between vlan and gmpls, you would want both at the same time. >=20 > /per >=20 >=20 > > > > /Loa > > > > P=E4r Mattsson wrote: > >>>hi par, one of the possibilities that has been considered to cope = with > >>>this requirement is to use a dedicated TPID for the Ethernet = labeled > >>>frames; this would allow differentiated processing with non-labeled > >>>framesthanks. > >> > >> > >> That seems to make more sence. If that frame is to be sized like a > >> 802.1q > >> frame. There is not that much space left to a label. Or is the = demand to > >> use jumboframes ? > >> Has there been any discussion on labelstacking, and mainly where to > >> place > >> the information? > >> > >> Regards. > >> Per > >> > >> > >> > > > > > > -- > > Loa Andersson > > > > Principal Networking Architect > > Acreo AB phone: +46 8 632 77 14 > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > Kista, Sweden email: loa.andersson@acreo.se > > loa@pi.se > > >=20 >=20 >=20 >=20 =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any reproduction, dissemination or distribution of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. Tellabs =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any reproduction, dissemination or distribution of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. Tellabs =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D ------_=_NextPart_001_01C58F03.6D2AEF23 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; = charset=3Diso-8859-1"> <TITLE>Message</TITLE> <META content=3D"MSHTML 6.00.2800.1106" name=3DGENERATOR> <STYLE></STYLE> </HEAD> <BODY bgColor=3D#ffffff> <DIV><SPAN class=3D689555920-22072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2>And me too Ben/Igor. But go look at MPLS when folks DID = define a=20 data-plane here....don't anyone dare tell me 'this is good' (or the PW = stuff it=20 spawned) because I know otherwise.</FONT></SPAN></DIV> <DIV><SPAN class=3D689555920-22072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2></FONT></SPAN> </DIV> <DIV><SPAN class=3D689555920-22072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2>reqards, Neil</FONT></SPAN></DIV> <BLOCKQUOTE dir=3Dltr=20 style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #800000 2px = solid; MARGIN-RIGHT: 0px"> <DIV></DIV> <DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr = align=3Dleft><FONT=20 face=3DTahoma size=3D2>-----Original Message-----<BR><B>From:</B>=20 owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] <B>On = Behalf Of=20 </B>Mack-Crane, T. Benjamin<BR><B>Sent:</B> 22 July 2005 = 19:25<BR><B>To:</B>=20 Igor Bryskin; Dimitri.Papadimitriou@alcatel.be<BR><B>Cc:</B> Adrian = Farrel;=20 P=E4r Mattsson; Loa Andersson; ccamp@ops.ietf.org<BR><B>Subject:</B> = RE:=20 Frameformat in a l2cs gmpls rnvironment.<BR><BR></FONT></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D935282018-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>Igor,</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D935282018-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2></FONT></SPAN> </DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D935282018-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>I, too, am baffled by the notion of = requirements for=20 control of an unspecified data plane.</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D935282018-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2></FONT></SPAN> </DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D935282018-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>Regards,</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D935282018-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>Ben</FONT></SPAN></DIV><BR> <BLOCKQUOTE dir=3Dltr=20 style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #0000ff 2px = solid; MARGIN-RIGHT: 0px"> <DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr = align=3Dleft> <HR tabIndex=3D-1> <FONT face=3DTahoma size=3D2><B>From:</B> Igor Bryskin=20 [mailto:ibryskin@movaz.com] <BR><B>Sent:</B> Friday, July 22, 2005 = 12:20=20 PM<BR><B>To:</B> Dimitri.Papadimitriou@alcatel.be<BR><B>Cc:</B> = Mack-Crane,=20 T. Benjamin; Adrian Farrel; P=E4r Mattsson; Loa Andersson;=20 ccamp@ops.ietf.org<BR><B>Subject:</B> Re: Frameformat in a l2cs = gmpls=20 rnvironment.<BR></FONT><BR></DIV> <DIV></DIV> <DIV><FONT face=3DArial size=3D2>Dimitri,</FONT></DIV> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <DIV><FONT face=3DArial size=3D2>The question is simple: is it = possible today to=20 statically provision L2LSPs that could, say, support e2e QoS? = If not,=20 may be it is to early to discuss aspects of dynamic provisioning of = such=20 LSPs? Is it possible/reasonable, in your opinion, to "<FONT=20 face=3D"Courier New" size=3D3>detail how forwarding information can = be exchanged=20 via the control plane (and then installed)" without having in = mind "a=20 specific forwarding behavior"?</FONT></FONT></DIV> <DIV><FONT face=3D"Courier New">How do you know which forwarding = information=20 is needed for the forwarding nobody has defined yet?</FONT></DIV> <DIV><FONT face=3D"Courier New">The other question is do we = need at all=20 e2e QoS, route control, fast recovery? Sounds like exciting idea, = but does=20 all that commercially make sense? So, I guess, we need two=20 things before we can move forward:</FONT></DIV> <DIV><FONT face=3D"Courier New">a) validation of the idea by the=20 providers;</FONT></DIV> <DIV><FONT face=3D"Courier New">b) definition of the data plane=20 behavior;</FONT></DIV> <DIV><FONT face=3D"Courier New"></FONT> </DIV> <DIV><FONT face=3D"Courier New">Igor</FONT></DIV> <BLOCKQUOTE dir=3Dltr=20 style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; = BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px"> <DIV style=3D"FONT: 10pt arial">----- Original Message ----- = </DIV> <DIV=20 style=3D"BACKGROUND: #e4e4e4; FONT: 10pt arial; font-color: = black"><B>From:</B>=20 <A title=3DDimitri.Papadimitriou@alcatel.be=20 = href=3D"mailto:Dimitri.Papadimitriou@alcatel.be">Dimitri.Papadimitriou@al= catel.be</A>=20 </DIV> <DIV style=3D"FONT: 10pt arial"><B>To:</B> <A = title=3Dibryskin@movaz.com=20 href=3D"mailto:ibryskin@movaz.com">Igor Bryskin</A> </DIV> <DIV style=3D"FONT: 10pt arial"><B>Cc:</B> <A=20 title=3DBen.Mack-Crane@tellabs.com=20 href=3D"mailto:Ben.Mack-Crane@tellabs.com">Mack-Crane, T. = Benjamin</A> ; <A=20 title=3Dadrian@olddog.co.uk = href=3D"mailto:adrian@olddog.co.uk">Adrian=20 Farrel</A> ; <A title=3Dper@defero.se = href=3D"mailto:per@defero.se">P=E4r=20 Mattsson</A> ; <A title=3Dloa@pi.se href=3D"mailto:loa@pi.se">Loa=20 Andersson</A> ; <A title=3Dccamp@ops.ietf.org=20 href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A> </DIV> <DIV style=3D"FONT: 10pt arial"><B>Sent:</B> Friday, July 22, 2005 = 12:21=20 PM</DIV> <DIV style=3D"FONT: 10pt arial"><B>Subject:</B> Re: Frameformat in = a l2cs=20 gmpls rnvironment.</DIV> <DIV><FONT face=3DArial size=3D2></FONT><BR></DIV> <P><FONT face=3DMonospace,Courier>igor - i am not sure which point = they=20 exactly have - the purpose of this document is to detail how = forwarding=20 information can be exchanged via the control plane (and then = installed)=20 not to define a specific forwarding behaviour</FONT></P> <P><FONT face=3DMonospace,Courier>thanks,</FONT></P> <P><FONT face=3DMonospace,Courier>- dimitri.</FONT><BR><BR><FONT=20 size=3D2><B>"Igor Bryskin" <<A=20 = href=3D"mailto:ibryskin@movaz.com">ibryskin@movaz.com</A>></B></FONT><= BR><FONT=20 size=3D2>Sent by: <A=20 = href=3D"mailto:owner-ccamp@ops.ietf.org">owner-ccamp@ops.ietf.org</A></FO= NT><BR><FONT=20 size=3D2>07/22/2005 11:10 AST</FONT><BR><BR><FONT = size=3D2>To:</FONT> <FONT=20 size=3D2>"Mack-Crane, T. Benjamin" <<A=20 = href=3D"mailto:Ben.Mack-Crane@tellabs.com">Ben.Mack-Crane@tellabs.com</A>= >,=20 "Adrian Farrel" <<A=20 href=3D"mailto:adrian@olddog.co.uk">adrian@olddog.co.uk</A>>, = P=E4r=20 Mattsson <<A = href=3D"mailto:per@defero.se">per@defero.se</A>>, "Loa=20 Andersson" <<A=20 href=3D"mailto:loa@pi.se">loa@pi.se</A>></FONT><BR><FONT=20 size=3D2>cc:</FONT> <FONT size=3D2><<A=20 = href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A>></FONT><BR><= FONT=20 size=3D2>bcc:</FONT> <BR><FONT size=3D2>Subject:</FONT> <FONT = size=3D2>Re:=20 Frameformat in a l2cs gmpls rnvironment.</FONT><BR><BR><BR></P> <P>Hi,<BR><BR>I think Ben and Juergen have a point here. There is = nothing=20 that could be dynamically provisioned that could not be also = provisioned=20 by management. In other words data plane is the King here. We need = data=20 plane standard(s) on how we encode a label, whether we need a = label stack=20 or not, how the labeled traffic is supposed to be treated, = how=20 labeled traffic co-exist with non-label traffic, etc. This is = something=20 that not for CCAMP to define. Take for example TDM networks. GMPLS = only=20 provides a way to dynamically provision G.707 networks. Hence = there is a=20 need in parallel standardization activities in ITU. Only after = that we can=20 discuss how all that could be dynamically provisioned, that is the = aspects=20 of control plane. <BR><BR>Igor<BR><FONT size=3D4>----- Original = Message=20 ----- </FONT><BR><FONT size=3D4><B>From:</B> </FONT><FONT = color=3Dblue=20 size=3D4><U><A = href=3D"mailto:Ben.Mack-Crane@tellabs.com">Mack-Crane, T.=20 Benjamin</A></U></FONT><FONT color=3Dblack size=3D4> = </FONT><BR><FONT=20 size=3D4><B>To:</B> </FONT><FONT color=3Dblue size=3D4><U><A=20 href=3D"mailto:adrian@olddog.co.uk">Adrian = Farrel</A></U></FONT><FONT=20 color=3Dblack size=3D4> ; </FONT><FONT color=3Dblue size=3D4><U><A = href=3D"mailto:per@defero.se">P=E4r Mattsson</A></U></FONT><FONT = color=3Dblack=20 size=3D4> ; </FONT><FONT color=3Dblue size=3D4><U><A = href=3D"mailto:loa@pi.se">Loa=20 Andersson</A></U></FONT><FONT color=3Dblack size=3D4> = </FONT><BR><FONT=20 size=3D4><B>Cc:</B> </FONT><FONT color=3Dblue size=3D4><U><A=20 = href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A></U></FONT><FONT= =20 color=3Dblack size=3D4> </FONT><BR><FONT size=3D4><B>Sent:</B> = Friday, July 22,=20 2005 10:23 AM</FONT><BR><FONT size=3D4><B>Subject:</B> RE: = Frameformat in a=20 l2cs gmpls rnvironment.</FONT><BR><BR>Hi,<BR><BR>I think Juergen = has=20 raised an important question. How frames are labeled (and = the=20 related data plane forwarding behavior) is not defined by the = control=20 plane. The control plane serves to provision the data plane, = not=20 define it. In the framework draft it is not clear what data = plane=20 standards are covered by the stated control plane = requirements. Some=20 references should be supplied. In any case, the labeling and = forwarding behavior should be defined by these referenced = standards,=20 not by GMPLS.<BR><BR>(I am assuming definition of new data = plane=20 standards is out of scope for CCAMP.)<BR><BR>Regards,<BR>Ben<BR> <HR align=3Dleft width=3D"99%" SIZE=3D4> <FONT color=3Dblack><BR><B>From:</B> owner-ccamp@ops.ietf.org=20 [mailto:owner-ccamp@ops.ietf.org] <B>On Behalf Of </B>Adrian=20 Farrel<BR><B>Sent:</B> Friday, July 22, 2005 8:37 AM<BR><B>To:</B> = P=E4r=20 Mattsson; Loa Andersson<BR><B>Cc:</B>=20 ccamp@ops.ietf.org<BR><B>Subject:</B> Re: Frameformat in a l2cs = gmpls=20 rnvironment.<BR><BR>Hi,<BR><BR>It is always refreshing how = engineers jump=20 straight to a discussion of the solution :-)<BR><BR>Perhaps we can = assume=20 that the framework draft is a good representation of the problem = space,=20 and that we are ready to start discussing the = solutions.<BR><BR>One=20 (perhaps the only?) significant question to answer is how the = frames will=20 be labelled. This question is one we must come to as soon as we = are=20 confident that the requirements need to be addressed at = all.<BR><BR>As has=20 been pointed out, there are several possibilities and to pick from = these=20 we need to understand:<BR>1. Do we need to support explicit label=20 stacking?<BR> Note that this is not supported in TDM, = LSC or=20 FSC.<BR>2. Do we need to be able to control (perhaps through=20 an<BR> external signaling controller) existing = hardware=20 and <BR> install LSPs through existing networks?<BR>3. = Do we=20 need to support existing function simultaneous to<BR> = the=20 support of GMPLS L2 LSPs?<BR><BR>I think that from a chair's = perspective I=20 can give some limited guidance.<BR><BR>A. These questions must be = raised=20 and answered in the framework I-D<BR>B. The answer to question 3 = is "yes".=20 This means that the use of labels<BR> must not = significantly=20 deplete any namespace used to support other<BR> =20 function.<BR>C. CCAMP is chartered to look at the control of = transport=20 networks. This<BR> includes Metro, but I am = unsure about=20 Campus. It does not cover <BR> signaling to the=20 desktop.<BR><BR>Thanks,<BR>Adrian<BR><BR>----- Original Message = -----=20 <BR>From: "P=E4r Mattsson" <</FONT><FONT color=3Dblue><U><A=20 href=3D"mailto:per@defero.se">per@defero.se</A></U></FONT><FONT=20 color=3Dblack>><BR>To: "Loa Andersson" <</FONT><FONT = color=3Dblue><U><A=20 href=3D"mailto:loa@pi.se">loa@pi.se</A></U></FONT><FONT=20 color=3Dblack>><BR>Cc: <</FONT><FONT color=3Dblue><U><A=20 = href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A></U></FONT><FONT= =20 color=3Dblack>><BR>Sent: Friday, July 22, 2005 11:41 = AM<BR>Subject: Re:=20 Frameformat in a l2cs gmpls rnvironment.<BR><BR><BR>> > Per = and=20 Dimitri,<BR>> ><BR>> > I would like to come down = stronger than=20 that, for me it is<BR>> > a very strong requirement that the = same=20 switch can handle<BR>> > both VLANs and GMPLs trafic = correctly. I=20 can't dsee how that<BR>> > could be done if using the VLAN = tpid to=20 indicate GMPLS<BR>> > traffic.<BR>> <BR>> If you ever = want=20 that same switch to handle traffic for a directly<BR>> = connected =20 host (not to uncommen) you would want that to use normal<BR>> = ethernet=20 macaddress switching. So of course you do not want to have = to<BR>>=20 choose between vlan and gmpls, you would want both at the same=20 time.<BR>> <BR>> /per<BR>> <BR>> <BR>> ><BR>> = >=20 /Loa<BR>> ><BR>> > P=E4r Mattsson wrote:<BR>> = >>>hi=20 par, one of the possibilities that has been considered to cope=20 with<BR>> >>>this requirement is to use a dedicated = TPID for=20 the Ethernet labeled<BR>> >>>frames; this would allow=20 differentiated processing with non-labeled<BR>>=20 >>>framesthanks.<BR>> >><BR>> = >><BR>>=20 >> That seems to make more sence. If that frame is to be = sized like=20 a<BR>> >> 802.1q<BR>> >> frame. There is not = that much=20 space left to a label. Or is the demand to<BR>> >> use=20 jumboframes ?<BR>> >> Has there been any discussion on=20 labelstacking, and mainly where to<BR>> >> place<BR>> = >>=20 the information?<BR>> >><BR>> >> = Regards.<BR>>=20 >> Per<BR>> >><BR>> >><BR>> = >><BR>>=20 ><BR>> ><BR>> > --<BR>> > Loa = Andersson<BR>>=20 ><BR>> > Principal Networking Architect<BR>> > = Acreo=20 = AB  = ; = =20 phone: +46 8 632 77 14<BR>> > Isafjordsgatan=20 = 22  = ; =20 mobile: +46 739 81 21 64<BR>> > Kista,=20 = Sweden &= nbsp; =20 email: </FONT><FONT color=3Dblue><U><A=20 = href=3D"mailto:loa.andersson@acreo.se">loa.andersson@acreo.se</A></U></FO= NT><FONT=20 color=3Dblack><BR>>=20 = > &nb= sp; &nbs= p;  = ; =20 </FONT><FONT color=3Dblue><U><A=20 href=3D"mailto:loa@pi.se">loa@pi.se</A></U></FONT><FONT = color=3Dblack><BR>>=20 ><BR>> <BR>> <BR>> <BR>> <BR><BR></FONT><FONT=20 face=3DMonospace,Courier=20 = size=3D4>=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D</FONT><BR><FONT=20 face=3DMonospace,Courier size=3D4>The information contained in = this message=20 may be privileged</FONT><BR><FONT face=3DMonospace,Courier = size=3D4>and=20 confidential and protected from disclosure. If the = reader</FONT><BR><FONT=20 face=3DMonospace,Courier size=3D4>of this message is not the = intended=20 recipient, or an employee</FONT><BR><FONT face=3DMonospace,Courier = size=3D4>or=20 agent responsible for delivering this message to = the</FONT><BR><FONT=20 face=3DMonospace,Courier size=3D4>intended recipient, you are = hereby notified=20 that any reproduction,</FONT><BR><FONT face=3DMonospace,Courier=20 size=3D4>dissemination or distribution of this communication is=20 strictly</FONT><BR><FONT face=3DMonospace,Courier = size=3D4>prohibited. If you=20 have received this communication in error,</FONT><BR><FONT=20 face=3DMonospace,Courier size=3D4>please notify us immediately by = replying to=20 the message and</FONT><BR><FONT face=3DMonospace,Courier = size=3D4>deleting it=20 from your computer. Thank you. Tellabs</FONT><BR><FONT=20 face=3DMonospace,Courier=20 = size=3D4>=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D</FONT><BR> = <P></P></BLOCKQUOTE></BLOCKQUOTE><PRE>=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any reproduction, dissemination or distribution of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. Tellabs =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D </PRE></BLOCKQUOTE></BODY></HTML> ------_=_NextPart_001_01C58F03.6D2AEF23-- Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 20:13:51 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01C58EF9.93F9EBAA" Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 21:11:43 +0100 Message-ID: <0536FC9B908BEC4597EE721BE6A353890A9F1E29@i2km07-ukbr.domain1.systemhost.net> Thread-Topic: Frameformat in a l2cs gmpls rnvironment. Thread-Index: AcWO1Vz5WDl4NyiSSoqZJ6lj4ozrFAAG4CmA From: <neil.2.harrison@bt.com> To: <Dimitri.Papadimitriou@alcatel.be>, <adrian@olddog.co.uk> Cc: <ccamp@ops.ietf.org> This is a multi-part message in MIME format. ------_=_NextPart_001_01C58EF9.93F9EBAA Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Dimitri, =20 =20 What do you mean by 'interferring elements'.....this sounds like some 'behind closed doors sect' at work. I thought Adrian's post was (as usual) plain/clear as to intent. =20 regards, Neil -----Original Message----- From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Dimitri.Papadimitriou@alcatel.be Sent: 22 July 2005 16:51 To: Adrian Farrel Cc: Dimitri.Papadimitriou@alcatel.be; ccamp@ops.ietf.org Subject: Re: Frameformat in a l2cs gmpls rnvironment. adrian - ok much thanks for the clarification -=20 i wanted to be sure that the working group consensus is going to be the element of decision for putting this item in the working list - and we won't have other interfering elements in the decision process -=20 "Adrian Farrel" <adrian@olddog.co.uk> Sent by: owner-ccamp@ops.ietf.org 07/22/2005 16:08 Please respond to "Adrian Farrel" To: Dimitri PAPADIMITRIOU/BE/ALCATEL@ALCATEL cc: <ccamp@ops.ietf.org> bcc:=20 Subject: Re: Frameformat in a l2cs gmpls rnvironment. Dimitri, > what do you EXACTLY mean with the following sentence > > "This question is one we must come to as soon as we are confident that the > requirements need to be addressed at all." > > please clarify - btw who is "we" in this context ? "We" means "the CCAMP working group" > this sentence reads to me as "CCAMP has not yet committed to work on this > topic" is my reading correct ? or do you mean something else with this ? Your reading is exactly correct as stated. I refer you to your email to the list on 19th July. You asked... > - if yes, is the interest in GMPLS for Ethernet in-/out-side the > CCAMP WG community strong enough to consider the corr. GMPLS > signaling and routing requirements (second part of the document) > as part of the CCAMP WG effort ? Thus, the purpose of the I-D was to crystalize the authors' thoughts on GMPLS control of Ethernet switching and to initiate discussion on whether the proposed deployment models are realistic and desirable. Following a positive answer and assuming support (by CCAMP) for the doing the work (in CCAMP) all that remains is to add the work to the charter (see previous charter discussions) and get on with it. Thanks, Adrian ------_=_NextPart_001_01C58EF9.93F9EBAA Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; = charset=3Dus-ascii"> <TITLE>Message</TITLE> <META content=3D"MSHTML 6.00.2800.1106" name=3DGENERATOR></HEAD> <BODY> <DIV><SPAN class=3D149210919-22072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2>Dimitri, </FONT></SPAN></DIV> <DIV><SPAN class=3D149210919-22072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2></FONT></SPAN> </DIV> <DIV><SPAN class=3D149210919-22072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2>What do you mean by 'interferring elements'.....this sounds = like some=20 'behind closed doors sect' at work. I thought Adrian's post was = (as usual)=20 plain/clear as to intent.</FONT></SPAN></DIV> <DIV><SPAN class=3D149210919-22072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2></FONT></SPAN> </DIV> <DIV><SPAN class=3D149210919-22072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2>regards, Neil</FONT></SPAN></DIV> <BLOCKQUOTE dir=3Dltr=20 style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #800000 2px = solid; MARGIN-RIGHT: 0px"> <DIV></DIV> <DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr = align=3Dleft><FONT=20 face=3DTahoma size=3D2>-----Original Message-----<BR><B>From:</B>=20 owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] <B>On = Behalf Of=20 </B>Dimitri.Papadimitriou@alcatel.be<BR><B>Sent:</B> 22 July 2005=20 16:51<BR><B>To:</B> Adrian Farrel<BR><B>Cc:</B>=20 Dimitri.Papadimitriou@alcatel.be; = ccamp@ops.ietf.org<BR><B>Subject:</B> Re:=20 Frameformat in a l2cs gmpls rnvironment.<BR><BR></FONT></DIV> <P><FONT face=3DMonospace,Courier>adrian - ok much thanks for the = clarification=20 - </FONT></P> <P><FONT face=3DMonospace,Courier>i wanted to be sure that the working = group=20 consensus is going to be the element of decision for putting this item = in the=20 working list - and we won't have other interfering elements in the = decision=20 process - </FONT><BR><BR><FONT size=3D2><B>"Adrian Farrel"=20 <adrian@olddog.co.uk></B></FONT><BR><FONT size=3D2>Sent by:=20 owner-ccamp@ops.ietf.org</FONT><BR><FONT size=3D2>07/22/2005=20 16:08</FONT><BR><FONT size=3D2>Please respond to "Adrian=20 Farrel"</FONT><BR><BR><FONT size=3D2>To:</FONT> <FONT size=3D2>Dimitri = PAPADIMITRIOU/BE/ALCATEL@ALCATEL</FONT><BR><FONT size=3D2>cc:</FONT> = <FONT=20 size=3D2><ccamp@ops.ietf.org></FONT><BR><FONT = size=3D2>bcc:</FONT> <BR><FONT=20 size=3D2>Subject:</FONT> <FONT size=3D2>Re: Frameformat in a l2cs = gmpls=20 rnvironment.</FONT><BR><BR><BR></P> <P><FONT face=3DMonospace,Courier>Dimitri,<BR></FONT><BR><FONT=20 face=3DMonospace,Courier>> what do you EXACTLY mean with the = following=20 sentence<BR>><BR>> "This question is one we must come to as soon = as we=20 are confident that<BR>the<BR>> requirements need to be addressed at = all."<BR>><BR>> please clarify - btw who is "we" in this context = ?<BR></FONT><BR><FONT face=3DMonospace,Courier>"We" means "the CCAMP = working=20 group"<BR></FONT><BR><FONT face=3DMonospace,Courier>> this sentence = reads to=20 me as "CCAMP has not yet committed to work on<BR>this<BR>> topic" = is my=20 reading correct ? or do you mean something else with this=20 ?<BR></FONT><BR><FONT face=3DMonospace,Courier>Your reading is exactly = correct=20 as stated.<BR>I refer you to your email to the list on 19th July. You=20 asked...<BR>> - if yes, is the interest in GMPLS for Ethernet = in-/out-side=20 the<BR>> CCAMP WG community strong enough to consider the = corr.=20 GMPLS<BR>> signaling and routing requirements (second part = of the=20 document)<BR>> as part of the CCAMP WG effort = ?<BR></FONT><BR><FONT=20 face=3DMonospace,Courier>Thus, the purpose of the I-D was to = crystalize the=20 authors' thoughts on<BR>GMPLS control of Ethernet switching and to = initiate=20 discussion on whether<BR>the proposed deployment models are realistic = and=20 desirable. Following a<BR>positive answer and assuming support (by = CCAMP) for=20 the doing the work (in<BR>CCAMP) all that remains is to add the work = to the=20 charter (see previous<BR>charter discussions) and get on with=20 it.<BR></FONT><BR><FONT=20 face=3DMonospace,Courier>Thanks,<BR>Adrian<BR></FONT></P></BLOCKQUOTE></B= ODY></HTML> ------_=_NextPart_001_01C58EF9.93F9EBAA-- Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 20:13:43 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01C58EF9.9504EFFA" Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 21:11:44 +0100 Message-ID: <0536FC9B908BEC4597EE721BE6A353890A9F1E2B@i2km07-ukbr.domain1.systemhost.net> Thread-Topic: Frameformat in a l2cs gmpls rnvironment. Thread-Index: AcWOz77GfzcCsQC6StanOHm687rebgAF3z2A From: <neil.2.harrison@bt.com> To: <ibryskin@movaz.com>, <Ben.Mack-Crane@tellabs.com>, <adrian@olddog.co.uk>, <per@defero.se>, <loa@pi.se> Cc: <ccamp@ops.ietf.org> This is a multi-part message in MIME format. ------_=_NextPart_001_01C58EF9.9504EFFA Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable And you are so right on this Igor. The nice thing about the co-cs mode = is that the data-plane has not been defined here and works great. = Adding a (forced) OOB control-plane is not a big deal.....in fact as we = move to the duct the control-plane is of much lesser interest/importance = (this is not to belittle it, its just a fact) Where the data-plane has = been defined however, for the co-ps mode in MPLS and the PW stuff its = unfortunately spawned, it's (how shall I put this politely) 'not so = great'. Can I make a plea NOT to screw-up ethernet by attempting to = replicate MPLS....in any case the ethernet data-plane belongs to IEEE. = I would advise folks here to leave this well alone. =20 regards, Neil -----Original Message----- From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On = Behalf Of Igor Bryskin Sent: 22 July 2005 16:11 To: Mack-Crane, T. Benjamin; Adrian Farrel; P=E4r Mattsson; Loa = Andersson Cc: ccamp@ops.ietf.org Subject: Re: Frameformat in a l2cs gmpls rnvironment. Hi, =20 I think Ben and Juergen have a point here. There is nothing that could = be dynamically provisioned that could not be also provisioned by = management. In other words data plane is the King here. We need data = plane standard(s) on how we encode a label, whether we need a label = stack or not, how the labeled traffic is supposed to be treated, how = labeled traffic co-exist with non-label traffic, etc. This is something = that not for CCAMP to define. Take for example TDM networks. GMPLS only = provides a way to dynamically provision G.707 networks. Hence there is a = need in parallel standardization activities in ITU. Only after that we = can discuss how all that could be dynamically provisioned, that is the = aspects of control plane.=20 =20 Igor ----- Original Message -----=20 From: Mack-Crane, T. Benjamin <mailto:Ben.Mack-Crane@tellabs.com> =20 To: Adrian Farrel <mailto:adrian@olddog.co.uk> ; P=E4r Mattsson = <mailto:per@defero.se> ; Loa Andersson <mailto:loa@pi.se> =20 Cc: ccamp@ops.ietf.org=20 Sent: Friday, July 22, 2005 10:23 AM Subject: RE: Frameformat in a l2cs gmpls rnvironment. Hi, =20 I think Juergen has raised an important question. How frames are = labeled (and the related data plane forwarding behavior) is not defined = by the control plane. The control plane serves to provision the data = plane, not define it. In the framework draft it is not clear what data = plane standards are covered by the stated control plane requirements. = Some references should be supplied. In any case, the labeling and = forwarding behavior should be defined by these referenced standards, not = by GMPLS. =20 (I am assuming definition of new data plane standards is out of scope = for CCAMP.) =20 Regards, Ben _____ =20 From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On = Behalf Of Adrian Farrel Sent: Friday, July 22, 2005 8:37 AM To: P=E4r Mattsson; Loa Andersson Cc: ccamp@ops.ietf.org Subject: Re: Frameformat in a l2cs gmpls rnvironment. Hi, =20 It is always refreshing how engineers jump straight to a discussion of = the solution :-) =20 Perhaps we can assume that the framework draft is a good representation = of the problem space, and that we are ready to start discussing the = solutions. =20 One (perhaps the only?) significant question to answer is how the frames = will be labelled. This question is one we must come to as soon as we are = confident that the requirements need to be addressed at all. =20 As has been pointed out, there are several possibilities and to pick = from these we need to understand: 1. Do we need to support explicit label stacking? Note that this is not supported in TDM, LSC or FSC. 2. Do we need to be able to control (perhaps through an external signaling controller) existing hardware and=20 install LSPs through existing networks? 3. Do we need to support existing function simultaneous to the support of GMPLS L2 LSPs? =20 I think that from a chair's perspective I can give some limited = guidance. =20 A. These questions must be raised and answered in the framework I-D B. The answer to question 3 is "yes". This means that the use of labels must not significantly deplete any namespace used to support other function. C. CCAMP is chartered to look at the control of transport networks. This includes Metro, but I am unsure about Campus. It does not cover=20 signaling to the desktop. =20 Thanks, Adrian =20 ----- Original Message -----=20 From: "P=E4r Mattsson" < <mailto:per@defero.se> per@defero.se> To: "Loa Andersson" < <mailto:loa@pi.se> loa@pi.se> Cc: < <mailto:ccamp@ops.ietf.org> ccamp@ops.ietf.org> Sent: Friday, July 22, 2005 11:41 AM Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > Per and Dimitri, > > > > I would like to come down stronger than that, for me it is > > a very strong requirement that the same switch can handle > > both VLANs and GMPLs trafic correctly. I can't dsee how that > > could be done if using the VLAN tpid to indicate GMPLS > > traffic. >=20 > If you ever want that same switch to handle traffic for a directly > connected host (not to uncommen) you would want that to use normal > ethernet macaddress switching. So of course you do not want to have to > choose between vlan and gmpls, you would want both at the same time. >=20 > /per >=20 >=20 > > > > /Loa > > > > P=E4r Mattsson wrote: > >>>hi par, one of the possibilities that has been considered to cope = with > >>>this requirement is to use a dedicated TPID for the Ethernet = labeled > >>>frames; this would allow differentiated processing with non-labeled > >>>framesthanks. > >> > >> > >> That seems to make more sence. If that frame is to be sized like a > >> 802.1q > >> frame. There is not that much space left to a label. Or is the = demand to > >> use jumboframes ? > >> Has there been any discussion on labelstacking, and mainly where to > >> place > >> the information? > >> > >> Regards. > >> Per > >> > >> > >> > > > > > > -- > > Loa Andersson > > > > Principal Networking Architect > > Acreo AB phone: +46 8 632 77 14 > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > Kista, Sweden email: = <mailto:loa.andersson@acreo.se> loa.andersson@acreo.se > > <mailto:loa@pi.se> = loa@pi.se > > >=20 >=20 >=20 >=20 =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any reproduction, dissemination or distribution of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. Tellabs =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D ------_=_NextPart_001_01C58EF9.9504EFFA Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; = charset=3Diso-8859-1"> <TITLE>Message</TITLE> <META content=3D"MSHTML 6.00.2800.1106" name=3DGENERATOR> <STYLE></STYLE> </HEAD> <BODY bgColor=3D#ffffff> <DIV><SPAN class=3D100240018-22072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2>And you are so right on this Igor. The nice thing about = the co-cs=20 mode is that the data-plane has not been defined here and works = great. =20 Adding a (forced) OOB control-plane is not a big deal.....in fact as we = move to=20 the duct the control-plane is of much lesser interest/importance (this = is not to=20 belittle it, its just a fact) Where the data-plane has been = defined=20 however, for the co-ps mode in MPLS and the PW stuff its unfortunately = spawned,=20 it's (how shall I put this politely) 'not so great'. Can I make a = plea NOT=20 to screw-up ethernet by attempting to replicate MPLS....in any case the = ethernet=20 data-plane belongs to IEEE. I would advise folks here to leave = this well=20 alone.</FONT></SPAN></DIV> <DIV><SPAN class=3D100240018-22072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2></FONT></SPAN> </DIV> <DIV><SPAN class=3D100240018-22072005><FONT face=3D"Comic Sans MS" = color=3D#800000=20 size=3D2>regards, Neil</FONT></SPAN></DIV> <BLOCKQUOTE dir=3Dltr=20 style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #800000 2px = solid; MARGIN-RIGHT: 0px"> <DIV></DIV> <DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr = align=3Dleft><FONT=20 face=3DTahoma size=3D2>-----Original Message-----<BR><B>From:</B>=20 owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] <B>On = Behalf Of=20 </B>Igor Bryskin<BR><B>Sent:</B> 22 July 2005 16:11<BR><B>To:</B> = Mack-Crane,=20 T. Benjamin; Adrian Farrel; P=E4r Mattsson; Loa = Andersson<BR><B>Cc:</B>=20 ccamp@ops.ietf.org<BR><B>Subject:</B> Re: Frameformat in a l2cs gmpls=20 rnvironment.<BR><BR></FONT></DIV> <DIV><FONT face=3DArial size=3D2>Hi,</FONT></DIV> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <DIV><FONT face=3DArial size=3D2>I think Ben and Juergen have a point = here. There=20 is nothing that could be dynamically provisioned that could not be = also=20 provisioned by management. In other words data plane is the King here. = We need=20 data plane standard(s) on how we encode a label, whether we need a = label stack=20 or not, how the labeled traffic is supposed to be treated, how = labeled=20 traffic co-exist with non-label traffic, etc. This is something that = not for=20 CCAMP to define. Take for example TDM networks. GMPLS only provides a = way to=20 dynamically provision G.707 networks. Hence there is a need in = parallel=20 standardization activities in ITU. Only after that we can discuss how = all that=20 could be dynamically provisioned, that is the aspects of control = plane.=20 </FONT></DIV> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <DIV><FONT face=3DArial size=3D2>Igor</FONT></DIV> <BLOCKQUOTE dir=3Dltr=20 style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; = BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px"> <DIV style=3D"FONT: 10pt arial">----- Original Message ----- </DIV> <DIV=20 style=3D"BACKGROUND: #e4e4e4; FONT: 10pt arial; font-color: = black"><B>From:</B>=20 <A title=3DBen.Mack-Crane@tellabs.com=20 href=3D"mailto:Ben.Mack-Crane@tellabs.com">Mack-Crane, T. = Benjamin</A> </DIV> <DIV style=3D"FONT: 10pt arial"><B>To:</B> <A = title=3Dadrian@olddog.co.uk=20 href=3D"mailto:adrian@olddog.co.uk">Adrian Farrel</A> ; <A = title=3Dper@defero.se=20 href=3D"mailto:per@defero.se">P=E4r Mattsson</A> ; <A = title=3Dloa@pi.se=20 href=3D"mailto:loa@pi.se">Loa Andersson</A> </DIV> <DIV style=3D"FONT: 10pt arial"><B>Cc:</B> <A = title=3Dccamp@ops.ietf.org=20 href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A> </DIV> <DIV style=3D"FONT: 10pt arial"><B>Sent:</B> Friday, July 22, 2005 = 10:23=20 AM</DIV> <DIV style=3D"FONT: 10pt arial"><B>Subject:</B> RE: Frameformat in a = l2cs=20 gmpls rnvironment.</DIV> <DIV><FONT face=3DArial size=3D2></FONT><BR></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>Hi,</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2></FONT></SPAN> </DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>I think Juergen has raised an important = question. =20 How frames are labeled (and the related data plane forwarding = behavior) is=20 not defined by the control plane. The control plane serves to=20 provision the data plane, not define it. In the framework = draft it is=20 not clear what data plane standards are covered by the stated = control plane=20 requirements. Some references should be supplied. In any = case,=20 the labeling and forwarding behavior should be defined by these = referenced standards, not by GMPLS.</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2></FONT></SPAN> </DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>(I am assuming definition of new data plane = standards=20 is out of scope for CCAMP.)</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2></FONT></SPAN> </DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>Regards,</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>Ben</FONT></SPAN></DIV><BR> <BLOCKQUOTE=20 style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #0000ff = 2px solid; MARGIN-RIGHT: 0px"> <DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr = align=3Dleft> <HR tabIndex=3D-1> <FONT face=3DTahoma size=3D2><B>From:</B> owner-ccamp@ops.ietf.org = [mailto:owner-ccamp@ops.ietf.org] <B>On Behalf Of </B>Adrian=20 Farrel<BR><B>Sent:</B> Friday, July 22, 2005 8:37 AM<BR><B>To:</B> = P=E4r=20 Mattsson; Loa Andersson<BR><B>Cc:</B>=20 ccamp@ops.ietf.org<BR><B>Subject:</B> Re: Frameformat in a l2cs = gmpls=20 rnvironment.<BR></FONT><BR></DIV> <DIV></DIV> <DIV><FONT face=3DCourier size=3D2>Hi,</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>It is always refreshing how = engineers jump=20 straight to a discussion of the solution :-)</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Perhaps we can assume that the = framework=20 draft is a good representation of the problem space, and that we = are ready=20 to start discussing the solutions.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>One (perhaps the only?) = significant=20 question to answer is how the frames will be labelled. This = question is=20 one we must come to as soon as we are confident that the = requirements need=20 to be addressed at all.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>As has been pointed out, there = are several=20 possibilities and to pick from these we need to = understand:</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>1. Do we need to support = explicit label=20 stacking?</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> Note that this is = not=20 supported in TDM, LSC or FSC.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>2. Do we need to be able to = control=20 (perhaps through an</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> external = signaling=20 controller) existing hardware and </FONT></DIV> <DIV><FONT face=3DCourier size=3D2> install LSPs = through existing=20 networks?</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>3. Do we need to support = existing function=20 simultaneous to</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> the support of = GMPLS L2=20 LSPs?</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>I think that from a chair's = perspective I=20 can give some limited guidance.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>A. These questions must be = raised and=20 answered in the framework I-D</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>B. The answer to question 3 is = "yes". This=20 means that the use of labels</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> must not = significantly deplete=20 any namespace used to support other</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> = function.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>C. CCAMP is chartered to look = at the=20 control of transport networks. This</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> includes = Metro, but I am=20 unsure about Campus. It does not cover </FONT></DIV> <DIV><FONT face=3DCourier size=3D2> signaling to the=20 desktop.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Thanks,</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Adrian</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>----- Original Message ----- = </FONT> <DIV><FONT face=3DCourier size=3D2>From: "P=E4r Mattsson" = <</FONT><A=20 href=3D"mailto:per@defero.se"><FONT face=3DCourier=20 size=3D2>per@defero.se</FONT></A><FONT face=3DCourier = size=3D2>></FONT></DIV> <DIV><FONT face=3DCourier size=3D2>To: "Loa Andersson" = <</FONT><A=20 href=3D"mailto:loa@pi.se"><FONT face=3DCourier=20 size=3D2>loa@pi.se</FONT></A><FONT face=3DCourier = size=3D2>></FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Cc: <</FONT><A=20 href=3D"mailto:ccamp@ops.ietf.org"><FONT face=3DCourier=20 size=3D2>ccamp@ops.ietf.org</FONT></A><FONT face=3DCourier=20 size=3D2>></FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Sent: Friday, July 22, 2005 = 11:41=20 AM</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Subject: Re: Frameformat in a = l2cs gmpls=20 rnvironment.</FONT></DIV></DIV> <DIV><FONT face=3DCourier><BR><FONT = size=3D2></FONT></FONT></DIV><FONT=20 face=3DCourier size=3D2>> > Per and Dimitri,<BR>> = ><BR>> > I=20 would like to come down stronger than that, for me it is<BR>> = > a=20 very strong requirement that the same switch can handle<BR>> = > both=20 VLANs and GMPLs trafic correctly. I can't dsee how that<BR>> = > could=20 be done if using the VLAN tpid to indicate GMPLS<BR>> >=20 traffic.<BR>> <BR>> If you ever want that same switch to = handle=20 traffic for a directly<BR>> connected host (not to = uncommen) you=20 would want that to use normal<BR>> ethernet macaddress = switching. So of=20 course you do not want to have to<BR>> choose between vlan and = gmpls,=20 you would want both at the same time.<BR>> <BR>> = /per<BR>>=20 <BR>> <BR>> ><BR>> > /Loa<BR>> ><BR>> > = P=E4r=20 Mattsson wrote:<BR>> >>>hi par, one of the = possibilities that=20 has been considered to cope with<BR>> >>>this = requirement is=20 to use a dedicated TPID for the Ethernet labeled<BR>>=20 >>>frames; this would allow differentiated processing = with=20 non-labeled<BR>> >>>framesthanks.<BR>> = >><BR>>=20 >><BR>> >> That seems to make more sence. If that = frame is=20 to be sized like a<BR>> >> 802.1q<BR>> >> frame. = There=20 is not that much space left to a label. Or is the demand = to<BR>>=20 >> use jumboframes ?<BR>> >> Has there been any = discussion=20 on labelstacking, and mainly where to<BR>> >> = place<BR>>=20 >> the information?<BR>> >><BR>> >>=20 Regards.<BR>> >> Per<BR>> >><BR>> = >><BR>>=20 >><BR>> ><BR>> ><BR>> > --<BR>> > = Loa=20 Andersson<BR>> ><BR>> > Principal Networking = Architect<BR>>=20 > Acreo=20 = AB  = ; = =20 phone: +46 8 632 77 14<BR>> > Isafjordsgatan=20 = 22  = ; =20 mobile: +46 739 81 21 64<BR>> > Kista,=20 = Sweden &= nbsp; =20 email: </FONT><A = href=3D"mailto:loa.andersson@acreo.se"><FONT=20 face=3DCourier size=3D2>loa.andersson@acreo.se</FONT></A><BR><FONT = face=3DCourier size=3D2>>=20 = > &nb= sp; &nbs= p;  = ; =20 </FONT><A href=3D"mailto:loa@pi.se"><FONT face=3DCourier=20 size=3D2>loa@pi.se</FONT></A><BR><FONT face=3DCourier = size=3D2>> ><BR>>=20 <BR>> <BR>> <BR>> = </FONT></BLOCKQUOTE><PRE>=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any reproduction, dissemination or distribution of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. Tellabs =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D </PRE></BLOCKQUOTE></BLOCKQUOTE></BODY></HTML> ------_=_NextPart_001_01C58EF9.9504EFFA-- Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 20:13:42 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 21:11:45 +0100 Message-ID: <0536FC9B908BEC4597EE721BE6A353890A9F1E2C@i2km07-ukbr.domain1.systemhost.net> Thread-Topic: Frameformat in a l2cs gmpls rnvironment. Thread-Index: AcWO0JK8kM7ppvqMRQihwOQp5MS+/AAGG3NA From: <neil.2.harrison@bt.com> To: <adrian@olddog.co.uk>, <Ben.Mack-Crane@tellabs.com> Cc: <ccamp@ops.ietf.org> Adrian....with an operators hat on ;-) Yes, you are right on most points here....few remarks in-(snipped)line: > > BMC=3D> How frames are labeled (and the related data plane = forwarding > > behavior) > is > > not defined by the control plane. >=20 > AF=3D> It is true to say that this is not defined by the control = plane.=20 > Nevertheless, the control plane passes labels around and it is=20 > important to define how those labels will be interpreted. I refer you=20 > to all of the non-packet labeling mechanisms already defined: you will > see that they do not define how the data is labeled, but they do=20 > define how the signaled label maps to the labeled data. NH=3D> Largely right Adrian....but on a wider note, I wish (and goodness knows how I wish, esp in MPLS/PWE3) we had a stronger understanding of func arch and the 3 networking modes. It is the very nature of the co-cs mode's 'label' that is key here, ie folks can't abuse this mode as the nature of label means it's self-protecting to a large degree (I'll ignore the 'topology on the fly' folks who quite clearly live in their own world). And remember a label is something that affects the client->server layer adaptation, BUT it is essentially owned/defined by the server layer and not the client. But all this labelling stuff is traffic data-plane oriented and has nothing to do with the control-plane. As Igor put it recently, the data-plane is King....and indeed its is, the control-plane is merely there to serve it. The role/significance of the control plane itself also changes in the 3 modes. In the cl-ps mode it is 'almost everything' (well distributed routing is THE issue here) but when we get down to SDH/OTNs and the co-cs mode its not so important. Here it's the management-plane that dominates everything, even IF there is a control-plane...and then its really mainly signalling. It's the co-ps mode that is the odd-man out. Let me point out that MPLS DOES define a new data-plane (so its not like GMPLS at all)....and its from an examination of THAT data-plane structure/behaviour that one uncovers a host of problems. I'll spare you a lengthy discourse on the topic but I think/hope most folks should grasp what I mean by this. >=20 > > BMC=3D>The control plane serves to provision the data plane, not=20 > > define it. >=20 > AF=3D>Yes. NH=3D> Exactly. >=20 > > BMC=3D>In the framework draft it is not clear what data plane > standards are > > covered by the stated control plane requirements. Some references > > should be supplied. >=20 > AF=3D>Such text would not be welcomed by the chairs who > specifically instructed it to be left out. NH=3D> They are wise IMO....don't screw-up the data-plane. Take a leaf from where this has been defined for the co-ps mode (ie MPLS/PWs)......I don't think the results are that good (understatement). > > BMC=3D>(I am assuming definition of new data plane standards is=20 > out of scope > for > > CCAMP.) >=20 > AF=3D>Are you? >=20 > I would be surprised if this work introduced new data plane standards. NH=3D> I would be horrified based on past experience....leave the data-plane which has been defined elsewhere alone please. regards, Neil >=20 > Thanks, > Adrian >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 19:02:18 +0000 To: "Mack-Crane, T. Benjamin" <Ben.Mack-Crane@tellabs.com> Cc: "Igor Bryskin" <ibryskin@movaz.com>, <Dimitri.Papadimitriou@alcatel.be>, "Adrian Farrel" <adrian@olddog.co.uk>, =?iso-8859-1?Q?P=E4r_Mattsson?= <per@defero.se>, "Loa Andersson" <loa@pi.se>, <ccamp@ops.ietf.org> MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 21:01:02 +0200 Message-ID: <OFADF90D64.2D5D897A-ONC1257046.00687692-C1257046.0068774F@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmFjdHVhbGx5IGkgZG9uJ3QgdGhpbmsg eW91IGhhdmUgZnVsbHkgdW5kZXJzdG9vZCB0aGUgcHVycG9zZSBvZiB0aGlzIGRvY3VtZW50PC9G T05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+c28gaSByZXBlYXQgYXNz dW1pbmcgdGhlcmUgaXMgbm8gb3IgdW5kZWZpbmVkIGRhdGEgcGxhbmUgZGVmaW5pdGlvbiBpcyBu b3QgY29ycmVjdCB3aGF0IHdlIGFyZSBkaXNjdXNzaW5nIGlzIHRoZSByb2xlIG9mIHRoZSBHTVBM UyBjb250cm9sIHBsYW5lIGluIGV4Y2hhbmdpbmcgdGhlIHJlcXVpcmVkIGZvcndhcmRpbmcgaW5m b3JtYXRpb24gZWxlbWVudHMsIHRoZSBxdWVzdGlvbiBpcyB0aHVzIHRvIGlkZW50aWZ5IHdoaWNo IGRhdGEgcGxhbmUgYmVoYXZpb3VyIHdlIHdvdWxkIGxpa2UgR01QTFMgdG8gY292ZXIgaW4gb3Jk ZXIgdG8gYWRkcmVzcyB0aGUgc3RyaW5nZW50IHJlcXVpcmVtZW50cyBhbmQgdGhlIHNjZW5hcmlv cyB0aGF0IGhhdmUgYmVlbiBwcm9kdWNlZCBpbiB0aGUgcHJvYmxlbSBzdGF0ZW1lbnQgLSA8L0ZP TlQ+PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj50byBiZSBldmVuIGNsZWFy ZXI6IHRoZSBkZWZpbml0aW9uIG9mIGFueSBkYXRhIHBsYW5lIGJlaGF2aW91ciBpcyBvdXRzaWRl IG9mIHRoZSBzY29wZSBvZiB0aGUgRFQgY2hhcnRlciBhbmQgdGhlIGRvY3VtZW50IGl0IHByb2R1 Y2VkIC08QlI+PC9GT05UPjxCUj48Rk9OVCBTSVpFPTI+PEI+JnF1b3Q7TWFjay1DcmFuZSwgVC4g QmVuamFtaW4mcXVvdDsgJmx0O0Jlbi5NYWNrLUNyYW5lQHRlbGxhYnMuY29tJmd0OzwvQj48L0ZP TlQ+PEJSPjxGT05UIFNJWkU9Mj4wNy8yMi8yMDA1IDEzOjI1IEVTVDwvRk9OVD48QlI+PEJSPiA8 Rk9OVCBTSVpFPTI+VG86PC9GT05UPiA8Rk9OVCBTSVpFPTI+JnF1b3Q7SWdvciBCcnlza2luJnF1 b3Q7ICZsdDtpYnJ5c2tpbkBtb3Zhei5jb20mZ3Q7LCBEaW1pdHJpIFBBUEFESU1JVFJJT1UvQkUv QUxDQVRFTEBBTENBVEVMPC9GT05UPjxCUj4gPEZPTlQgU0laRT0yPmNjOjwvRk9OVD4gPEZPTlQg U0laRT0yPiZxdW90O0FkcmlhbiBGYXJyZWwmcXVvdDsgJmx0O2FkcmlhbkBvbGRkb2cuY28udWsm Z3Q7LCBQJmF1bWw7ciBNYXR0c3NvbiAmbHQ7cGVyQGRlZmVyby5zZSZndDssICZxdW90O0xvYSBB bmRlcnNzb24mcXVvdDsgJmx0O2xvYUBwaS5zZSZndDssICZsdDtjY2FtcEBvcHMuaWV0Zi5vcmcm Z3Q7PC9GT05UPjxCUj4gPEZPTlQgU0laRT0yPmJjYzo8L0ZPTlQ+IDxCUj4gPEZPTlQgU0laRT0y PlN1YmplY3Q6PC9GT05UPiA8Rk9OVCBTSVpFPTI+UkU6IEZyYW1lZm9ybWF0IGluIGEgbDJjcyBn bXBscyBybnZpcm9ubWVudC48L0ZPTlQ+PEJSPiA8QlI+PEJSPjwvUD48UD5JZ29yLDxCUj48QlI+ SSwgdG9vLCBhbSBiYWZmbGVkIGJ5IHRoZSBub3Rpb24gb2YgcmVxdWlyZW1lbnRzIGZvciBjb250 cm9sIG9mIGFuIHVuc3BlY2lmaWVkIGRhdGEgcGxhbmUuPEJSPjxCUj5SZWdhcmRzLDxCUj5CZW48 QlI+PEhSIFdJRFRIPSI5OSUiIFNJWkU9IjQiIEFMSUdOPSJsZWZ0Ij48Rk9OVCBDT0xPUj1CTEFD Sz48QlI+PEI+RnJvbTo8L0I+IElnb3IgQnJ5c2tpbiBbbWFpbHRvOmlicnlza2luQG1vdmF6LmNv bV0gPEJSPjxCPlNlbnQ6PC9CPiBGcmlkYXksIEp1bHkgMjIsIDIwMDUgMTI6MjAgUE08QlI+PEI+ VG86PC9CPiBEaW1pdHJpLlBhcGFkaW1pdHJpb3VAYWxjYXRlbC5iZTxCUj48Qj5DYzo8L0I+IE1h Y2stQ3JhbmUsIFQuIEJlbmphbWluOyBBZHJpYW4gRmFycmVsOyBQJmF1bWw7ciBNYXR0c3Nvbjsg TG9hIEFuZGVyc3NvbjsgY2NhbXBAb3BzLmlldGYub3JnPEJSPjxCPlN1YmplY3Q6PC9CPiBSZTog RnJhbWVmb3JtYXQgaW4gYSBsMmNzIGdtcGxzIHJudmlyb25tZW50LjxCUj48QlI+RGltaXRyaSw8 QlI+PEJSPlRoZSBxdWVzdGlvbiBpcyBzaW1wbGU6IGlzIGl0IHBvc3NpYmxlIHRvZGF5IHRvIHN0 YXRpY2FsbHkgcHJvdmlzaW9uIEwyTFNQcyB0aGF0IGNvdWxkLCBzYXksJm5ic3A7c3VwcG9ydCBl MmUgUW9TPyBJZiBub3QsIG1heSBiZSBpdCBpcyB0byBlYXJseSB0byBkaXNjdXNzIGFzcGVjdHMg b2YgZHluYW1pYyBwcm92aXNpb25pbmcgb2Ygc3VjaCBMU1BzPyBJcyBpdCBwb3NzaWJsZS9yZWFz b25hYmxlLCBpbiB5b3VyIG9waW5pb24sIHRvICZxdW90OzwvRk9OVD48Rk9OVCBTSVpFPTQ+ZGV0 YWlsIGhvdyBmb3J3YXJkaW5nIGluZm9ybWF0aW9uIGNhbiBiZSBleGNoYW5nZWQgdmlhIHRoZSBj b250cm9sIHBsYW5lIChhbmQgdGhlbiBpbnN0YWxsZWQpJnF1b3Q7IHdpdGhvdXQmbmJzcDtoYXZp bmcgaW4gbWluZCAmcXVvdDthIHNwZWNpZmljIGZvcndhcmRpbmcgYmVoYXZpb3ImcXVvdDs/PC9G T05UPjxCUj48Rk9OVCBTSVpFPTQ+SG93IGRvIHlvdSBrbm93IHdoaWNoIGZvcndhcmRpbmcgaW5m b3JtYXRpb24gaXMgbmVlZGVkIGZvciB0aGUgZm9yd2FyZGluZyBub2JvZHkgaGFzIGRlZmluZWQg eWV0PzwvRk9OVD48QlI+PEZPTlQgU0laRT00PlRoZSBvdGhlciBxdWVzdGlvbiBpcyZuYnNwO2Rv IHdlIG5lZWQgYXQgYWxsIGUyZSBRb1MsIHJvdXRlIGNvbnRyb2wsIGZhc3QgcmVjb3Zlcnk/IFNv dW5kcyBsaWtlIGV4Y2l0aW5nIGlkZWEsIGJ1dCBkb2VzIGFsbCB0aGF0IGNvbW1lcmNpYWxseSBt YWtlIHNlbnNlPyBTbywgSSBndWVzcywgd2UgbmVlZCB0d28gdGhpbmdzJm5ic3A7YmVmb3JlIHdl IGNhbiZuYnNwO21vdmUgZm9yd2FyZDo8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND5hKSB2YWxpZGF0 aW9uIG9mIHRoZSBpZGVhIGJ5IHRoZSBwcm92aWRlcnM7PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+ YikgZGVmaW5pdGlvbiBvZiB0aGUgZGF0YSBwbGFuZSBiZWhhdmlvcjs8L0ZPTlQ+PEJSPjxCUj48 Rk9OVCBTSVpFPTQ+SWdvcjwvRk9OVD48QlI+PEZPTlQgU0laRT00Pi0tLS0tIE9yaWdpbmFsIE1l c3NhZ2UgLS0tLS0gPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+PEI+RnJvbTo8L0I+IDwvRk9OVD48 Rk9OVCBTSVpFPTQgQ09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0bzpEaW1pdHJpLlBhcGFkaW1p dHJpb3VAYWxjYXRlbC5iZT5EaW1pdHJpLlBhcGFkaW1pdHJpb3VAYWxjYXRlbC5iZTwvQT48L1U+ PC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTEFDSz4gPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+ PEI+VG86PC9CPiA8L0ZPTlQ+PEZPTlQgU0laRT00IENPTE9SPUJMVUU+PFU+PEEgSFJFRj1tYWls dG86aWJyeXNraW5AbW92YXouY29tPklnb3IgQnJ5c2tpbjwvQT48L1U+PC9GT05UPjxGT05UIFNJ WkU9NCBDT0xPUj1CTEFDSz4gPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+PEI+Q2M6PC9CPiA8L0ZP TlQ+PEZPTlQgU0laRT00IENPTE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86QmVuLk1hY2stQ3Jh bmVAdGVsbGFicy5jb20+TWFjay1DcmFuZSwgVC4gQmVuamFtaW48L0E+PC9VPjwvRk9OVD48Rk9O VCBTSVpFPTQgQ09MT1I9QkxBQ0s+IDsgPC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTFVFPjxV PjxBIEhSRUY9bWFpbHRvOmFkcmlhbkBvbGRkb2cuY28udWs+QWRyaWFuIEZhcnJlbDwvQT48L1U+ PC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTEFDSz4gOyA8L0ZPTlQ+PEZPTlQgU0laRT00IENP TE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86cGVyQGRlZmVyby5zZT5QJmF1bWw7ciBNYXR0c3Nv bjwvQT48L1U+PC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTEFDSz4gOyA8L0ZPTlQ+PEZPTlQg U0laRT00IENPTE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86bG9hQHBpLnNlPkxvYSBBbmRlcnNz b248L0E+PC9VPjwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxBQ0s+IDsgPC9GT05UPjxGT05U IFNJWkU9NCBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOmNjYW1wQG9wcy5pZXRmLm9yZz5j Y2FtcEBvcHMuaWV0Zi5vcmc8L0E+PC9VPjwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxBQ0s+ IDwvRk9OVD48QlI+PEZPTlQgU0laRT00PjxCPlNlbnQ6PC9CPiBGcmlkYXksIEp1bHkgMjIsIDIw MDUgMTI6MjEgUE08L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND48Qj5TdWJqZWN0OjwvQj4gUmU6IEZy YW1lZm9ybWF0IGluIGEgbDJjcyBnbXBscyBybnZpcm9ubWVudC48L0ZPTlQ+PEJSPjxCUj48QlI+ PEZPTlQgU0laRT00Pmlnb3IgLSBpIGFtIG5vdCBzdXJlIHdoaWNoIHBvaW50IHRoZXkgZXhhY3Rs eSBoYXZlIC0gdGhlIHB1cnBvc2Ugb2YgdGhpcyBkb2N1bWVudCBpcyB0byBkZXRhaWwgaG93IGZv cndhcmRpbmcgaW5mb3JtYXRpb24gY2FuIGJlIGV4Y2hhbmdlZCB2aWEgdGhlIGNvbnRyb2wgcGxh bmUgKGFuZCB0aGVuIGluc3RhbGxlZCkgbm90IHRvIGRlZmluZSBhIHNwZWNpZmljIGZvcndhcmRp bmcgYmVoYXZpb3VyPC9GT05UPjxCUj48QlI+PEZPTlQgU0laRT00PnRoYW5rcyw8L0ZPTlQ+PEJS PjxCUj48Rk9OVCBTSVpFPTQ+LSBkaW1pdHJpLjwvRk9OVD48QlI+PEJSPjxCPiZxdW90O0lnb3Ig QnJ5c2tpbiZxdW90OyAmbHQ7PC9CPjxGT05UIENPTE9SPUJMVUU+PEI+PFU+PEEgSFJFRj1tYWls dG86aWJyeXNraW5AbW92YXouY29tPmlicnlza2luQG1vdmF6LmNvbTwvQT48L1U+PC9CPjwvRk9O VD48Rk9OVCBDT0xPUj1CTEFDSz48Qj4mZ3Q7PC9CPjxCUj5TZW50IGJ5OiA8L0ZPTlQ+PEZPTlQg Q09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0bzpvd25lci1jY2FtcEBvcHMuaWV0Zi5vcmc+b3du ZXItY2NhbXBAb3BzLmlldGYub3JnPC9BPjwvVT48L0ZPTlQ+PEZPTlQgQ09MT1I9QkxBQ0s+PEJS PjA3LzIyLzIwMDUgMTE6MTAgQVNUPEJSPjxCUj5Ubzo8L0ZPTlQ+PEZPTlQgU0laRT00PiA8L0ZP TlQ+JnF1b3Q7TWFjay1DcmFuZSwgVC4gQmVuamFtaW4mcXVvdDsgJmx0OzxGT05UIENPTE9SPUJM VUU+PFU+PEEgSFJFRj1tYWlsdG86QmVuLk1hY2stQ3JhbmVAdGVsbGFicy5jb20+QmVuLk1hY2st Q3JhbmVAdGVsbGFicy5jb208L0E+PC9VPjwvRk9OVD48Rk9OVCBDT0xPUj1CTEFDSz4mZ3Q7LCAm cXVvdDtBZHJpYW4gRmFycmVsJnF1b3Q7ICZsdDs8L0ZPTlQ+PEZPTlQgQ09MT1I9QkxVRT48VT48 QSBIUkVGPW1haWx0bzphZHJpYW5Ab2xkZG9nLmNvLnVrPmFkcmlhbkBvbGRkb2cuY28udWs8L0E+ PC9VPjwvRk9OVD48Rk9OVCBDT0xPUj1CTEFDSz4mZ3Q7LCBQJmF1bWw7ciBNYXR0c3NvbiAmbHQ7 PC9GT05UPjxGT05UIENPTE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86cGVyQGRlZmVyby5zZT5w ZXJAZGVmZXJvLnNlPC9BPjwvVT48L0ZPTlQ+PEZPTlQgQ09MT1I9QkxBQ0s+Jmd0OywgJnF1b3Q7 TG9hIEFuZGVyc3NvbiZxdW90OyAmbHQ7PC9GT05UPjxGT05UIENPTE9SPUJMVUU+PFU+PEEgSFJF Rj1tYWlsdG86bG9hQHBpLnNlPmxvYUBwaS5zZTwvQT48L1U+PC9GT05UPjxGT05UIENPTE9SPUJM QUNLPiZndDs8QlI+Y2M6PC9GT05UPjxGT05UIFNJWkU9ND4gPC9GT05UPiZsdDs8Rk9OVCBDT0xP Uj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOmNjYW1wQG9wcy5pZXRmLm9yZz5jY2FtcEBvcHMuaWV0 Zi5vcmc8L0E+PC9VPjwvRk9OVD48Rk9OVCBDT0xPUj1CTEFDSz4mZ3Q7PEJSPmJjYzo8L0ZPTlQ+ PEZPTlQgU0laRT00PiA8L0ZPTlQ+PEJSPlN1YmplY3Q6PEZPTlQgU0laRT00PiA8L0ZPTlQ+UmU6 IEZyYW1lZm9ybWF0IGluIGEgbDJjcyBnbXBscyBybnZpcm9ubWVudC48QlI+PEJSPjxCUj48Rk9O VCBTSVpFPTQ+SGksPC9GT05UPjxCUj48QlI+PEZPTlQgU0laRT00PkkgdGhpbmsgQmVuIGFuZCBK dWVyZ2VuIGhhdmUgYSBwb2ludCBoZXJlLiBUaGVyZSBpcyBub3RoaW5nIHRoYXQgY291bGQgYmUg ZHluYW1pY2FsbHkgcHJvdmlzaW9uZWQgdGhhdCBjb3VsZCBub3QgYmUgYWxzbyBwcm92aXNpb25l ZCBieSBtYW5hZ2VtZW50LiBJbiBvdGhlciB3b3JkcyBkYXRhIHBsYW5lIGlzIHRoZSBLaW5nIGhl cmUuIFdlIG5lZWQgZGF0YSBwbGFuZSBzdGFuZGFyZChzKSBvbiBob3cgd2UgZW5jb2RlIGEgbGFi ZWwsIHdoZXRoZXIgd2UgbmVlZCBhIGxhYmVsIHN0YWNrIG9yIG5vdCwgaG93IHRoZSBsYWJlbGVk IHRyYWZmaWMgaXMgc3VwcG9zZWQgdG8gYmUmbmJzcDt0cmVhdGVkLCBob3cgbGFiZWxlZCB0cmFm ZmljIGNvLWV4aXN0IHdpdGggbm9uLWxhYmVsIHRyYWZmaWMsIGV0Yy4gVGhpcyBpcyBzb21ldGhp bmcgdGhhdCBub3QgZm9yIENDQU1QIHRvIGRlZmluZS4gVGFrZSBmb3IgZXhhbXBsZSBURE0gbmV0 d29ya3MuIEdNUExTIG9ubHkgcHJvdmlkZXMgYSB3YXkgdG8gZHluYW1pY2FsbHkgcHJvdmlzaW9u IEcuNzA3IG5ldHdvcmtzLiBIZW5jZSB0aGVyZSBpcyBhIG5lZWQgaW4gcGFyYWxsZWwgc3RhbmRh cmRpemF0aW9uIGFjdGl2aXRpZXMgaW4gSVRVLiBPbmx5IGFmdGVyIHRoYXQgd2UgY2FuIGRpc2N1 c3MgaG93IGFsbCB0aGF0IGNvdWxkIGJlIGR5bmFtaWNhbGx5IHByb3Zpc2lvbmVkLCB0aGF0IGlz IHRoZSBhc3BlY3RzIG9mIGNvbnRyb2wgcGxhbmUuIDwvRk9OVD48QlI+PEJSPjxGT05UIFNJWkU9 ND5JZ29yPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+LS0tLS0gT3JpZ2luYWwgTWVzc2FnZSAtLS0t LSA8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9NT48Qj5Gcm9tOjwvQj4gPC9GT05UPjxGT05UIFNJWkU9 NSBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOkJlbi5NYWNrLUNyYW5lQHRlbGxhYnMuY29t Pk1hY2stQ3JhbmUsIFQuIEJlbmphbWluPC9BPjwvVT48L0ZPTlQ+PEZPTlQgU0laRT01IENPTE9S PUJMQUNLPiA8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9NT48Qj5Ubzo8L0I+IDwvRk9OVD48Rk9OVCBT SVpFPTUgQ09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0bzphZHJpYW5Ab2xkZG9nLmNvLnVrPkFk cmlhbiBGYXJyZWw8L0E+PC9VPjwvRk9OVD48Rk9OVCBTSVpFPTUgQ09MT1I9QkxBQ0s+IDsgPC9G T05UPjxGT05UIFNJWkU9NSBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOnBlckBkZWZlcm8u c2U+UCZhdW1sO3IgTWF0dHNzb248L0E+PC9VPjwvRk9OVD48Rk9OVCBTSVpFPTUgQ09MT1I9QkxB Q0s+IDsgPC9GT05UPjxGT05UIFNJWkU9NSBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOmxv YUBwaS5zZT5Mb2EgQW5kZXJzc29uPC9BPjwvVT48L0ZPTlQ+PEZPTlQgU0laRT01IENPTE9SPUJM QUNLPiA8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9NT48Qj5DYzo8L0I+IDwvRk9OVD48Rk9OVCBTSVpF PTUgQ09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0bzpjY2FtcEBvcHMuaWV0Zi5vcmc+Y2NhbXBA b3BzLmlldGYub3JnPC9BPjwvVT48L0ZPTlQ+PEZPTlQgU0laRT01IENPTE9SPUJMQUNLPiA8L0ZP TlQ+PEJSPjxGT05UIFNJWkU9NT48Qj5TZW50OjwvQj4gRnJpZGF5LCBKdWx5IDIyLCAyMDA1IDEw OjIzIEFNPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+PEI+U3ViamVjdDo8L0I+IFJFOiBGcmFtZWZv cm1hdCBpbiBhIGwyY3MgZ21wbHMgcm52aXJvbm1lbnQuPC9GT05UPjxCUj48QlI+PEZPTlQgU0la RT00PkhpLDwvRk9OVD48QlI+PEJSPjxGT05UIFNJWkU9ND5JIHRoaW5rIEp1ZXJnZW4gaGFzIHJh aXNlZCBhbiBpbXBvcnRhbnQgcXVlc3Rpb24uJm5ic3A7IEhvdyBmcmFtZXMgYXJlIGxhYmVsZWQg KGFuZCB0aGUgcmVsYXRlZCBkYXRhIHBsYW5lIGZvcndhcmRpbmcgYmVoYXZpb3IpIGlzIG5vdCBk ZWZpbmVkIGJ5IHRoZSBjb250cm9sIHBsYW5lLiZuYnNwOyBUaGUgY29udHJvbCBwbGFuZSBzZXJ2 ZXMgdG8gcHJvdmlzaW9uIHRoZSBkYXRhIHBsYW5lLCBub3QgZGVmaW5lIGl0LiZuYnNwOyBJbiB0 aGUgZnJhbWV3b3JrIGRyYWZ0IGl0IGlzIG5vdCBjbGVhciB3aGF0IGRhdGEgcGxhbmUgc3RhbmRh cmRzIGFyZSBjb3ZlcmVkIGJ5IHRoZSBzdGF0ZWQgY29udHJvbCBwbGFuZSByZXF1aXJlbWVudHMu Jm5ic3A7IFNvbWUgcmVmZXJlbmNlcyBzaG91bGQgYmUgc3VwcGxpZWQuJm5ic3A7IEluIGFueSBj YXNlLCB0aGUgbGFiZWxpbmcgYW5kIGZvcndhcmRpbmcgYmVoYXZpb3Igc2hvdWxkIGJlIGRlZmlu ZWQmbmJzcDtieSB0aGVzZSByZWZlcmVuY2VkIHN0YW5kYXJkcywgbm90Jm5ic3A7YnkgR01QTFMu PC9GT05UPjxCUj48QlI+PEZPTlQgU0laRT00PihJIGFtIGFzc3VtaW5nIGRlZmluaXRpb24gb2Yg bmV3IGRhdGEgcGxhbmUgc3RhbmRhcmRzIGlzIG91dCBvZiBzY29wZSBmb3IgQ0NBTVAuKTwvRk9O VD48QlI+PEJSPjxGT05UIFNJWkU9ND5SZWdhcmRzLDwvRk9OVD48QlI+PEZPTlQgU0laRT00PkJl bjwvRk9OVD48QlI+PEhSIFdJRFRIPSI5OSUiIFNJWkU9IjgiIEFMSUdOPSJsZWZ0Ij48Rk9OVCBD T0xPUj1CTEFDSz48QlI+PC9GT05UPjxGT05UIFNJWkU9ND48Qj5Gcm9tOjwvQj4gb3duZXItY2Nh bXBAb3BzLmlldGYub3JnIFttYWlsdG86b3duZXItY2NhbXBAb3BzLmlldGYub3JnXSA8Qj5PbiBC ZWhhbGYgT2YgPC9CPkFkcmlhbiBGYXJyZWw8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND48Qj5TZW50 OjwvQj4gRnJpZGF5LCBKdWx5IDIyLCAyMDA1IDg6MzcgQU08L0ZPTlQ+PEJSPjxGT05UIFNJWkU9 ND48Qj5Ubzo8L0I+IFAmYXVtbDtyIE1hdHRzc29uOyBMb2EgQW5kZXJzc29uPC9GT05UPjxCUj48 Rk9OVCBTSVpFPTQ+PEI+Q2M6PC9CPiBjY2FtcEBvcHMuaWV0Zi5vcmc8L0ZPTlQ+PEJSPjxGT05U IFNJWkU9ND48Qj5TdWJqZWN0OjwvQj4gUmU6IEZyYW1lZm9ybWF0IGluIGEgbDJjcyBnbXBscyBy bnZpcm9ubWVudC48L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTQ+SGksPC9GT05UPjxCUj48QlI+ PEZPTlQgU0laRT00Pkl0IGlzIGFsd2F5cyByZWZyZXNoaW5nIGhvdyBlbmdpbmVlcnMganVtcCBz dHJhaWdodCB0byBhIGRpc2N1c3Npb24gb2YgdGhlIHNvbHV0aW9uIDotKTwvRk9OVD48QlI+PEJS PjxGT05UIFNJWkU9ND5QZXJoYXBzIHdlIGNhbiBhc3N1bWUgdGhhdCB0aGUgZnJhbWV3b3JrIGRy YWZ0IGlzIGEgZ29vZCByZXByZXNlbnRhdGlvbiBvZiB0aGUgcHJvYmxlbSBzcGFjZSwgYW5kIHRo YXQgd2UgYXJlIHJlYWR5IHRvIHN0YXJ0IGRpc2N1c3NpbmcgdGhlIHNvbHV0aW9ucy48L0ZPTlQ+ PEJSPjxCUj48Rk9OVCBTSVpFPTQ+T25lIChwZXJoYXBzIHRoZSBvbmx5Pykgc2lnbmlmaWNhbnQg cXVlc3Rpb24gdG8gYW5zd2VyIGlzIGhvdyB0aGUgZnJhbWVzIHdpbGwgYmUgbGFiZWxsZWQuIFRo aXMgcXVlc3Rpb24gaXMgb25lIHdlIG11c3QgY29tZSB0byBhcyBzb29uIGFzIHdlIGFyZSBjb25m aWRlbnQgdGhhdCB0aGUgcmVxdWlyZW1lbnRzIG5lZWQgdG8gYmUgYWRkcmVzc2VkIGF0IGFsbC48 L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTQ+QXMgaGFzIGJlZW4gcG9pbnRlZCBvdXQsIHRoZXJl IGFyZSBzZXZlcmFsIHBvc3NpYmlsaXRpZXMgYW5kIHRvIHBpY2sgZnJvbSB0aGVzZSB3ZSBuZWVk IHRvIHVuZGVyc3RhbmQ6PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+MS4gRG8gd2UgbmVlZCB0byBz dXBwb3J0IGV4cGxpY2l0IGxhYmVsIHN0YWNraW5nPzwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZu YnNwOyZuYnNwOyBOb3RlIHRoYXQgdGhpcyBpcyBub3Qgc3VwcG9ydGVkIGluIFRETSwgTFNDIG9y IEZTQy48L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4yLiBEbyB3ZSBuZWVkIHRvIGJlIGFibGUgdG8g Y29udHJvbCAocGVyaGFwcyB0aHJvdWdoIGFuPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jm5ic3A7 Jm5ic3A7Jm5ic3A7ZXh0ZXJuYWwgc2lnbmFsaW5nIGNvbnRyb2xsZXIpIGV4aXN0aW5nIGhhcmR3 YXJlIGFuZCA8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mbmJzcDsmbmJzcDsgaW5zdGFsbCBMU1Bz IHRocm91Z2ggZXhpc3RpbmcgbmV0d29ya3M/PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+My4gRG8g d2UgbmVlZCB0byBzdXBwb3J0IGV4aXN0aW5nIGZ1bmN0aW9uIHNpbXVsdGFuZW91cyB0bzwvRk9O VD48QlI+PEZPTlQgU0laRT00PiZuYnNwOyZuYnNwOyB0aGUgc3VwcG9ydCBvZiBHTVBMUyBMMiBM U1BzPzwvRk9OVD48QlI+PEJSPjxGT05UIFNJWkU9ND5JIHRoaW5rIHRoYXQgZnJvbSBhIGNoYWly J3MgcGVyc3BlY3RpdmUgSSBjYW4gZ2l2ZSBzb21lIGxpbWl0ZWQgZ3VpZGFuY2UuPC9GT05UPjxC Uj48QlI+PEZPTlQgU0laRT00PkEuIFRoZXNlIHF1ZXN0aW9ucyBtdXN0IGJlIHJhaXNlZCBhbmQg YW5zd2VyZWQgaW4gdGhlIGZyYW1ld29yayBJLUQ8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND5CLiBU aGUgYW5zd2VyIHRvIHF1ZXN0aW9uIDMgaXMgJnF1b3Q7eWVzJnF1b3Q7LiBUaGlzIG1lYW5zIHRo YXQgdGhlIHVzZSBvZiBsYWJlbHM8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mbmJzcDsmbmJzcDsg bXVzdCBub3Qgc2lnbmlmaWNhbnRseSBkZXBsZXRlIGFueSBuYW1lc3BhY2UgdXNlZCB0byBzdXBw b3J0IG90aGVyPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jm5ic3A7Jm5ic3A7IGZ1bmN0aW9uLjwv Rk9OVD48QlI+PEZPTlQgU0laRT00PkMuIENDQU1QIGlzIGNoYXJ0ZXJlZCB0byBsb29rIGF0IHRo ZSBjb250cm9sIG9mIHRyYW5zcG9ydCBuZXR3b3Jrcy4gVGhpczwvRk9OVD48QlI+PEZPTlQgU0la RT00PiZuYnNwOyZuYnNwOyZuYnNwO2luY2x1ZGVzIE1ldHJvLCBidXQgSSBhbSB1bnN1cmUgYWJv dXQgQ2FtcHVzLiBJdCBkb2VzIG5vdCZuYnNwO2NvdmVyIDwvRk9OVD48QlI+PEZPTlQgU0laRT00 PiZuYnNwOyZuYnNwOyBzaWduYWxpbmcgdG8gdGhlIGRlc2t0b3AuPC9GT05UPjxCUj48QlI+PEZP TlQgU0laRT00PlRoYW5rcyw8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND5BZHJpYW48L0ZPTlQ+PEJS PjxCUj48Rk9OVCBTSVpFPTQ+LS0tLS0gT3JpZ2luYWwgTWVzc2FnZSAtLS0tLSA8L0ZPTlQ+PEJS PjxGT05UIFNJWkU9ND5Gcm9tOiAmcXVvdDtQJmF1bWw7ciBNYXR0c3NvbiZxdW90OyAmbHQ7PC9G T05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOnBlckBkZWZlcm8u c2U+cGVyQGRlZmVyby5zZTwvQT48L1U+PC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTEFDSz4m Z3Q7PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+VG86ICZxdW90O0xvYSBBbmRlcnNzb24mcXVvdDsg Jmx0OzwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0bzpsb2FA cGkuc2U+bG9hQHBpLnNlPC9BPjwvVT48L0ZPTlQ+PEZPTlQgU0laRT00IENPTE9SPUJMQUNLPiZn dDs8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND5DYzogJmx0OzwvRk9OVD48Rk9OVCBTSVpFPTQgQ09M T1I9QkxVRT48VT48QSBIUkVGPW1haWx0bzpjY2FtcEBvcHMuaWV0Zi5vcmc+Y2NhbXBAb3BzLmll dGYub3JnPC9BPjwvVT48L0ZPTlQ+PEZPTlQgU0laRT00IENPTE9SPUJMQUNLPiZndDs8L0ZPTlQ+ PEJSPjxGT05UIFNJWkU9ND5TZW50OiBGcmlkYXksIEp1bHkgMjIsIDIwMDUgMTE6NDEgQU08L0ZP TlQ+PEJSPjxGT05UIFNJWkU9ND5TdWJqZWN0OiBSZTogRnJhbWVmb3JtYXQgaW4gYSBsMmNzIGdt cGxzIHJudmlyb25tZW50LjwvRk9OVD48QlI+PEJSPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7 IFBlciBhbmQgRGltaXRyaSw8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDs8L0ZPTlQ+ PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsgSSB3b3VsZCBsaWtlIHRvIGNvbWUgZG93biBzdHJv bmdlciB0aGFuIHRoYXQsIGZvciBtZSBpdCBpczwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsg Jmd0OyBhIHZlcnkgc3Ryb25nIHJlcXVpcmVtZW50IHRoYXQgdGhlIHNhbWUgc3dpdGNoIGNhbiBo YW5kbGU8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsgYm90aCBWTEFOcyBhbmQgR01Q THMgdHJhZmljIGNvcnJlY3RseS4gSSBjYW4ndCBkc2VlIGhvdyB0aGF0PC9GT05UPjxCUj48Rk9O VCBTSVpFPTQ+Jmd0OyAmZ3Q7IGNvdWxkIGJlIGRvbmUgaWYgdXNpbmcgdGhlIFZMQU4gdHBpZCB0 byBpbmRpY2F0ZSBHTVBMUzwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyB0cmFmZmlj LjwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0 OyBJZiB5b3UgZXZlciB3YW50IHRoYXQgc2FtZSBzd2l0Y2ggdG8gaGFuZGxlIHRyYWZmaWMgZm9y IGEgZGlyZWN0bHk8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7IGNvbm5lY3RlZCZuYnNwOyBo b3N0IChub3QgdG8gdW5jb21tZW4pIHlvdSB3b3VsZCB3YW50IHRoYXQgdG8gdXNlIG5vcm1hbDwv Rk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgZXRoZXJuZXQgbWFjYWRkcmVzcyBzd2l0Y2hpbmcu IFNvIG9mIGNvdXJzZSB5b3UgZG8gbm90IHdhbnQgdG8gaGF2ZSB0bzwvRk9OVD48QlI+PEZPTlQg U0laRT00PiZndDsgY2hvb3NlIGJldHdlZW4gdmxhbiBhbmQgZ21wbHMsIHlvdSB3b3VsZCB3YW50 IGJvdGggYXQgdGhlIHNhbWUgdGltZS48L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7IDwvRk9O VD48QlI+PEZPTlQgU0laRT00PiZndDsgL3BlcjwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsg PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyA8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7 ICZndDs8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsgL0xvYTwvRk9OVD48QlI+PEZP TlQgU0laRT00PiZndDsgJmd0OzwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyBQJmF1 bWw7ciBNYXR0c3NvbiB3cm90ZTo8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsmZ3Q7 Jmd0O2hpIHBhciwgb25lIG9mIHRoZSBwb3NzaWJpbGl0aWVzIHRoYXQgaGFzIGJlZW4gY29uc2lk ZXJlZCB0byBjb3BlIHdpdGg8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsmZ3Q7Jmd0 O3RoaXMgcmVxdWlyZW1lbnQgaXMgdG8gdXNlIGEgZGVkaWNhdGVkIFRQSUQgZm9yIHRoZSBFdGhl cm5ldCBsYWJlbGVkPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7Jmd0OyZndDtmcmFt ZXM7IHRoaXMgd291bGQgYWxsb3cgZGlmZmVyZW50aWF0ZWQgcHJvY2Vzc2luZyB3aXRoIG5vbi1s YWJlbGVkPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7Jmd0OyZndDtmcmFtZXN0aGFu a3MuPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7Jmd0OzwvRk9OVD48QlI+PEZPTlQg U0laRT00PiZndDsgJmd0OyZndDs8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsmZ3Q7 IFRoYXQgc2VlbXMgdG8gbWFrZSBtb3JlIHNlbmNlLiBJZiB0aGF0IGZyYW1lIGlzIHRvIGJlIHNp emVkIGxpa2UgYTwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyZndDsgODAyLjFxPC9G T05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7Jmd0OyBmcmFtZS4gVGhlcmUgaXMgbm90IHRo YXQgbXVjaCBzcGFjZSBsZWZ0IHRvIGEgbGFiZWwuIE9yIGlzIHRoZSBkZW1hbmQgdG88L0ZPTlQ+ PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsmZ3Q7IHVzZSBqdW1ib2ZyYW1lcyA/PC9GT05UPjxC Uj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7Jmd0OyBIYXMgdGhlcmUgYmVlbiBhbnkgZGlzY3Vzc2lv biBvbiBsYWJlbHN0YWNraW5nLCBhbmQgbWFpbmx5IHdoZXJlIHRvPC9GT05UPjxCUj48Rk9OVCBT SVpFPTQ+Jmd0OyAmZ3Q7Jmd0OyBwbGFjZTwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0 OyZndDsgdGhlIGluZm9ybWF0aW9uPzwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyZn dDs8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsmZ3Q7IFJlZ2FyZHMuPC9GT05UPjxC Uj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7Jmd0OyBQZXI8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4m Z3Q7ICZndDsmZ3Q7PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7Jmd0OzwvRk9OVD48 QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyZndDs8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7 ICZndDs8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDs8L0ZPTlQ+PEJSPjxGT05UIFNJ WkU9ND4mZ3Q7ICZndDsgLS08L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsgTG9hIEFu ZGVyc3NvbjwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OzwvRk9OVD48QlI+PEZPTlQg U0laRT00PiZndDsgJmd0OyBQcmluY2lwYWwgTmV0d29ya2luZyBBcmNoaXRlY3Q8L0ZPTlQ+PEJS PjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsgQWNyZW8gQUImbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsm bmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJz cDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsm bmJzcDsmbmJzcDsmbmJzcDsgcGhvbmU6Jm5ic3A7ICs0NiA4IDYzMiA3NyAxNDwvRk9OVD48QlI+ PEZPTlQgU0laRT00PiZndDsgJmd0OyBJc2Fmam9yZHNnYXRhbiAyMiZuYnNwOyZuYnNwOyZuYnNw OyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZu YnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyBtb2JpbGU6ICs0NiA3MzkgODEgMjEgNjQ8L0ZP TlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsgS2lzdGEsIFN3ZWRlbiZuYnNwOyZuYnNwOyZu YnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNw OyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyBl bWFpbDombmJzcDsgPC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFp bHRvOmxvYS5hbmRlcnNzb25AYWNyZW8uc2U+bG9hLmFuZGVyc3NvbkBhY3Jlby5zZTwvQT48L1U+ PC9GT05UPjxGT05UIENPTE9SPUJMQUNLPjxCUj48L0ZPTlQ+PEZPTlQgU0laRT00PiZndDsgJmd0 OyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZu YnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNw OyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZu YnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNw OyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyZuYnNwOyA8L0ZPTlQ+PEZPTlQgU0laRT00 IENPTE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86bG9hQHBpLnNlPmxvYUBwaS5zZTwvQT48L1U+ PC9GT05UPjxGT05UIENPTE9SPUJMQUNLPjxCUj48L0ZPTlQ+PEZPTlQgU0laRT00PiZndDsgJmd0 OzwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0 OyA8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7IDwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZn dDsgPC9GT05UPjxCUj48QlI+PEZPTlQgU0laRT01Pj09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PTwvRk9OVD48QlI+PEZPTlQgU0laRT01 PlRoZSBpbmZvcm1hdGlvbiBjb250YWluZWQgaW4gdGhpcyBtZXNzYWdlIG1heSBiZSBwcml2aWxl Z2VkPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+YW5kIGNvbmZpZGVudGlhbCBhbmQgcHJvdGVjdGVk IGZyb20gZGlzY2xvc3VyZS4gSWYgdGhlIHJlYWRlcjwvRk9OVD48QlI+PEZPTlQgU0laRT01Pm9m IHRoaXMgbWVzc2FnZSBpcyBub3QgdGhlIGludGVuZGVkIHJlY2lwaWVudCwgb3IgYW4gZW1wbG95 ZWU8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9NT5vciBhZ2VudCByZXNwb25zaWJsZSBmb3IgZGVsaXZl cmluZyB0aGlzIG1lc3NhZ2UgdG8gdGhlPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+aW50ZW5kZWQg cmVjaXBpZW50LCB5b3UgYXJlIGhlcmVieSBub3RpZmllZCB0aGF0IGFueSByZXByb2R1Y3Rpb24s PC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+ZGlzc2VtaW5hdGlvbiBvciBkaXN0cmlidXRpb24gb2Yg dGhpcyBjb21tdW5pY2F0aW9uIGlzIHN0cmljdGx5PC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+cHJv aGliaXRlZC4gSWYgeW91IGhhdmUgcmVjZWl2ZWQgdGhpcyBjb21tdW5pY2F0aW9uIGluIGVycm9y LDwvRk9OVD48QlI+PEZPTlQgU0laRT01PnBsZWFzZSBub3RpZnkgdXMgaW1tZWRpYXRlbHkgYnkg cmVwbHlpbmcgdG8gdGhlIG1lc3NhZ2UgYW5kPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+ZGVsZXRp bmcgaXQgZnJvbSB5b3VyIGNvbXB1dGVyLiBUaGFuayB5b3UuIFRlbGxhYnM8L0ZPTlQ+PEJSPjxG T05UIFNJWkU9NT49PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT08L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTQgRkFDRT0iTW9ub3NwYWNl LENvdXJpZXIiPj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PTwvRk9OVD48QlI+PEZPTlQgU0laRT00IEZBQ0U9Ik1vbm9zcGFjZSxDb3Vy aWVyIj5UaGUgaW5mb3JtYXRpb24gY29udGFpbmVkIGluIHRoaXMgbWVzc2FnZSBtYXkgYmUgcHJp dmlsZWdlZDwvRk9OVD48QlI+PEZPTlQgU0laRT00IEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5h bmQgY29uZmlkZW50aWFsIGFuZCBwcm90ZWN0ZWQgZnJvbSBkaXNjbG9zdXJlLiBJZiB0aGUgcmVh ZGVyPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPm9mIHRo aXMgbWVzc2FnZSBpcyBub3QgdGhlIGludGVuZGVkIHJlY2lwaWVudCwgb3IgYW4gZW1wbG95ZWU8 L0ZPTlQ+PEJSPjxGT05UIFNJWkU9NCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+b3IgYWdlbnQg cmVzcG9uc2libGUgZm9yIGRlbGl2ZXJpbmcgdGhpcyBtZXNzYWdlIHRvIHRoZTwvRk9OVD48QlI+ PEZPTlQgU0laRT00IEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5pbnRlbmRlZCByZWNpcGllbnQs IHlvdSBhcmUgaGVyZWJ5IG5vdGlmaWVkIHRoYXQgYW55IHJlcHJvZHVjdGlvbiw8L0ZPTlQ+PEJS PjxGT05UIFNJWkU9NCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+ZGlzc2VtaW5hdGlvbiBvciBk aXN0cmlidXRpb24gb2YgdGhpcyBjb21tdW5pY2F0aW9uIGlzIHN0cmljdGx5PC9GT05UPjxCUj48 Rk9OVCBTSVpFPTQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPnByb2hpYml0ZWQuIElmIHlvdSBo YXZlIHJlY2VpdmVkIHRoaXMgY29tbXVuaWNhdGlvbiBpbiBlcnJvciw8L0ZPTlQ+PEJSPjxGT05U IFNJWkU9NCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+cGxlYXNlIG5vdGlmeSB1cyBpbW1lZGlh dGVseSBieSByZXBseWluZyB0byB0aGUgbWVzc2FnZSBhbmQ8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9 NCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+ZGVsZXRpbmcgaXQgZnJvbSB5b3VyIGNvbXB1dGVy LiBUaGFuayB5b3UuIFRlbGxhYnM8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9NCBGQUNFPSJNb25vc3Bh Y2UsQ291cmllciI+PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PC9GT05UPjxCUj48L1A+ Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 18:28:24 +0000 To: "Igor Bryskin" <ibryskin@movaz.com> Cc: <Dimitri.Papadimitriou@alcatel.be>, "Mack-Crane, T. Benjamin" <Ben.Mack-Crane@tellabs.com>, "Adrian Farrel" <adrian@olddog.co.uk>, =?iso-8859-1?Q?P=E4r_Mattsson?= <per@defero.se>, "Loa Andersson" <loa@pi.se>, <ccamp@ops.ietf.org> MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: Re: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 20:27:34 +0200 Message-ID: <OF50B02F59.9B130229-ONC1257046.00656637-C1257046.006566B4@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmlnb3IgPEJSPjxCUj4mZ3Q7IFRoZSBx dWVzdGlvbiBpcyBzaW1wbGU6IGlzIGl0IHBvc3NpYmxlIHRvZGF5IHRvIHN0YXRpY2FsbHkgcHJv dmlzaW9uIEwyTFNQcyB0aGF0IGNvdWxkLCBzYXksJm5ic3A7c3VwcG9ydCBlMmUgUW9TPzwvRk9O VD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPltkcF0gaSBzdWdnZXN0IHlv dSB0YWtlIGEgbG9vayBhdCB0aGUgcHJvcG9zZWQgc2NlbmFyaW9zIG9mIHRoaXMgZG9jdW1lbnQ8 L0ZPTlQ+PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj4mZ3Q7IElmIG5vdCwg bWF5IGJlIGl0IGlzIHRvIGVhcmx5IHRvIGRpc2N1c3MgYXNwZWN0cyBvZiBkeW5hbWljIHByb3Zp c2lvbmluZyBvZiBzdWNoIExTUHM/IDwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNl LENvdXJpZXIiPiZndDsgSXMgaXQgcG9zc2libGUvcmVhc29uYWJsZSwgaW4geW91ciBvcGluaW9u LCB0byAmcXVvdDtkZXRhaWwgaG93IGZvcndhcmRpbmcgaW5mb3JtYXRpb24gY2FuIGJlIGV4Y2hh bmdlZCB2aWEgdGhlIDwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIi PiZndDsgY29udHJvbCBwbGFuZSAoYW5kIHRoZW4gaW5zdGFsbGVkKSZxdW90OyB3aXRob3V0Jm5i c3A7aGF2aW5nIGluIG1pbmQgJnF1b3Q7YSBzcGVjaWZpYyBmb3J3YXJkaW5nIGJlaGF2aW9yJnF1 b3Q7PzwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPltkcF0gaSBk b24ndCB0aGluayBhbnlvbmUgaGFzIHNhaWQgdGhlIGNvbnRyYXJ5IGJ1dCBpIGRvbid0IHRoaXMg bWVhbnMgZGVmaW5lIGEgbmV3IGZvcndhcmRpbmcgYmVoYXZpb3VyPC9GT05UPjwvUD48UD48Rk9O VCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+Jmd0OyBIb3cgZG8geW91IGtub3cgd2hpY2ggZm9y d2FyZGluZyBpbmZvcm1hdGlvbiBpcyBuZWVkZWQgZm9yIHRoZSBmb3J3YXJkaW5nIG5vYm9keSBo YXMgZGVmaW5lZCB5ZXQ/PC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmll ciI+Jmd0OyBUaGUgb3RoZXIgcXVlc3Rpb24gaXMmbmJzcDtkbyB3ZSBuZWVkIGF0IGFsbCBlMmUg UW9TLCByb3V0ZSBjb250cm9sLCBmYXN0IHJlY292ZXJ5PyA8L0ZPTlQ+PC9QPjxQPjxGT05UIEZB Q0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5bZHBdIHdoZXJlIGluIHRoZSBkb2N1bWVudCBpdCBoYXMg YmVlbiBtZW50aW9uZWQgZTJlIFFvUyBvciBzb21ldGhpbmcgYXJvdW5kIGVuZC10by1lbmQgYmVo YXZpb3VyIGJ1dCBwcm9iYWJseSBpIHNob3VsZCBpIGFzayB3aGF0IHlvdSBtZWFuIHdpdGggJnF1 b3Q7ZW5kLXRvLWVuZCZxdW90OyBoZXJlIDwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3Nw YWNlLENvdXJpZXIiPiZndDsgU291bmRzIGxpa2UgZXhjaXRpbmcgaWRlYSwgYnV0IGRvZXMgYWxs IHRoYXQgY29tbWVyY2lhbGx5IG1ha2Ugc2Vuc2U/IDwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0i TW9ub3NwYWNlLENvdXJpZXIiPiZndDsgU28sIEkgZ3Vlc3MsIHdlIG5lZWQgdHdvIHRoaW5ncyZu YnNwO2JlZm9yZSB3ZSBjYW4mbmJzcDttb3ZlIGZvcndhcmQ6PEJSPiZndDsgYSkgdmFsaWRhdGlv biBvZiB0aGUgaWRlYSBieSB0aGUgcHJvdmlkZXJzOzwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0i TW9ub3NwYWNlLENvdXJpZXIiPltkcF0gaW5kZWVkLCB3ZSBkZWZpbml0ZWx5IG5lZWQgZmVlZGJh Y2sgZnJvbSB0aGUgdXNlciBjb21tdW5pdHkgYnV0IGFsc28ga2VlcCB0aGlzIGRpc2N1c3Npb24g aW4gdGhlIGlldGYgY29udGV4dCAoZnlpOiAxLzMgb2YgdGhlIERUIGlzIGNvbnN0aXR1ZWQgYnkg aW5kaXZpZHVhbHMgcGFydCBvZiB0aGUgdXNlciBjb21tdW5pdHkpPC9GT05UPjwvUD48UD48Rk9O VCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+Jmd0OyBiKSBkZWZpbml0aW9uIG9mIHRoZSBkYXRh IHBsYW5lIGJlaGF2aW9yOzwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJp ZXIiPltkcF0gd2UgZG8gbm90IG5lZWQgYSBkZWZpbml0aW9uIG9mIHRoZSBkYXRhIHBsYW5lIGJl aGF2aW91ciB3ZSBkbyBuZWVkIHRvIGlkZW50aWZ5IHRoZSByZWxldmFudCBiZWhhdmlvdXIgZm9y IHdoaWNoIEdNUExTIGNvbnRyb2wgcGxhbmUgd2lsbCBiZSB1c2VkIGZvciBleGNoYW5naW5nIHRo ZSBmb3J3YXJkaW5nIGluZm9ybWF0aW9uIDxCUj48L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4tLS0t LSBPcmlnaW5hbCBNZXNzYWdlIC0tLS0tIDwvRk9OVD48QlI+PEZPTlQgU0laRT00PjxCPkZyb206 PC9CPiA8L0ZPTlQ+PEZPTlQgU0laRT00IENPTE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86RGlt aXRyaS5QYXBhZGltaXRyaW91QGFsY2F0ZWwuYmU+RGltaXRyaS5QYXBhZGltaXRyaW91QGFsY2F0 ZWwuYmU8L0E+PC9VPjwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxBQ0s+IDwvRk9OVD48QlI+ PEZPTlQgU0laRT00PjxCPlRvOjwvQj4gPC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTFVFPjxV PjxBIEhSRUY9bWFpbHRvOmlicnlza2luQG1vdmF6LmNvbT5JZ29yIEJyeXNraW48L0E+PC9VPjwv Rk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxBQ0s+IDwvRk9OVD48QlI+PEZPTlQgU0laRT00PjxC PkNjOjwvQj4gPC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRv OkJlbi5NYWNrLUNyYW5lQHRlbGxhYnMuY29tPk1hY2stQ3JhbmUsIFQuIEJlbmphbWluPC9BPjwv VT48L0ZPTlQ+PEZPTlQgU0laRT00IENPTE9SPUJMQUNLPiA7IDwvRk9OVD48Rk9OVCBTSVpFPTQg Q09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0bzphZHJpYW5Ab2xkZG9nLmNvLnVrPkFkcmlhbiBG YXJyZWw8L0E+PC9VPjwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxBQ0s+IDsgPC9GT05UPjxG T05UIFNJWkU9NCBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOnBlckBkZWZlcm8uc2U+UCZh dW1sO3IgTWF0dHNzb248L0E+PC9VPjwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxBQ0s+IDsg PC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOmxvYUBwaS5z ZT5Mb2EgQW5kZXJzc29uPC9BPjwvVT48L0ZPTlQ+PEZPTlQgU0laRT00IENPTE9SPUJMQUNLPiA7 IDwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0bzpjY2FtcEBv cHMuaWV0Zi5vcmc+Y2NhbXBAb3BzLmlldGYub3JnPC9BPjwvVT48L0ZPTlQ+PEZPTlQgU0laRT00 IENPTE9SPUJMQUNLPiA8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND48Qj5TZW50OjwvQj4gRnJpZGF5 LCBKdWx5IDIyLCAyMDA1IDEyOjIxIFBNPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+PEI+U3ViamVj dDo8L0I+IFJlOiBGcmFtZWZvcm1hdCBpbiBhIGwyY3MgZ21wbHMgcm52aXJvbm1lbnQuPC9GT05U PjxCUj48QlI+PEJSPjxGT05UIFNJWkU9ND5pZ29yIC0gaSBhbSBub3Qgc3VyZSB3aGljaCBwb2lu dCB0aGV5IGV4YWN0bHkgaGF2ZSAtIHRoZSBwdXJwb3NlIG9mIHRoaXMgZG9jdW1lbnQgaXMgdG8g ZGV0YWlsIGhvdyBmb3J3YXJkaW5nIGluZm9ybWF0aW9uIGNhbiBiZSBleGNoYW5nZWQgdmlhIHRo ZSBjb250cm9sIHBsYW5lIChhbmQgdGhlbiBpbnN0YWxsZWQpIG5vdCB0byBkZWZpbmUgYSBzcGVj aWZpYyBmb3J3YXJkaW5nIGJlaGF2aW91cjwvRk9OVD48QlI+PEJSPjxGT05UIFNJWkU9ND50aGFu a3MsPC9GT05UPjxCUj48QlI+PEZPTlQgU0laRT00Pi0gZGltaXRyaS48L0ZPTlQ+PEJSPjxCUj48 Qj4mcXVvdDtJZ29yIEJyeXNraW4mcXVvdDsgJmx0OzwvQj48Rk9OVCBDT0xPUj1CTFVFPjxCPjxV PjxBIEhSRUY9bWFpbHRvOmlicnlza2luQG1vdmF6LmNvbT5pYnJ5c2tpbkBtb3Zhei5jb208L0E+ PC9VPjwvQj48L0ZPTlQ+PEZPTlQgQ09MT1I9QkxBQ0s+PEI+Jmd0OzwvQj48QlI+U2VudCBieTog PC9GT05UPjxGT05UIENPTE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86b3duZXItY2NhbXBAb3Bz LmlldGYub3JnPm93bmVyLWNjYW1wQG9wcy5pZXRmLm9yZzwvQT48L1U+PC9GT05UPjxGT05UIENP TE9SPUJMQUNLPjxCUj4wNy8yMi8yMDA1IDExOjEwIEFTVDxCUj48QlI+VG86PC9GT05UPjxGT05U IFNJWkU9ND4gPC9GT05UPiZxdW90O01hY2stQ3JhbmUsIFQuIEJlbmphbWluJnF1b3Q7ICZsdDs8 Rk9OVCBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOkJlbi5NYWNrLUNyYW5lQHRlbGxhYnMu Y29tPkJlbi5NYWNrLUNyYW5lQHRlbGxhYnMuY29tPC9BPjwvVT48L0ZPTlQ+PEZPTlQgQ09MT1I9 QkxBQ0s+Jmd0OywgJnF1b3Q7QWRyaWFuIEZhcnJlbCZxdW90OyAmbHQ7PC9GT05UPjxGT05UIENP TE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86YWRyaWFuQG9sZGRvZy5jby51az5hZHJpYW5Ab2xk ZG9nLmNvLnVrPC9BPjwvVT48L0ZPTlQ+PEZPTlQgQ09MT1I9QkxBQ0s+Jmd0OywgUCZhdW1sO3Ig TWF0dHNzb24gJmx0OzwvRk9OVD48Rk9OVCBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOnBl ckBkZWZlcm8uc2U+cGVyQGRlZmVyby5zZTwvQT48L1U+PC9GT05UPjxGT05UIENPTE9SPUJMQUNL PiZndDssICZxdW90O0xvYSBBbmRlcnNzb24mcXVvdDsgJmx0OzwvRk9OVD48Rk9OVCBDT0xPUj1C TFVFPjxVPjxBIEhSRUY9bWFpbHRvOmxvYUBwaS5zZT5sb2FAcGkuc2U8L0E+PC9VPjwvRk9OVD48 Rk9OVCBDT0xPUj1CTEFDSz4mZ3Q7PEJSPmNjOjwvRk9OVD48Rk9OVCBTSVpFPTQ+IDwvRk9OVD4m bHQ7PEZPTlQgQ09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0bzpjY2FtcEBvcHMuaWV0Zi5vcmc+ Y2NhbXBAb3BzLmlldGYub3JnPC9BPjwvVT48L0ZPTlQ+PEZPTlQgQ09MT1I9QkxBQ0s+Jmd0OzxC Uj5iY2M6PC9GT05UPjxGT05UIFNJWkU9ND4gPC9GT05UPjxCUj5TdWJqZWN0OjxGT05UIFNJWkU9 ND4gPC9GT05UPlJlOiBGcmFtZWZvcm1hdCBpbiBhIGwyY3MgZ21wbHMgcm52aXJvbm1lbnQuPEJS PjxCUj48QlI+PEZPTlQgU0laRT00PkhpLDwvRk9OVD48QlI+PEJSPjxGT05UIFNJWkU9ND5JIHRo aW5rIEJlbiBhbmQgSnVlcmdlbiBoYXZlIGEgcG9pbnQgaGVyZS4gVGhlcmUgaXMgbm90aGluZyB0 aGF0IGNvdWxkIGJlIGR5bmFtaWNhbGx5IHByb3Zpc2lvbmVkIHRoYXQgY291bGQgbm90IGJlIGFs c28gcHJvdmlzaW9uZWQgYnkgbWFuYWdlbWVudC4gSW4gb3RoZXIgd29yZHMgZGF0YSBwbGFuZSBp cyB0aGUgS2luZyBoZXJlLiBXZSBuZWVkIGRhdGEgcGxhbmUgc3RhbmRhcmQocykgb24gaG93IHdl IGVuY29kZSBhIGxhYmVsLCB3aGV0aGVyIHdlIG5lZWQgYSBsYWJlbCBzdGFjayBvciBub3QsIGhv dyB0aGUgbGFiZWxlZCB0cmFmZmljIGlzIHN1cHBvc2VkIHRvIGJlJm5ic3A7dHJlYXRlZCwgaG93 IGxhYmVsZWQgdHJhZmZpYyBjby1leGlzdCB3aXRoIG5vbi1sYWJlbCB0cmFmZmljLCBldGMuIFRo aXMgaXMgc29tZXRoaW5nIHRoYXQgbm90IGZvciBDQ0FNUCB0byBkZWZpbmUuIFRha2UgZm9yIGV4 YW1wbGUgVERNIG5ldHdvcmtzLiBHTVBMUyBvbmx5IHByb3ZpZGVzIGEgd2F5IHRvIGR5bmFtaWNh bGx5IHByb3Zpc2lvbiBHLjcwNyBuZXR3b3Jrcy4gSGVuY2UgdGhlcmUgaXMgYSBuZWVkIGluIHBh cmFsbGVsIHN0YW5kYXJkaXphdGlvbiBhY3Rpdml0aWVzIGluIElUVS4gT25seSBhZnRlciB0aGF0 IHdlIGNhbiBkaXNjdXNzIGhvdyBhbGwgdGhhdCBjb3VsZCBiZSBkeW5hbWljYWxseSBwcm92aXNp b25lZCwgdGhhdCBpcyB0aGUgYXNwZWN0cyBvZiBjb250cm9sIHBsYW5lLiA8L0ZPTlQ+PEJSPjxC Uj48Rk9OVCBTSVpFPTQ+SWdvcjwvRk9OVD48QlI+PEZPTlQgU0laRT01Pi0tLS0tIE9yaWdpbmFs IE1lc3NhZ2UgLS0tLS0gPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+PEI+RnJvbTo8L0I+IDwvRk9O VD48Rk9OVCBTSVpFPTUgQ09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0bzpCZW4uTWFjay1DcmFu ZUB0ZWxsYWJzLmNvbT5NYWNrLUNyYW5lLCBULiBCZW5qYW1pbjwvQT48L1U+PC9GT05UPjxGT05U IFNJWkU9NSBDT0xPUj1CTEFDSz4gPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+PEI+VG86PC9CPiA8 L0ZPTlQ+PEZPTlQgU0laRT01IENPTE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86YWRyaWFuQG9s ZGRvZy5jby51az5BZHJpYW4gRmFycmVsPC9BPjwvVT48L0ZPTlQ+PEZPTlQgU0laRT01IENPTE9S PUJMQUNLPiA7IDwvRk9OVD48Rk9OVCBTSVpFPTUgQ09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0 bzpwZXJAZGVmZXJvLnNlPlAmYXVtbDtyIE1hdHRzc29uPC9BPjwvVT48L0ZPTlQ+PEZPTlQgU0la RT01IENPTE9SPUJMQUNLPiA7IDwvRk9OVD48Rk9OVCBTSVpFPTUgQ09MT1I9QkxVRT48VT48QSBI UkVGPW1haWx0bzpsb2FAcGkuc2U+TG9hIEFuZGVyc3NvbjwvQT48L1U+PC9GT05UPjxGT05UIFNJ WkU9NSBDT0xPUj1CTEFDSz4gPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+PEI+Q2M6PC9CPiA8L0ZP TlQ+PEZPTlQgU0laRT01IENPTE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86Y2NhbXBAb3BzLmll dGYub3JnPmNjYW1wQG9wcy5pZXRmLm9yZzwvQT48L1U+PC9GT05UPjxGT05UIFNJWkU9NSBDT0xP Uj1CTEFDSz4gPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+PEI+U2VudDo8L0I+IEZyaWRheSwgSnVs eSAyMiwgMjAwNSAxMDoyMyBBTTwvRk9OVD48QlI+PEZPTlQgU0laRT01PjxCPlN1YmplY3Q6PC9C PiBSRTogRnJhbWVmb3JtYXQgaW4gYSBsMmNzIGdtcGxzIHJudmlyb25tZW50LjwvRk9OVD48QlI+ PEJSPjxGT05UIFNJWkU9ND5IaSw8L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTQ+SSB0aGluayBK dWVyZ2VuIGhhcyByYWlzZWQgYW4gaW1wb3J0YW50IHF1ZXN0aW9uLiZuYnNwOyBIb3cgZnJhbWVz IGFyZSBsYWJlbGVkIChhbmQgdGhlIHJlbGF0ZWQgZGF0YSBwbGFuZSBmb3J3YXJkaW5nIGJlaGF2 aW9yKSBpcyBub3QgZGVmaW5lZCBieSB0aGUgY29udHJvbCBwbGFuZS4mbmJzcDsgVGhlIGNvbnRy b2wgcGxhbmUgc2VydmVzIHRvIHByb3Zpc2lvbiB0aGUgZGF0YSBwbGFuZSwgbm90IGRlZmluZSBp dC4mbmJzcDsgSW4gdGhlIGZyYW1ld29yayBkcmFmdCBpdCBpcyBub3QgY2xlYXIgd2hhdCBkYXRh IHBsYW5lIHN0YW5kYXJkcyBhcmUgY292ZXJlZCBieSB0aGUgc3RhdGVkIGNvbnRyb2wgcGxhbmUg cmVxdWlyZW1lbnRzLiZuYnNwOyBTb21lIHJlZmVyZW5jZXMgc2hvdWxkIGJlIHN1cHBsaWVkLiZu YnNwOyBJbiBhbnkgY2FzZSwgdGhlIGxhYmVsaW5nIGFuZCBmb3J3YXJkaW5nIGJlaGF2aW9yIHNo b3VsZCBiZSBkZWZpbmVkJm5ic3A7YnkgdGhlc2UgcmVmZXJlbmNlZCBzdGFuZGFyZHMsIG5vdCZu YnNwO2J5IEdNUExTLjwvRk9OVD48QlI+PEJSPjxGT05UIFNJWkU9ND4oSSBhbSBhc3N1bWluZyBk ZWZpbml0aW9uIG9mIG5ldyBkYXRhIHBsYW5lIHN0YW5kYXJkcyBpcyBvdXQgb2Ygc2NvcGUgZm9y IENDQU1QLik8L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTQ+UmVnYXJkcyw8L0ZPTlQ+PEJSPjxG T05UIFNJWkU9ND5CZW48L0ZPTlQ+PEJSPjxIUiBXSURUSD0iOTklIiBTSVpFPSI4IiBBTElHTj0i bGVmdCI+PEZPTlQgQ09MT1I9QkxBQ0s+PEJSPjwvRk9OVD48Rk9OVCBTSVpFPTQ+PEI+RnJvbTo8 L0I+IG93bmVyLWNjYW1wQG9wcy5pZXRmLm9yZyBbbWFpbHRvOm93bmVyLWNjYW1wQG9wcy5pZXRm Lm9yZ10gPEI+T24gQmVoYWxmIE9mIDwvQj5BZHJpYW4gRmFycmVsPC9GT05UPjxCUj48Rk9OVCBT SVpFPTQ+PEI+U2VudDo8L0I+IEZyaWRheSwgSnVseSAyMiwgMjAwNSA4OjM3IEFNPC9GT05UPjxC Uj48Rk9OVCBTSVpFPTQ+PEI+VG86PC9CPiBQJmF1bWw7ciBNYXR0c3NvbjsgTG9hIEFuZGVyc3Nv bjwvRk9OVD48QlI+PEZPTlQgU0laRT00PjxCPkNjOjwvQj4gY2NhbXBAb3BzLmlldGYub3JnPC9G T05UPjxCUj48Rk9OVCBTSVpFPTQ+PEI+U3ViamVjdDo8L0I+IFJlOiBGcmFtZWZvcm1hdCBpbiBh IGwyY3MgZ21wbHMgcm52aXJvbm1lbnQuPC9GT05UPjxCUj48QlI+PEZPTlQgU0laRT00PkhpLDwv Rk9OVD48QlI+PEJSPjxGT05UIFNJWkU9ND5JdCBpcyBhbHdheXMgcmVmcmVzaGluZyBob3cgZW5n aW5lZXJzIGp1bXAgc3RyYWlnaHQgdG8gYSBkaXNjdXNzaW9uIG9mIHRoZSBzb2x1dGlvbiA6LSk8 L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTQ+UGVyaGFwcyB3ZSBjYW4gYXNzdW1lIHRoYXQgdGhl IGZyYW1ld29yayBkcmFmdCBpcyBhIGdvb2QgcmVwcmVzZW50YXRpb24gb2YgdGhlIHByb2JsZW0g c3BhY2UsIGFuZCB0aGF0IHdlIGFyZSByZWFkeSB0byBzdGFydCBkaXNjdXNzaW5nIHRoZSBzb2x1 dGlvbnMuPC9GT05UPjxCUj48QlI+PEZPTlQgU0laRT00Pk9uZSAocGVyaGFwcyB0aGUgb25seT8p IHNpZ25pZmljYW50IHF1ZXN0aW9uIHRvIGFuc3dlciBpcyBob3cgdGhlIGZyYW1lcyB3aWxsIGJl IGxhYmVsbGVkLiBUaGlzIHF1ZXN0aW9uIGlzIG9uZSB3ZSBtdXN0IGNvbWUgdG8gYXMgc29vbiBh cyB3ZSBhcmUgY29uZmlkZW50IHRoYXQgdGhlIHJlcXVpcmVtZW50cyBuZWVkIHRvIGJlIGFkZHJl c3NlZCBhdCBhbGwuPC9GT05UPjxCUj48QlI+PEZPTlQgU0laRT00PkFzIGhhcyBiZWVuIHBvaW50 ZWQgb3V0LCB0aGVyZSBhcmUgc2V2ZXJhbCBwb3NzaWJpbGl0aWVzIGFuZCB0byBwaWNrIGZyb20g dGhlc2Ugd2UgbmVlZCB0byB1bmRlcnN0YW5kOjwvRk9OVD48QlI+PEZPTlQgU0laRT00PjEuIERv IHdlIG5lZWQgdG8gc3VwcG9ydCBleHBsaWNpdCBsYWJlbCBzdGFja2luZz88L0ZPTlQ+PEJSPjxG T05UIFNJWkU9ND4mbmJzcDsmbmJzcDsgTm90ZSB0aGF0IHRoaXMgaXMgbm90IHN1cHBvcnRlZCBp biBURE0sIExTQyBvciBGU0MuPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Mi4gRG8gd2UgbmVlZCB0 byBiZSBhYmxlIHRvIGNvbnRyb2wgKHBlcmhhcHMgdGhyb3VnaCBhbjwvRk9OVD48QlI+PEZPTlQg U0laRT00PiZuYnNwOyZuYnNwOyZuYnNwO2V4dGVybmFsIHNpZ25hbGluZyBjb250cm9sbGVyKSBl eGlzdGluZyBoYXJkd2FyZSBhbmQgPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jm5ic3A7Jm5ic3A7 IGluc3RhbGwgTFNQcyB0aHJvdWdoIGV4aXN0aW5nIG5ldHdvcmtzPzwvRk9OVD48QlI+PEZPTlQg U0laRT00PjMuIERvIHdlIG5lZWQgdG8gc3VwcG9ydCBleGlzdGluZyBmdW5jdGlvbiBzaW11bHRh bmVvdXMgdG88L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mbmJzcDsmbmJzcDsgdGhlIHN1cHBvcnQg b2YgR01QTFMgTDIgTFNQcz88L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTQ+SSB0aGluayB0aGF0 IGZyb20gYSBjaGFpcidzIHBlcnNwZWN0aXZlIEkgY2FuIGdpdmUgc29tZSBsaW1pdGVkIGd1aWRh bmNlLjwvRk9OVD48QlI+PEJSPjxGT05UIFNJWkU9ND5BLiBUaGVzZSBxdWVzdGlvbnMgbXVzdCBi ZSByYWlzZWQgYW5kIGFuc3dlcmVkIGluIHRoZSBmcmFtZXdvcmsgSS1EPC9GT05UPjxCUj48Rk9O VCBTSVpFPTQ+Qi4gVGhlIGFuc3dlciB0byBxdWVzdGlvbiAzIGlzICZxdW90O3llcyZxdW90Oy4g VGhpcyBtZWFucyB0aGF0IHRoZSB1c2Ugb2YgbGFiZWxzPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+ Jm5ic3A7Jm5ic3A7IG11c3Qgbm90IHNpZ25pZmljYW50bHkgZGVwbGV0ZSBhbnkgbmFtZXNwYWNl IHVzZWQgdG8gc3VwcG9ydCBvdGhlcjwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZuYnNwOyZuYnNw OyBmdW5jdGlvbi48L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND5DLiBDQ0FNUCBpcyBjaGFydGVyZWQg dG8gbG9vayBhdCB0aGUgY29udHJvbCBvZiB0cmFuc3BvcnQgbmV0d29ya3MuIFRoaXM8L0ZPTlQ+ PEJSPjxGT05UIFNJWkU9ND4mbmJzcDsmbmJzcDsmbmJzcDtpbmNsdWRlcyBNZXRybywgYnV0IEkg YW0gdW5zdXJlIGFib3V0IENhbXB1cy4gSXQgZG9lcyBub3QmbmJzcDtjb3ZlciA8L0ZPTlQ+PEJS PjxGT05UIFNJWkU9ND4mbmJzcDsmbmJzcDsgc2lnbmFsaW5nIHRvIHRoZSBkZXNrdG9wLjwvRk9O VD48QlI+PEJSPjxGT05UIFNJWkU9ND5UaGFua3MsPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+QWRy aWFuPC9GT05UPjxCUj48QlI+PEZPTlQgU0laRT00Pi0tLS0tIE9yaWdpbmFsIE1lc3NhZ2UgLS0t LS0gPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+RnJvbTogJnF1b3Q7UCZhdW1sO3IgTWF0dHNzb24m cXVvdDsgJmx0OzwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0 bzpwZXJAZGVmZXJvLnNlPnBlckBkZWZlcm8uc2U8L0E+PC9VPjwvRk9OVD48Rk9OVCBTSVpFPTQg Q09MT1I9QkxBQ0s+Jmd0OzwvRk9OVD48QlI+PEZPTlQgU0laRT00PlRvOiAmcXVvdDtMb2EgQW5k ZXJzc29uJnF1b3Q7ICZsdDs8L0ZPTlQ+PEZPTlQgU0laRT00IENPTE9SPUJMVUU+PFU+PEEgSFJF Rj1tYWlsdG86bG9hQHBpLnNlPmxvYUBwaS5zZTwvQT48L1U+PC9GT05UPjxGT05UIFNJWkU9NCBD T0xPUj1CTEFDSz4mZ3Q7PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Q2M6ICZsdDs8L0ZPTlQ+PEZP TlQgU0laRT00IENPTE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86Y2NhbXBAb3BzLmlldGYub3Jn PmNjYW1wQG9wcy5pZXRmLm9yZzwvQT48L1U+PC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTEFD Sz4mZ3Q7PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+U2VudDogRnJpZGF5LCBKdWx5IDIyLCAyMDA1 IDExOjQxIEFNPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+U3ViamVjdDogUmU6IEZyYW1lZm9ybWF0 IGluIGEgbDJjcyBnbXBscyBybnZpcm9ubWVudC48L0ZPTlQ+PEJSPjxCUj48QlI+PEZPTlQgU0la RT00PiZndDsgJmd0OyBQZXIgYW5kIERpbWl0cmksPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0 OyAmZ3Q7PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7IEkgd291bGQgbGlrZSB0byBj b21lIGRvd24gc3Ryb25nZXIgdGhhbiB0aGF0LCBmb3IgbWUgaXQgaXM8L0ZPTlQ+PEJSPjxGT05U IFNJWkU9ND4mZ3Q7ICZndDsgYSB2ZXJ5IHN0cm9uZyByZXF1aXJlbWVudCB0aGF0IHRoZSBzYW1l IHN3aXRjaCBjYW4gaGFuZGxlPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7IGJvdGgg VkxBTnMgYW5kIEdNUExzIHRyYWZpYyBjb3JyZWN0bHkuIEkgY2FuJ3QgZHNlZSBob3cgdGhhdDwv Rk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyBjb3VsZCBiZSBkb25lIGlmIHVzaW5nIHRo ZSBWTEFOIHRwaWQgdG8gaW5kaWNhdGUgR01QTFM8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7 ICZndDsgdHJhZmZpYy48L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7IDwvRk9OVD48QlI+PEZP TlQgU0laRT00PiZndDsgSWYgeW91IGV2ZXIgd2FudCB0aGF0IHNhbWUgc3dpdGNoIHRvIGhhbmRs ZSB0cmFmZmljIGZvciBhIGRpcmVjdGx5PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyBjb25u ZWN0ZWQmbmJzcDsgaG9zdCAobm90IHRvIHVuY29tbWVuKSB5b3Ugd291bGQgd2FudCB0aGF0IHRv IHVzZSBub3JtYWw8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7IGV0aGVybmV0IG1hY2FkZHJl c3Mgc3dpdGNoaW5nLiBTbyBvZiBjb3Vyc2UgeW91IGRvIG5vdCB3YW50IHRvIGhhdmUgdG88L0ZP TlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7IGNob29zZSBiZXR3ZWVuIHZsYW4gYW5kIGdtcGxzLCB5 b3Ugd291bGQgd2FudCBib3RoIGF0IHRoZSBzYW1lIHRpbWUuPC9GT05UPjxCUj48Rk9OVCBTSVpF PTQ+Jmd0OyA8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7IC9wZXI8L0ZPTlQ+PEJSPjxGT05U IFNJWkU9ND4mZ3Q7IDwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgPC9GT05UPjxCUj48Rk9O VCBTSVpFPTQ+Jmd0OyAmZ3Q7PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7IC9Mb2E8 L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDs8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4m Z3Q7ICZndDsgUCZhdW1sO3IgTWF0dHNzb24gd3JvdGU6PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+ Jmd0OyAmZ3Q7Jmd0OyZndDtoaSBwYXIsIG9uZSBvZiB0aGUgcG9zc2liaWxpdGllcyB0aGF0IGhh cyBiZWVuIGNvbnNpZGVyZWQgdG8gY29wZSB3aXRoPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0 OyAmZ3Q7Jmd0OyZndDt0aGlzIHJlcXVpcmVtZW50IGlzIHRvIHVzZSBhIGRlZGljYXRlZCBUUElE IGZvciB0aGUgRXRoZXJuZXQgbGFiZWxlZDwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0 OyZndDsmZ3Q7ZnJhbWVzOyB0aGlzIHdvdWxkIGFsbG93IGRpZmZlcmVudGlhdGVkIHByb2Nlc3Np bmcgd2l0aCBub24tbGFiZWxlZDwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyZndDsm Z3Q7ZnJhbWVzdGhhbmtzLjwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyZndDs8L0ZP TlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsmZ3Q7PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+ Jmd0OyAmZ3Q7Jmd0OyBUaGF0IHNlZW1zIHRvIG1ha2UgbW9yZSBzZW5jZS4gSWYgdGhhdCBmcmFt ZSBpcyB0byBiZSBzaXplZCBsaWtlIGE8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsm Z3Q7IDgwMi4xcTwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyZndDsgZnJhbWUuIFRo ZXJlIGlzIG5vdCB0aGF0IG11Y2ggc3BhY2UgbGVmdCB0byBhIGxhYmVsLiBPciBpcyB0aGUgZGVt YW5kIHRvPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7Jmd0OyB1c2UganVtYm9mcmFt ZXMgPzwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyZndDsgSGFzIHRoZXJlIGJlZW4g YW55IGRpc2N1c3Npb24gb24gbGFiZWxzdGFja2luZywgYW5kIG1haW5seSB3aGVyZSB0bzwvRk9O VD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyZndDsgcGxhY2U8L0ZPTlQ+PEJSPjxGT05UIFNJ WkU9ND4mZ3Q7ICZndDsmZ3Q7IHRoZSBpbmZvcm1hdGlvbj88L0ZPTlQ+PEJSPjxGT05UIFNJWkU9 ND4mZ3Q7ICZndDsmZ3Q7PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7Jmd0OyBSZWdh cmRzLjwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0OyZndDsgUGVyPC9GT05UPjxCUj48 Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7Jmd0OzwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDsgJmd0 OyZndDs8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsmZ3Q7PC9GT05UPjxCUj48Rk9O VCBTSVpFPTQ+Jmd0OyAmZ3Q7PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7PC9GT05U PjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7IC0tPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0 OyAmZ3Q7IExvYSBBbmRlcnNzb248L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDs8L0ZP TlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsgUHJpbmNpcGFsIE5ldHdvcmtpbmcgQXJjaGl0 ZWN0PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7IEFjcmVvIEFCJm5ic3A7Jm5ic3A7 Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5i c3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7 Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7IHBob25lOiZuYnNwOyArNDYgOCA2MzIgNzcg MTQ8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ICZndDsgSXNhZmpvcmRzZ2F0YW4gMjImbmJz cDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsm bmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsgbW9iaWxlOiArNDYgNzM5 IDgxIDIxIDY0PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyAmZ3Q7IEtpc3RhLCBTd2VkZW4m bmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJz cDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsm bmJzcDsmbmJzcDsgZW1haWw6Jm5ic3A7IDwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxVRT48 VT48QSBIUkVGPW1haWx0bzpsb2EuYW5kZXJzc29uQGFjcmVvLnNlPmxvYS5hbmRlcnNzb25AYWNy ZW8uc2U8L0E+PC9VPjwvRk9OVD48Rk9OVCBDT0xPUj1CTEFDSz48QlI+PC9GT05UPjxGT05UIFNJ WkU9ND4mZ3Q7ICZndDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsm bmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJz cDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsm bmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJz cDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsgPC9GT05U PjxGT05UIFNJWkU9NCBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOmxvYUBwaS5zZT5sb2FA cGkuc2U8L0E+PC9VPjwvRk9OVD48Rk9OVCBDT0xPUj1CTEFDSz48QlI+PC9GT05UPjxGT05UIFNJ WkU9ND4mZ3Q7ICZndDs8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7IDwvRk9OVD48QlI+PEZP TlQgU0laRT00PiZndDsgPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OyA8L0ZPTlQ+PEJSPjxG T05UIFNJWkU9ND4mZ3Q7IDwvRk9OVD48QlI+PEJSPjxGT05UIFNJWkU9NT49PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT08L0ZPTlQ+PEJS PjxGT05UIFNJWkU9NT5UaGUgaW5mb3JtYXRpb24gY29udGFpbmVkIGluIHRoaXMgbWVzc2FnZSBt YXkgYmUgcHJpdmlsZWdlZDwvRk9OVD48QlI+PEZPTlQgU0laRT01PmFuZCBjb25maWRlbnRpYWwg YW5kIHByb3RlY3RlZCBmcm9tIGRpc2Nsb3N1cmUuIElmIHRoZSByZWFkZXI8L0ZPTlQ+PEJSPjxG T05UIFNJWkU9NT5vZiB0aGlzIG1lc3NhZ2UgaXMgbm90IHRoZSBpbnRlbmRlZCByZWNpcGllbnQs IG9yIGFuIGVtcGxveWVlPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+b3IgYWdlbnQgcmVzcG9uc2li bGUgZm9yIGRlbGl2ZXJpbmcgdGhpcyBtZXNzYWdlIHRvIHRoZTwvRk9OVD48QlI+PEZPTlQgU0la RT01PmludGVuZGVkIHJlY2lwaWVudCwgeW91IGFyZSBoZXJlYnkgbm90aWZpZWQgdGhhdCBhbnkg cmVwcm9kdWN0aW9uLDwvRk9OVD48QlI+PEZPTlQgU0laRT01PmRpc3NlbWluYXRpb24gb3IgZGlz dHJpYnV0aW9uIG9mIHRoaXMgY29tbXVuaWNhdGlvbiBpcyBzdHJpY3RseTwvRk9OVD48QlI+PEZP TlQgU0laRT01PnByb2hpYml0ZWQuIElmIHlvdSBoYXZlIHJlY2VpdmVkIHRoaXMgY29tbXVuaWNh dGlvbiBpbiBlcnJvciw8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9NT5wbGVhc2Ugbm90aWZ5IHVzIGlt bWVkaWF0ZWx5IGJ5IHJlcGx5aW5nIHRvIHRoZSBtZXNzYWdlIGFuZDwvRk9OVD48QlI+PEZPTlQg U0laRT01PmRlbGV0aW5nIGl0IGZyb20geW91ciBjb21wdXRlci4gVGhhbmsgeW91LiBUZWxsYWJz PC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PC9GT05UPjxCUj48L1A+ Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 18:27:45 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01C58EEA.DBC1A8EF" Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 13:25:06 -0500 Message-ID: <A1A52203CA93634BA1748887B9993AEA01046A08@USNVEX1.tellabs-west.tellabsinc.net> Thread-Topic: Frameformat in a l2cs gmpls rnvironment. Thread-Index: AcWO4Z/sn5bUTdc8RiiIwKwSk2UxnAACGn2Q From: "Mack-Crane, T. Benjamin" <Ben.Mack-Crane@tellabs.com> To: "Igor Bryskin" <ibryskin@movaz.com>, <Dimitri.Papadimitriou@alcatel.be> Cc: "Adrian Farrel" <adrian@olddog.co.uk>, =?iso-8859-1?Q?P=E4r_Mattsson?= <per@defero.se>, "Loa Andersson" <loa@pi.se>, <ccamp@ops.ietf.org> This is a multi-part message in MIME format. ------_=_NextPart_001_01C58EEA.DBC1A8EF Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Igor, =20 I, too, am baffled by the notion of requirements for control of an unspecif= ied data plane. =20 Regards, Ben ________________________________ From: Igor Bryskin [mailto:ibryskin@movaz.com]=20 Sent: Friday, July 22, 2005 12:20 PM To: Dimitri.Papadimitriou@alcatel.be Cc: Mack-Crane, T. Benjamin; Adrian Farrel; P=E4r Mattsson; Loa Andersson;= ccamp@ops.ietf.org Subject: Re: Frameformat in a l2cs gmpls rnvironment. =09 =09 Dimitri, =20 The question is simple: is it possible today to statically provision L2LSP= s that could, say, support e2e QoS? If not, may be it is to early to discus= s aspects of dynamic provisioning of such LSPs? Is it possible/reasonable, = in your opinion, to "detail how forwarding information can be exchanged via= the control plane (and then installed)" without having in mind "a specific= forwarding behavior"? How do you know which forwarding information is needed for the forwarding = nobody has defined yet? The other question is do we need at all e2e QoS, route control, fast recov= ery? Sounds like exciting idea, but does all that commercially make sense? = So, I guess, we need two things before we can move forward: a) validation of the idea by the providers; b) definition of the data plane behavior; =20 Igor ----- Original Message -----=20 From: Dimitri.Papadimitriou@alcatel.be=20 To: Igor Bryskin <mailto:ibryskin@movaz.com> =20 Cc: Mack-Crane, T. Benjamin <mailto:Ben.Mack-Crane@tellabs.com> ; Adrian= Farrel <mailto:adrian@olddog.co.uk> ; P=E4r Mattsson <mailto:per@defero.s= e> ; Loa Andersson <mailto:loa@pi.se> ; ccamp@ops.ietf.org=20 Sent: Friday, July 22, 2005 12:21 PM Subject: Re: Frameformat in a l2cs gmpls rnvironment. =09 =09 igor - i am not sure which point they exactly have - the purpose of this = document is to detail how forwarding information can be exchanged via the c= ontrol plane (and then installed) not to define a specific forwarding behav= iour thanks, - dimitri. =09 "Igor Bryskin" <ibryskin@movaz.com> Sent by: owner-ccamp@ops.ietf.org 07/22/2005 11:10 AST =09 To: "Mack-Crane, T. Benjamin" <Ben.Mack-Crane@tellabs.com>, "Adrian Farre= l" <adrian@olddog.co.uk>, P=E4r Mattsson <per@defero.se>, "Loa Andersson" <= loa@pi.se> cc: <ccamp@ops.ietf.org> bcc:=20 Subject: Re: Frameformat in a l2cs gmpls rnvironment. =09 =09 =09 Hi, =09 I think Ben and Juergen have a point here. There is nothing that could be= dynamically provisioned that could not be also provisioned by management. = In other words data plane is the King here. We need data plane standard(s) = on how we encode a label, whether we need a label stack or not, how the lab= eled traffic is supposed to be treated, how labeled traffic co-exist with n= on-label traffic, etc. This is something that not for CCAMP to define. Take= for example TDM networks. GMPLS only provides a way to dynamically provisi= on G.707 networks. Hence there is a need in parallel standardization activi= ties in ITU. Only after that we can discuss how all that could be dynamical= ly provisioned, that is the aspects of control plane.=20 =09 Igor ----- Original Message -----=20 From: Mack-Crane, T. Benjamin <mailto:Ben.Mack-Crane@tellabs.com> =20 To: Adrian Farrel <mailto:adrian@olddog.co.uk> ; P=E4r Mattsson <mailto:= per@defero.se> ; Loa Andersson <mailto:loa@pi.se> =20 Cc: ccamp@ops.ietf.org=20 Sent: Friday, July 22, 2005 10:23 AM Subject: RE: Frameformat in a l2cs gmpls rnvironment. =09 Hi, =09 I think Juergen has raised an important question. How frames are labeled= (and the related data plane forwarding behavior) is not defined by the con= trol plane. The control plane serves to provision the data plane, not defi= ne it. In the framework draft it is not clear what data plane standards ar= e covered by the stated control plane requirements. Some references should= be supplied. In any case, the labeling and forwarding behavior should be = defined by these referenced standards, not by GMPLS. =09 (I am assuming definition of new data plane standards is out of scope for= CCAMP.) =09 Regards, Ben =09 ________________________________ =09 From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On Behal= f Of Adrian Farrel Sent: Friday, July 22, 2005 8:37 AM To: P=E4r Mattsson; Loa Andersson Cc: ccamp@ops.ietf.org Subject: Re: Frameformat in a l2cs gmpls rnvironment. =09 Hi, =09 It is always refreshing how engineers jump straight to a discussion of th= e solution :-) =09 Perhaps we can assume that the framework draft is a good representation o= f the problem space, and that we are ready to start discussing the solution= s=2E =09 One (perhaps the only?) significant question to answer is how the frames = will be labelled. This question is one we must come to as soon as we are co= nfident that the requirements need to be addressed at all. =09 As has been pointed out, there are several possibilities and to pick from= these we need to understand: 1. Do we need to support explicit label stacking? Note that this is not supported in TDM, LSC or FSC. 2. Do we need to be able to control (perhaps through an external signaling controller) existing hardware and=20 install LSPs through existing networks? 3. Do we need to support existing function simultaneous to the support of GMPLS L2 LSPs? =09 I think that from a chair's perspective I can give some limited guidance. =09 A. These questions must be raised and answered in the framework I-D B. The answer to question 3 is "yes". This means that the use of labels must not significantly deplete any namespace used to support other function. C. CCAMP is chartered to look at the control of transport networks. This includes Metro, but I am unsure about Campus. It does not cover=20 signaling to the desktop. =09 Thanks, Adrian =09 ----- Original Message -----=20 From: "P=E4r Mattsson" <per@defero.se> To: "Loa Andersson" <loa@pi.se> Cc: <ccamp@ops.ietf.org> Sent: Friday, July 22, 2005 11:41 AM Subject: Re: Frameformat in a l2cs gmpls rnvironment. =09 =09 > > Per and Dimitri, > > > > I would like to come down stronger than that, for me it is > > a very strong requirement that the same switch can handle > > both VLANs and GMPLs trafic correctly. I can't dsee how that > > could be done if using the VLAN tpid to indicate GMPLS > > traffic. >=20 > If you ever want that same switch to handle traffic for a directly > connected host (not to uncommen) you would want that to use normal > ethernet macaddress switching. So of course you do not want to have to > choose between vlan and gmpls, you would want both at the same time. >=20 > /per >=20 >=20 > > > > /Loa > > > > P=E4r Mattsson wrote: > >>>hi par, one of the possibilities that has been considered to cope wi= th > >>>this requirement is to use a dedicated TPID for the Ethernet labeled > >>>frames; this would allow differentiated processing with non-labeled > >>>framesthanks. > >> > >> > >> That seems to make more sence. If that frame is to be sized like a > >> 802.1q > >> frame. There is not that much space left to a label. Or is the deman= d to > >> use jumboframes ? > >> Has there been any discussion on labelstacking, and mainly where to > >> place > >> the information? > >> > >> Regards. > >> Per > >> > >> > >> > > > > > > -- > > Loa Andersson > > > > Principal Networking Architect > > Acreo AB phone: +46 8 632 77 14 > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > Kista, Sweden email: loa.andersson@acreo.se > > loa@pi.se > > >=20 >=20 >=20 >=20 =09 =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any reproduction, dissemination or distribution of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. Tellabs =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D =09 =09 =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any reproduction, dissemination or distribution of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. Tellabs =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D ------_=_NextPart_001_01C58EEA.DBC1A8EF Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META http-equiv=3DContent-Type content=3D"text/html; charset=3Diso-8859-1"> <META content=3D"MSHTML 6.00.2800.1106" name=3DGENERATOR> <STYLE></STYLE> </HEAD> <BODY bgColor=3D#ffffff> <DIV dir=3Dltr align=3Dleft><SPAN class=3D935282018-22072005><FONT face=3DA= rial=20 color=3D#0000ff size=3D2>Igor,</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D935282018-22072005><FONT face=3DA= rial=20 color=3D#0000ff size=3D2></FONT></SPAN> </DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D935282018-22072005><FONT face=3DA= rial=20 color=3D#0000ff size=3D2>I, too, am baffled by the notion of requirements f= or=20 control of an unspecified data plane.</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D935282018-22072005><FONT face=3DA= rial=20 color=3D#0000ff size=3D2></FONT></SPAN> </DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D935282018-22072005><FONT face=3DA= rial=20 color=3D#0000ff size=3D2>Regards,</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D935282018-22072005><FONT face=3DA= rial=20 color=3D#0000ff size=3D2>Ben</FONT></SPAN></DIV><BR> <BLOCKQUOTE dir=3Dltr=20 style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #0000ff 2px soli= d; MARGIN-RIGHT: 0px"> <DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr align=3Dleft> <HR tabIndex=3D-1> <FONT face=3DTahoma size=3D2><B>From:</B> Igor Bryskin [mailto:ibryskin@m= ovaz.com]=20 <BR><B>Sent:</B> Friday, July 22, 2005 12:20 PM<BR><B>To:</B>=20 Dimitri.Papadimitriou@alcatel.be<BR><B>Cc:</B> Mack-Crane, T. Benjamin; A= drian=20 Farrel; P=E4r Mattsson; Loa Andersson; ccamp@ops.ietf.org<BR><B>Subject:<= /B> Re:=20 Frameformat in a l2cs gmpls rnvironment.<BR></FONT><BR></DIV> <DIV></DIV> <DIV><FONT face=3DArial size=3D2>Dimitri,</FONT></DIV> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <DIV><FONT face=3DArial size=3D2>The question is simple: is it possible t= oday to=20 statically provision L2LSPs that could, say, support e2e QoS? If not= , may=20 be it is to early to discuss aspects of dynamic provisioning of such LSPs= ? Is=20 it possible/reasonable, in your opinion, to "<FONT face=3D"Courier New"=20 size=3D3>detail how forwarding information can be exchanged via the contr= ol=20 plane (and then installed)" without having in mind "a specific forwa= rding=20 behavior"?</FONT></FONT></DIV> <DIV><FONT face=3D"Courier New">How do you know which forwarding informat= ion is=20 needed for the forwarding nobody has defined yet?</FONT></DIV> <DIV><FONT face=3D"Courier New">The other question is do we need at = all e2e=20 QoS, route control, fast recovery? Sounds like exciting idea, but does al= l=20 that commercially make sense? So, I guess, we need two things before= we=20 can move forward:</FONT></DIV> <DIV><FONT face=3D"Courier New">a) validation of the idea by the=20 providers;</FONT></DIV> <DIV><FONT face=3D"Courier New">b) definition of the data plane=20 behavior;</FONT></DIV> <DIV><FONT face=3D"Courier New"></FONT> </DIV> <DIV><FONT face=3D"Courier New">Igor</FONT></DIV> <BLOCKQUOTE dir=3Dltr=20 style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-= LEFT: #000000 2px solid; MARGIN-RIGHT: 0px"> <DIV style=3D"FONT: 10pt arial">----- Original Message ----- </DIV> <DIV=20 style=3D"BACKGROUND: #e4e4e4; FONT: 10pt arial; font-color: black"><B>F= rom:</B>=20 <A title=3DDimitri.Papadimitriou@alcatel.be=20 href=3D"mailto:Dimitri.Papadimitriou@alcatel.be">Dimitri.Papadimitriou@= alcatel.be</A>=20 </DIV> <DIV style=3D"FONT: 10pt arial"><B>To:</B> <A title=3Dibryskin@movaz.co= m=20 href=3D"mailto:ibryskin@movaz.com">Igor Bryskin</A> </DIV> <DIV style=3D"FONT: 10pt arial"><B>Cc:</B> <A title=3DBen.Mack-Crane@te= llabs.com=20 href=3D"mailto:Ben.Mack-Crane@tellabs.com">Mack-Crane, T. Benjamin</A> = ; <A=20 title=3Dadrian@olddog.co.uk href=3D"mailto:adrian@olddog.co.uk">Adrian=20 Farrel</A> ; <A title=3Dper@defero.se href=3D"mailto:per@defero.se">P= =E4r=20 Mattsson</A> ; <A title=3Dloa@pi.se href=3D"mailto:loa@pi.se">Loa Ander= sson</A>=20 ; <A title=3Dccamp@ops.ietf.org=20 href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A> </DIV> <DIV style=3D"FONT: 10pt arial"><B>Sent:</B> Friday, July 22, 2005 12:2= 1=20 PM</DIV> <DIV style=3D"FONT: 10pt arial"><B>Subject:</B> Re: Frameformat in a l2= cs=20 gmpls rnvironment.</DIV> <DIV><FONT face=3DArial size=3D2></FONT><BR></DIV> <P><FONT face=3DMonospace,Courier>igor - i am not sure which point they=20 exactly have - the purpose of this document is to detail how forwarding=20 information can be exchanged via the control plane (and then installed)= not=20 to define a specific forwarding behaviour</FONT></P> <P><FONT face=3DMonospace,Courier>thanks,</FONT></P> <P><FONT face=3DMonospace,Courier>- dimitri.</FONT><BR><BR><FONT=20 size=3D2><B>"Igor Bryskin" <<A=20 href=3D"mailto:ibryskin@movaz.com">ibryskin@movaz.com</A>></B></FONT= ><BR><FONT=20 size=3D2>Sent by: <A=20 href=3D"mailto:owner-ccamp@ops.ietf.org">owner-ccamp@ops.ietf.org</A></= FONT><BR><FONT=20 size=3D2>07/22/2005 11:10 AST</FONT><BR><BR><FONT size=3D2>To:</FONT> <= FONT=20 size=3D2>"Mack-Crane, T. Benjamin" <<A=20 href=3D"mailto:Ben.Mack-Crane@tellabs.com">Ben.Mack-Crane@tellabs.com</= A>>,=20 "Adrian Farrel" <<A=20 href=3D"mailto:adrian@olddog.co.uk">adrian@olddog.co.uk</A>>, P=E4r = Mattsson=20 <<A href=3D"mailto:per@defero.se">per@defero.se</A>>, "Loa Anders= son"=20 <<A href=3D"mailto:loa@pi.se">loa@pi.se</A>></FONT><BR><FONT=20 size=3D2>cc:</FONT> <FONT size=3D2><<A=20 href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A>></FONT><BR= ><FONT=20 size=3D2>bcc:</FONT> <BR><FONT size=3D2>Subject:</FONT> <FONT size=3D2>= Re:=20 Frameformat in a l2cs gmpls rnvironment.</FONT><BR><BR><BR></P> <P>Hi,<BR><BR>I think Ben and Juergen have a point here. There is nothi= ng=20 that could be dynamically provisioned that could not be also provisione= d by=20 management. In other words data plane is the King here. We need data pl= ane=20 standard(s) on how we encode a label, whether we need a label stack or = not,=20 how the labeled traffic is supposed to be treated, how labeled tra= ffic=20 co-exist with non-label traffic, etc. This is something that not for CC= AMP=20 to define. Take for example TDM networks. GMPLS only provides a way to=20 dynamically provision G.707 networks. Hence there is a need in parallel=20 standardization activities in ITU. Only after that we can discuss how a= ll=20 that could be dynamically provisioned, that is the aspects of control p= lane.=20 <BR><BR>Igor<BR><FONT size=3D4>----- Original Message ----- </FONT><BR>= <FONT=20 size=3D4><B>From:</B> </FONT><FONT color=3Dblue size=3D4><U><A=20 href=3D"mailto:Ben.Mack-Crane@tellabs.com">Mack-Crane, T.=20 Benjamin</A></U></FONT><FONT color=3Dblack size=3D4> </FONT><BR><FONT=20 size=3D4><B>To:</B> </FONT><FONT color=3Dblue size=3D4><U><A=20 href=3D"mailto:adrian@olddog.co.uk">Adrian Farrel</A></U></FONT><FONT=20 color=3Dblack size=3D4> ; </FONT><FONT color=3Dblue size=3D4><U><A=20 href=3D"mailto:per@defero.se">P=E4r Mattsson</A></U></FONT><FONT color= =3Dblack=20 size=3D4> ; </FONT><FONT color=3Dblue size=3D4><U><A href=3D"mailto:loa= @pi.se">Loa=20 Andersson</A></U></FONT><FONT color=3Dblack size=3D4> </FONT><BR><FONT=20 size=3D4><B>Cc:</B> </FONT><FONT color=3Dblue size=3D4><U><A=20 href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A></U></FONT><FO= NT=20 color=3Dblack size=3D4> </FONT><BR><FONT size=3D4><B>Sent:</B> Friday, = July 22,=20 2005 10:23 AM</FONT><BR><FONT size=3D4><B>Subject:</B> RE: Frameformat = in a=20 l2cs gmpls rnvironment.</FONT><BR><BR>Hi,<BR><BR>I think Juergen has ra= ised=20 an important question. How frames are labeled (and the related da= ta=20 plane forwarding behavior) is not defined by the control plane. T= he=20 control plane serves to provision the data plane, not define it. = In=20 the framework draft it is not clear what data plane standards are cover= ed by=20 the stated control plane requirements. Some references should be=20 supplied. In any case, the labeling and forwarding behavior shoul= d be=20 defined by these referenced standards, not by GMPLS.<BR><BR>(= I am=20 assuming definition of new data plane standards is out of scope for=20 CCAMP.)<BR><BR>Regards,<BR>Ben<BR> <HR align=3Dleft width=3D"99%" SIZE=3D4> <FONT color=3Dblack><BR><B>From:</B> owner-ccamp@ops.ietf.org=20 [mailto:owner-ccamp@ops.ietf.org] <B>On Behalf Of </B>Adrian=20 Farrel<BR><B>Sent:</B> Friday, July 22, 2005 8:37 AM<BR><B>To:</B> P=E4= r=20 Mattsson; Loa Andersson<BR><B>Cc:</B> ccamp@ops.ietf.org<BR><B>Subject:= </B>=20 Re: Frameformat in a l2cs gmpls rnvironment.<BR><BR>Hi,<BR><BR>It is al= ways=20 refreshing how engineers jump straight to a discussion of the solution=20 :-)<BR><BR>Perhaps we can assume that the framework draft is a good=20 representation of the problem space, and that we are ready to start=20 discussing the solutions.<BR><BR>One (perhaps the only?) significant=20 question to answer is how the frames will be labelled. This question is= one=20 we must come to as soon as we are confident that the requirements need = to be=20 addressed at all.<BR><BR>As has been pointed out, there are several=20 possibilities and to pick from these we need to understand:<BR>1. Do we= need=20 to support explicit label stacking?<BR> Note that this is n= ot=20 supported in TDM, LSC or FSC.<BR>2. Do we need to be able to control=20 (perhaps through an<BR> external signaling controller)=20 existing hardware and <BR> install LSPs through existing=20 networks?<BR>3. Do we need to support existing function simultaneous=20 to<BR> the support of GMPLS L2 LSPs?<BR><BR>I think that fr= om a=20 chair's perspective I can give some limited guidance.<BR><BR>A. These=20 questions must be raised and answered in the framework I-D<BR>B. The an= swer=20 to question 3 is "yes". This means that the use of labels<BR> &nbs= p;=20 must not significantly deplete any namespace used to support=20 other<BR> function.<BR>C. CCAMP is chartered to look at the=20 control of transport networks. This<BR> includes Metro= , but=20 I am unsure about Campus. It does not cover <BR> signa= ling=20 to the desktop.<BR><BR>Thanks,<BR>Adrian<BR><BR>----- Original Message = -----=20 <BR>From: "P=E4r Mattsson" <</FONT><FONT color=3Dblue><U><A=20 href=3D"mailto:per@defero.se">per@defero.se</A></U></FONT><FONT=20 color=3Dblack>><BR>To: "Loa Andersson" <</FONT><FONT color=3Dblue= ><U><A=20 href=3D"mailto:loa@pi.se">loa@pi.se</A></U></FONT><FONT=20 color=3Dblack>><BR>Cc: <</FONT><FONT color=3Dblue><U><A=20 href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A></U></FONT><FO= NT=20 color=3Dblack>><BR>Sent: Friday, July 22, 2005 11:41 AM<BR>Subject: = Re:=20 Frameformat in a l2cs gmpls rnvironment.<BR><BR><BR>> > Per and=20 Dimitri,<BR>> ><BR>> > I would like to come down stronger t= han=20 that, for me it is<BR>> > a very strong requirement that the same=20 switch can handle<BR>> > both VLANs and GMPLs trafic correctly. I=20 can't dsee how that<BR>> > could be done if using the VLAN tpid t= o=20 indicate GMPLS<BR>> > traffic.<BR>> <BR>> If you ever want = that=20 same switch to handle traffic for a directly<BR>> connected ho= st=20 (not to uncommen) you would want that to use normal<BR>> ethernet=20 macaddress switching. So of course you do not want to have to<BR>> c= hoose=20 between vlan and gmpls, you would want both at the same time.<BR>>=20 <BR>> /per<BR>> <BR>> <BR>> ><BR>> > /Loa<BR>>=20 ><BR>> > P=E4r Mattsson wrote:<BR>> >>>hi par, one= of the=20 possibilities that has been considered to cope with<BR>> >>>= ;this=20 requirement is to use a dedicated TPID for the Ethernet labeled<BR>>=20 >>>frames; this would allow differentiated processing with=20 non-labeled<BR>> >>>framesthanks.<BR>> >><BR>>=20 >><BR>> >> That seems to make more sence. If that frame = is to=20 be sized like a<BR>> >> 802.1q<BR>> >> frame. There i= s not=20 that much space left to a label. Or is the demand to<BR>> >> u= se=20 jumboframes ?<BR>> >> Has there been any discussion on=20 labelstacking, and mainly where to<BR>> >> place<BR>> >&= gt;=20 the information?<BR>> >><BR>> >> Regards.<BR>> >= ;>=20 Per<BR>> >><BR>> >><BR>> >><BR>> ><BR>= >=20 ><BR>> > --<BR>> > Loa Andersson<BR>> ><BR>> &g= t;=20 Principal Networking Architect<BR>> > Acreo=20 AB &nb= sp; = =20 phone: +46 8 632 77 14<BR>> > Isafjordsgatan=20 22 &nb= sp; =20 mobile: +46 739 81 21 64<BR>> > Kista,=20 Sweden  = ; =20 email: </FONT><FONT color=3Dblue><U><A=20 href=3D"mailto:loa.andersson@acreo.se">loa.andersson@acreo.se</A></U></= FONT><FONT=20 color=3Dblack><BR>>=20 > &= nbsp; &nbs= p; &= nbsp; =20 </FONT><FONT color=3Dblue><U><A=20 href=3D"mailto:loa@pi.se">loa@pi.se</A></U></FONT><FONT color=3Dblack><= BR>>=20 ><BR>> <BR>> <BR>> <BR>> <BR><BR></FONT><FONT=20 face=3DMonospace,Courier=20 size=3D4>=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D</FONT><BR><FONT=20 face=3DMonospace,Courier size=3D4>The information contained in this mes= sage may=20 be privileged</FONT><BR><FONT face=3DMonospace,Courier size=3D4>and con= fidential=20 and protected from disclosure. If the reader</FONT><BR><FONT=20 face=3DMonospace,Courier size=3D4>of this message is not the intended r= ecipient,=20 or an employee</FONT><BR><FONT face=3DMonospace,Courier size=3D4>or age= nt=20 responsible for delivering this message to the</FONT><BR><FONT=20 face=3DMonospace,Courier size=3D4>intended recipient, you are hereby no= tified=20 that any reproduction,</FONT><BR><FONT face=3DMonospace,Courier=20 size=3D4>dissemination or distribution of this communication is=20 strictly</FONT><BR><FONT face=3DMonospace,Courier size=3D4>prohibited. = If you=20 have received this communication in error,</FONT><BR><FONT=20 face=3DMonospace,Courier size=3D4>please notify us immediately by reply= ing to=20 the message and</FONT><BR><FONT face=3DMonospace,Courier size=3D4>delet= ing it=20 from your computer. Thank you. Tellabs</FONT><BR><FONT=20 face=3DMonospace,Courier=20 size=3D4>=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D</FONT><BR> <P></P></BLOCKQUOTE></BLOCKQUOTE><pre>=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any reproduction, dissemination or distribution of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. Tellabs =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D </pre></BODY></HTML> ------_=_NextPart_001_01C58EEA.DBC1A8EF-- Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 18:12:35 +0000 Message-ID: <001901c58ee9$0be728f0$b7849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <richard.spencer@bt.com>, <juergen.heiles@siemens.com>, <loa@pi.se>, <per@defero.se> Cc: <ccamp@ops.ietf.org> Subject: Ethernet Control Plane [Was: Re: Frameformat in a l2cs gmpls rnvironment.] Date: Fri, 22 Jul 2005 19:13:11 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 8bit Whoa there! GMPLS assumes an IP control plane. I am aware that there is work afoot in other SDOs to assign a dedicated VLAN tag to specify the control plane traffic. This is not what we are doing. As far as I am aware, everyone in the IETF is comfortable with the idea of using IP packets to carry control plane data through an IP-enabled DCN. In Ethernet the issue will be whether the control plane is in band (use IP address and MAC address) or out of band (use distinct DCN). The use of a separate VLAN ID to signify the control plane is out of scope for CCAMP. Thanks, Adrian ----- Original Message ----- From: <richard.spencer@bt.com> To: <juergen.heiles@siemens.com>; <loa@pi.se>; <per@defero.se> Cc: <ccamp@ops.ietf.org> Sent: Friday, July 22, 2005 4:33 PM Subject: RE: Frameformat in a l2cs gmpls rnvironment. Juergen, Before you can use GMPLS to set up VLANs a data plane must first exist to carry the GMPLS control plane traffic. Yes you can use the standard Ethernet data plane for this purpose (and in fact must do so for the solution to be viable), but its not as simple as that. You cannot assume that just because two GMPLS peers are directly connected that they can communicate with each other. To exchange GMPLS control plane packets between a pair of GMPLS peers, the peers will need to agree to use either untagged control packets or to tag them with a common VLAN ID. The use of a separate VLAN for GMPLS control traffic is analogous to using a separate VLAN for management traffic today. If two switches belong to the same management VLAN then you can use ICMP ping to verify connectivity between them. However, if one switch has been put into the wrong management VLAN then the ping will obviously fail. If you want to use a separate VLAN for GMPLS control plane traffic, then you either need to (i) define a reserved GMPLS signalling VLAN ID, or (ii) leave it up to the operator to choose a VLAN ID. If you leave it up to the operator to decide what VLAN ID to use, then you need to consider what the implications in the inter-operator case will be, i.e. what VLAN ID range should be used for control plane traffic and what VLAN ID range should be used for user plane traffic. I'm not saying that defining how GMPLS control plane packets are forwarded in the Ethernet data plane will be an issue, simply that this is something that must be addressed to develop interoperable solutions. Regards, Richard > -----Original Message----- > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On > Behalf Of Heiles Juergen > Sent: 22 July 2005 14:17 > To: Loa Andersson; Pär Mattsson > Cc: ccamp@ops.ietf.org > Subject: RE: Frameformat in a l2cs gmpls rnvironment. > > > GMPLS for SONET/SDH has no impact on the SONET/SDH data > plane. It introduces a control plane for connection setup > instead of a network management based connection setup. GMPLS > for Ethernet can work in the same way. Ethernet as such is > cl, but VLANs are setup and GMPLS can provide a control plane > for VLAN setup. The Ethernet data plane is not changed and > MAC address learning and forwarding is still be done within a > VLAN. So no change to the Ethernet data plane. No GMPLS > traffic is introduced, it is still VLAN traffic. > Introducing a dedicated GMPLS label for Ethernet (GMPLS > traffic) goes beyond GMPLS as a control plane technique. > > Regards > > Juergen > > > > > > -----Original Message----- > > From: owner-ccamp@ops.ietf.org > > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson > > Sent: Friday, July 22, 2005 12:05 PM > > To: Pär Mattsson > > Cc: ccamp@ops.ietf.org > > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > > > Per and Dimitri, > > > > I would like to come down stronger than that, for me it is > > a very strong requirement that the same switch can handle > > both VLANs and GMPLs trafic correctly. I can't dsee how that > > could be done if using the VLAN tpid to indicate GMPLS > > traffic. > > > > /Loa > > > > Pär Mattsson wrote: > > >>hi par, one of the possibilities that has been considered > > to cope with > > >>this requirement is to use a dedicated TPID for the > Ethernet labeled > > >>frames; this would allow differentiated processing with > non-labeled > > >>framesthanks. > > > > > > > > > That seems to make more sence. If that frame is to be sized > > like a 802.1q > > > frame. There is not that much space left to a label. Or is > > the demand to > > > use jumboframes ? > > > Has there been any discussion on labelstacking, and mainly > > where to place > > > the information? > > > > > > Regards. > > > Per > > > > > > > > > > > > > > > -- > > Loa Andersson > > > > Principal Networking Architect > > Acreo AB phone: +46 8 632 77 14 > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > Kista, Sweden email: loa.andersson@acreo.se > > loa@pi.se > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 17:21:21 +0000 Message-ID: <017f01c58ee1$a1afdc90$6501a8c0@movaz.com> From: "Igor Bryskin" <ibryskin@movaz.com> To: <Dimitri.Papadimitriou@alcatel.be> Cc: "Mack-Crane, T. Benjamin" <Ben.Mack-Crane@tellabs.com>, "Adrian Farrel" <adrian@olddog.co.uk>, =?iso-8859-1?Q?P=E4r_Mattsson?= <per@defero.se>, "Loa Andersson" <loa@pi.se>, <ccamp@ops.ietf.org> Subject: Re: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 13:19:53 -0400 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_017C_01C58EC0.0C02EAF0" This is a multi-part message in MIME format. ------=_NextPart_000_017C_01C58EC0.0C02EAF0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Dimitri, The question is simple: is it possible today to statically provision = L2LSPs that could, say, support e2e QoS? If not, may be it is to early = to discuss aspects of dynamic provisioning of such LSPs? Is it = possible/reasonable, in your opinion, to "detail how forwarding = information can be exchanged via the control plane (and then installed)" = without having in mind "a specific forwarding behavior"? How do you know which forwarding information is needed for the = forwarding nobody has defined yet? The other question is do we need at all e2e QoS, route control, fast = recovery? Sounds like exciting idea, but does all that commercially make = sense? So, I guess, we need two things before we can move forward: a) validation of the idea by the providers; b) definition of the data plane behavior; Igor ----- Original Message -----=20 From: Dimitri.Papadimitriou@alcatel.be=20 To: Igor Bryskin=20 Cc: Mack-Crane, T. Benjamin ; Adrian Farrel ; P=E4r Mattsson ; Loa = Andersson ; ccamp@ops.ietf.org=20 Sent: Friday, July 22, 2005 12:21 PM Subject: Re: Frameformat in a l2cs gmpls rnvironment. igor - i am not sure which point they exactly have - the purpose of = this document is to detail how forwarding information can be exchanged = via the control plane (and then installed) not to define a specific = forwarding behaviour thanks, - dimitri. "Igor Bryskin" <ibryskin@movaz.com> Sent by: owner-ccamp@ops.ietf.org 07/22/2005 11:10 AST To: "Mack-Crane, T. Benjamin" <Ben.Mack-Crane@tellabs.com>, "Adrian = Farrel" <adrian@olddog.co.uk>, P=E4r Mattsson <per@defero.se>, "Loa = Andersson" <loa@pi.se> cc: <ccamp@ops.ietf.org> bcc:=20 Subject: Re: Frameformat in a l2cs gmpls rnvironment. Hi, I think Ben and Juergen have a point here. There is nothing that could = be dynamically provisioned that could not be also provisioned by = management. In other words data plane is the King here. We need data = plane standard(s) on how we encode a label, whether we need a label = stack or not, how the labeled traffic is supposed to be treated, how = labeled traffic co-exist with non-label traffic, etc. This is something = that not for CCAMP to define. Take for example TDM networks. GMPLS only = provides a way to dynamically provision G.707 networks. Hence there is a = need in parallel standardization activities in ITU. Only after that we = can discuss how all that could be dynamically provisioned, that is the = aspects of control plane.=20 Igor ----- Original Message -----=20 From: Mack-Crane, T. Benjamin=20 To: Adrian Farrel ; P=E4r Mattsson ; Loa Andersson=20 Cc: ccamp@ops.ietf.org=20 Sent: Friday, July 22, 2005 10:23 AM Subject: RE: Frameformat in a l2cs gmpls rnvironment. Hi, I think Juergen has raised an important question. How frames are = labeled (and the related data plane forwarding behavior) is not defined = by the control plane. The control plane serves to provision the data = plane, not define it. In the framework draft it is not clear what data = plane standards are covered by the stated control plane requirements. = Some references should be supplied. In any case, the labeling and = forwarding behavior should be defined by these referenced standards, not = by GMPLS. (I am assuming definition of new data plane standards is out of scope = for CCAMP.) Regards, Ben -------------------------------------------------------------------------= ----- From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On = Behalf Of Adrian Farrel Sent: Friday, July 22, 2005 8:37 AM To: P=E4r Mattsson; Loa Andersson Cc: ccamp@ops.ietf.org Subject: Re: Frameformat in a l2cs gmpls rnvironment. Hi, It is always refreshing how engineers jump straight to a discussion of = the solution :-) Perhaps we can assume that the framework draft is a good = representation of the problem space, and that we are ready to start = discussing the solutions. One (perhaps the only?) significant question to answer is how the = frames will be labelled. This question is one we must come to as soon as = we are confident that the requirements need to be addressed at all. As has been pointed out, there are several possibilities and to pick = from these we need to understand: 1. Do we need to support explicit label stacking? Note that this is not supported in TDM, LSC or FSC. 2. Do we need to be able to control (perhaps through an external signaling controller) existing hardware and=20 install LSPs through existing networks? 3. Do we need to support existing function simultaneous to the support of GMPLS L2 LSPs? I think that from a chair's perspective I can give some limited = guidance. A. These questions must be raised and answered in the framework I-D B. The answer to question 3 is "yes". This means that the use of = labels must not significantly deplete any namespace used to support other function. C. CCAMP is chartered to look at the control of transport networks. = This includes Metro, but I am unsure about Campus. It does not cover=20 signaling to the desktop. Thanks, Adrian ----- Original Message -----=20 From: "P=E4r Mattsson" <per@defero.se> To: "Loa Andersson" <loa@pi.se> Cc: <ccamp@ops.ietf.org> Sent: Friday, July 22, 2005 11:41 AM Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > Per and Dimitri, > > > > I would like to come down stronger than that, for me it is > > a very strong requirement that the same switch can handle > > both VLANs and GMPLs trafic correctly. I can't dsee how that > > could be done if using the VLAN tpid to indicate GMPLS > > traffic. >=20 > If you ever want that same switch to handle traffic for a directly > connected host (not to uncommen) you would want that to use normal > ethernet macaddress switching. So of course you do not want to have = to > choose between vlan and gmpls, you would want both at the same time. >=20 > /per >=20 >=20 > > > > /Loa > > > > P=E4r Mattsson wrote: > >>>hi par, one of the possibilities that has been considered to cope = with > >>>this requirement is to use a dedicated TPID for the Ethernet = labeled > >>>frames; this would allow differentiated processing with = non-labeled > >>>framesthanks. > >> > >> > >> That seems to make more sence. If that frame is to be sized like = a > >> 802.1q > >> frame. There is not that much space left to a label. Or is the = demand to > >> use jumboframes ? > >> Has there been any discussion on labelstacking, and mainly where = to > >> place > >> the information? > >> > >> Regards. > >> Per > >> > >> > >> > > > > > > -- > > Loa Andersson > > > > Principal Networking Architect > > Acreo AB phone: +46 8 632 77 14 > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > Kista, Sweden email: loa.andersson@acreo.se > > loa@pi.se > > >=20 >=20 >=20 >=20 = =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any reproduction, dissemination or distribution of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. Tellabs = =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D ------=_NextPart_000_017C_01C58EC0.0C02EAF0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META http-equiv=3DContent-Type content=3D"text/html; = charset=3Diso-8859-1"> <META content=3D"MSHTML 6.00.2800.1400" name=3DGENERATOR> <STYLE></STYLE> </HEAD> <BODY bgColor=3D#ffffff> <DIV><FONT face=3DArial size=3D2>Dimitri,</FONT></DIV> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <DIV><FONT face=3DArial size=3D2>The question is simple: is it possible = today to=20 statically provision L2LSPs that could, say, support e2e QoS? If = not, may=20 be it is to early to discuss aspects of dynamic provisioning of such = LSPs? Is it=20 possible/reasonable, in your opinion, to "<FONT face=3D"Courier New" = size=3D3>detail=20 how forwarding information can be exchanged via the control plane (and = then=20 installed)" without having in mind "a specific forwarding=20 behavior"?</FONT></FONT></DIV> <DIV><FONT face=3D"Courier New">How do you know which forwarding = information is=20 needed for the forwarding nobody has defined yet?</FONT></DIV> <DIV><FONT face=3D"Courier New">The other question is do we need at = all e2e=20 QoS, route control, fast recovery? Sounds like exciting idea, but does = all that=20 commercially make sense? So, I guess, we need two things before we=20 can move forward:</FONT></DIV> <DIV><FONT face=3D"Courier New">a) validation of the idea by the=20 providers;</FONT></DIV> <DIV><FONT face=3D"Courier New">b) definition of the data plane=20 behavior;</FONT></DIV> <DIV><FONT face=3D"Courier New"></FONT> </DIV> <DIV><FONT face=3D"Courier New">Igor</FONT></DIV> <BLOCKQUOTE dir=3Dltr=20 style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; = BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px"> <DIV style=3D"FONT: 10pt arial">----- Original Message ----- </DIV> <DIV=20 style=3D"BACKGROUND: #e4e4e4; FONT: 10pt arial; font-color: = black"><B>From:</B>=20 <A title=3DDimitri.Papadimitriou@alcatel.be=20 = href=3D"mailto:Dimitri.Papadimitriou@alcatel.be">Dimitri.Papadimitriou@al= catel.be</A>=20 </DIV> <DIV style=3D"FONT: 10pt arial"><B>To:</B> <A = title=3Dibryskin@movaz.com=20 href=3D"mailto:ibryskin@movaz.com">Igor Bryskin</A> </DIV> <DIV style=3D"FONT: 10pt arial"><B>Cc:</B> <A = title=3DBen.Mack-Crane@tellabs.com=20 href=3D"mailto:Ben.Mack-Crane@tellabs.com">Mack-Crane, T. Benjamin</A> = ; <A=20 title=3Dadrian@olddog.co.uk href=3D"mailto:adrian@olddog.co.uk">Adrian = Farrel</A>=20 ; <A title=3Dper@defero.se href=3D"mailto:per@defero.se">P=E4r = Mattsson</A> ; <A=20 title=3Dloa@pi.se href=3D"mailto:loa@pi.se">Loa Andersson</A> ; <A=20 title=3Dccamp@ops.ietf.org=20 href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A> </DIV> <DIV style=3D"FONT: 10pt arial"><B>Sent:</B> Friday, July 22, 2005 = 12:21=20 PM</DIV> <DIV style=3D"FONT: 10pt arial"><B>Subject:</B> Re: Frameformat in a = l2cs gmpls=20 rnvironment.</DIV> <DIV><FONT face=3DArial size=3D2></FONT><BR></DIV> <P><FONT face=3DMonospace,Courier>igor - i am not sure which point = they exactly=20 have - the purpose of this document is to detail how forwarding = information=20 can be exchanged via the control plane (and then installed) not to = define a=20 specific forwarding behaviour</FONT></P> <P><FONT face=3DMonospace,Courier>thanks,</FONT></P> <P><FONT face=3DMonospace,Courier>- dimitri.</FONT><BR><BR><FONT = size=3D2><B>"Igor=20 Bryskin" <<A=20 = href=3D"mailto:ibryskin@movaz.com">ibryskin@movaz.com</A>></B></FONT><= BR><FONT=20 size=3D2>Sent by: <A=20 = href=3D"mailto:owner-ccamp@ops.ietf.org">owner-ccamp@ops.ietf.org</A></FO= NT><BR><FONT=20 size=3D2>07/22/2005 11:10 AST</FONT><BR><BR><FONT size=3D2>To:</FONT> = <FONT=20 size=3D2>"Mack-Crane, T. Benjamin" <<A=20 = href=3D"mailto:Ben.Mack-Crane@tellabs.com">Ben.Mack-Crane@tellabs.com</A>= >,=20 "Adrian Farrel" <<A=20 href=3D"mailto:adrian@olddog.co.uk">adrian@olddog.co.uk</A>>, P=E4r = Mattsson=20 <<A href=3D"mailto:per@defero.se">per@defero.se</A>>, "Loa = Andersson"=20 <<A href=3D"mailto:loa@pi.se">loa@pi.se</A>></FONT><BR><FONT=20 size=3D2>cc:</FONT> <FONT size=3D2><<A=20 = href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A>></FONT><BR><= FONT=20 size=3D2>bcc:</FONT> <BR><FONT size=3D2>Subject:</FONT> <FONT = size=3D2>Re:=20 Frameformat in a l2cs gmpls rnvironment.</FONT><BR><BR><BR></P> <P>Hi,<BR><BR>I think Ben and Juergen have a point here. There is = nothing that=20 could be dynamically provisioned that could not be also provisioned by = management. In other words data plane is the King here. We need data = plane=20 standard(s) on how we encode a label, whether we need a label stack or = not,=20 how the labeled traffic is supposed to be treated, how labeled = traffic=20 co-exist with non-label traffic, etc. This is something that not for = CCAMP to=20 define. Take for example TDM networks. GMPLS only provides a way to=20 dynamically provision G.707 networks. Hence there is a need in = parallel=20 standardization activities in ITU. Only after that we can discuss how = all that=20 could be dynamically provisioned, that is the aspects of control = plane.=20 <BR><BR>Igor<BR><FONT size=3D4>----- Original Message ----- = </FONT><BR><FONT=20 size=3D4><B>From:</B> </FONT><FONT color=3Dblue size=3D4><U><A=20 href=3D"mailto:Ben.Mack-Crane@tellabs.com">Mack-Crane, T.=20 Benjamin</A></U></FONT><FONT color=3Dblack size=3D4> </FONT><BR><FONT=20 size=3D4><B>To:</B> </FONT><FONT color=3Dblue size=3D4><U><A=20 href=3D"mailto:adrian@olddog.co.uk">Adrian Farrel</A></U></FONT><FONT=20 color=3Dblack size=3D4> ; </FONT><FONT color=3Dblue size=3D4><U><A=20 href=3D"mailto:per@defero.se">P=E4r Mattsson</A></U></FONT><FONT = color=3Dblack=20 size=3D4> ; </FONT><FONT color=3Dblue size=3D4><U><A = href=3D"mailto:loa@pi.se">Loa=20 Andersson</A></U></FONT><FONT color=3Dblack size=3D4> </FONT><BR><FONT = size=3D4><B>Cc:</B> </FONT><FONT color=3Dblue size=3D4><U><A=20 = href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A></U></FONT><FONT= =20 color=3Dblack size=3D4> </FONT><BR><FONT size=3D4><B>Sent:</B> Friday, = July 22, 2005=20 10:23 AM</FONT><BR><FONT size=3D4><B>Subject:</B> RE: Frameformat in a = l2cs=20 gmpls rnvironment.</FONT><BR><BR>Hi,<BR><BR>I think Juergen has raised = an=20 important question. How frames are labeled (and the related data = plane=20 forwarding behavior) is not defined by the control plane. The = control=20 plane serves to provision the data plane, not define it. In the=20 framework draft it is not clear what data plane standards are covered = by the=20 stated control plane requirements. Some references should be=20 supplied. In any case, the labeling and forwarding behavior = should be=20 defined by these referenced standards, not by = GMPLS.<BR><BR>(I am=20 assuming definition of new data plane standards is out of scope for=20 CCAMP.)<BR><BR>Regards,<BR>Ben<BR> <HR align=3Dleft width=3D"99%" SIZE=3D4> <FONT color=3Dblack><BR><B>From:</B> owner-ccamp@ops.ietf.org=20 [mailto:owner-ccamp@ops.ietf.org] <B>On Behalf Of </B>Adrian=20 Farrel<BR><B>Sent:</B> Friday, July 22, 2005 8:37 AM<BR><B>To:</B> = P=E4r=20 Mattsson; Loa Andersson<BR><B>Cc:</B> = ccamp@ops.ietf.org<BR><B>Subject:</B>=20 Re: Frameformat in a l2cs gmpls rnvironment.<BR><BR>Hi,<BR><BR>It is = always=20 refreshing how engineers jump straight to a discussion of the solution = :-)<BR><BR>Perhaps we can assume that the framework draft is a good=20 representation of the problem space, and that we are ready to start = discussing=20 the solutions.<BR><BR>One (perhaps the only?) significant question to = answer=20 is how the frames will be labelled. This question is one we must come = to as=20 soon as we are confident that the requirements need to be addressed at = all.<BR><BR>As has been pointed out, there are several possibilities = and to=20 pick from these we need to understand:<BR>1. Do we need to support = explicit=20 label stacking?<BR> Note that this is not supported in = TDM, LSC or=20 FSC.<BR>2. Do we need to be able to control (perhaps through=20 an<BR> external signaling controller) existing = hardware and=20 <BR> install LSPs through existing networks?<BR>3. Do we = need to=20 support existing function simultaneous to<BR> the support = of GMPLS=20 L2 LSPs?<BR><BR>I think that from a chair's perspective I can give = some=20 limited guidance.<BR><BR>A. These questions must be raised and = answered in the=20 framework I-D<BR>B. The answer to question 3 is "yes". This means that = the use=20 of labels<BR> must not significantly deplete any namespace = used to=20 support other<BR> function.<BR>C. CCAMP is chartered to = look at=20 the control of transport networks. This<BR> includes = Metro,=20 but I am unsure about Campus. It does not cover <BR> =20 signaling to the desktop.<BR><BR>Thanks,<BR>Adrian<BR><BR>----- = Original=20 Message ----- <BR>From: "P=E4r Mattsson" <</FONT><FONT = color=3Dblue><U><A=20 href=3D"mailto:per@defero.se">per@defero.se</A></U></FONT><FONT=20 color=3Dblack>><BR>To: "Loa Andersson" <</FONT><FONT = color=3Dblue><U><A=20 href=3D"mailto:loa@pi.se">loa@pi.se</A></U></FONT><FONT = color=3Dblack>><BR>Cc:=20 <</FONT><FONT color=3Dblue><U><A=20 = href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A></U></FONT><FONT= =20 color=3Dblack>><BR>Sent: Friday, July 22, 2005 11:41 AM<BR>Subject: = Re:=20 Frameformat in a l2cs gmpls rnvironment.<BR><BR><BR>> > Per and=20 Dimitri,<BR>> ><BR>> > I would like to come down stronger = than=20 that, for me it is<BR>> > a very strong requirement that the = same switch=20 can handle<BR>> > both VLANs and GMPLs trafic correctly. I can't = dsee=20 how that<BR>> > could be done if using the VLAN tpid to indicate = GMPLS<BR>> > traffic.<BR>> <BR>> If you ever want that = same switch=20 to handle traffic for a directly<BR>> connected host (not to=20 uncommen) you would want that to use normal<BR>> ethernet = macaddress=20 switching. So of course you do not want to have to<BR>> choose = between vlan=20 and gmpls, you would want both at the same time.<BR>> <BR>> = /per<BR>>=20 <BR>> <BR>> ><BR>> > /Loa<BR>> ><BR>> > = P=E4r=20 Mattsson wrote:<BR>> >>>hi par, one of the possibilities = that has=20 been considered to cope with<BR>> >>>this requirement is = to use a=20 dedicated TPID for the Ethernet labeled<BR>> >>>frames; = this would=20 allow differentiated processing with non-labeled<BR>>=20 >>>framesthanks.<BR>> >><BR>> >><BR>> = >>=20 That seems to make more sence. If that frame is to be sized like = a<BR>>=20 >> 802.1q<BR>> >> frame. There is not that much space = left to a=20 label. Or is the demand to<BR>> >> use jumboframes ?<BR>> = >>=20 Has there been any discussion on labelstacking, and mainly where = to<BR>>=20 >> place<BR>> >> the information?<BR>> = >><BR>>=20 >> Regards.<BR>> >> Per<BR>> >><BR>>=20 >><BR>> >><BR>> ><BR>> ><BR>> > = --<BR>>=20 > Loa Andersson<BR>> ><BR>> > Principal Networking=20 Architect<BR>> > Acreo=20 = AB  = ; = =20 phone: +46 8 632 77 14<BR>> > Isafjordsgatan=20 = 22  = ; =20 mobile: +46 739 81 21 64<BR>> > Kista,=20 = Sweden &= nbsp; =20 email: </FONT><FONT color=3Dblue><U><A=20 = href=3D"mailto:loa.andersson@acreo.se">loa.andersson@acreo.se</A></U></FO= NT><FONT=20 color=3Dblack><BR>>=20 = > &nb= sp; &nbs= p;  = ; =20 </FONT><FONT color=3Dblue><U><A=20 href=3D"mailto:loa@pi.se">loa@pi.se</A></U></FONT><FONT = color=3Dblack><BR>>=20 ><BR>> <BR>> <BR>> <BR>> <BR><BR></FONT><FONT=20 face=3DMonospace,Courier=20 = size=3D4>=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D</FONT><BR><FONT=20 face=3DMonospace,Courier size=3D4>The information contained in this = message may be=20 privileged</FONT><BR><FONT face=3DMonospace,Courier size=3D4>and = confidential and=20 protected from disclosure. If the reader</FONT><BR><FONT=20 face=3DMonospace,Courier size=3D4>of this message is not the intended = recipient,=20 or an employee</FONT><BR><FONT face=3DMonospace,Courier size=3D4>or = agent=20 responsible for delivering this message to the</FONT><BR><FONT=20 face=3DMonospace,Courier size=3D4>intended recipient, you are hereby = notified that=20 any reproduction,</FONT><BR><FONT face=3DMonospace,Courier = size=3D4>dissemination=20 or distribution of this communication is strictly</FONT><BR><FONT=20 face=3DMonospace,Courier size=3D4>prohibited. If you have received = this=20 communication in error,</FONT><BR><FONT face=3DMonospace,Courier = size=3D4>please=20 notify us immediately by replying to the message and</FONT><BR><FONT=20 face=3DMonospace,Courier size=3D4>deleting it from your computer. = Thank you.=20 Tellabs</FONT><BR><FONT face=3DMonospace,Courier=20 = size=3D4>=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D</FONT><BR> <P></P></BLOCKQUOTE></BODY></HTML> ------=_NextPart_000_017C_01C58EC0.0C02EAF0-- Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 16:23:38 +0000 To: "Igor Bryskin" <ibryskin@movaz.com> Cc: "Mack-Crane, T. Benjamin" <Ben.Mack-Crane@tellabs.com>, "Adrian Farrel" <adrian@olddog.co.uk>, =?iso-8859-1?Q?P=E4r_Mattsson?= <per@defero.se>, "Loa Andersson" <loa@pi.se>, <ccamp@ops.ietf.org> MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: Re: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 18:21:14 +0200 Message-ID: <OFF1031856.32FE1CC9-ONC1257046.0059D553-C1257046.0059D5F2@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmlnb3IgLSBpIGFtIG5vdCBzdXJlIHdo aWNoIHBvaW50IHRoZXkgZXhhY3RseSBoYXZlIC0gdGhlIHB1cnBvc2Ugb2YgdGhpcyBkb2N1bWVu dCBpcyB0byBkZXRhaWwgaG93IGZvcndhcmRpbmcgaW5mb3JtYXRpb24gY2FuIGJlIGV4Y2hhbmdl ZCB2aWEgdGhlIGNvbnRyb2wgcGxhbmUgKGFuZCB0aGVuIGluc3RhbGxlZCkgbm90IHRvIGRlZmlu ZSBhIHNwZWNpZmljIGZvcndhcmRpbmcgYmVoYXZpb3VyPC9GT05UPjwvUD48UD48Rk9OVCBGQUNF PSJNb25vc3BhY2UsQ291cmllciI+dGhhbmtzLDwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9u b3NwYWNlLENvdXJpZXIiPi0gZGltaXRyaS48L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTI+PEI+ JnF1b3Q7SWdvciBCcnlza2luJnF1b3Q7ICZsdDtpYnJ5c2tpbkBtb3Zhei5jb20mZ3Q7PC9CPjwv Rk9OVD48QlI+PEZPTlQgU0laRT0yPlNlbnQgYnk6IG93bmVyLWNjYW1wQG9wcy5pZXRmLm9yZzwv Rk9OVD48QlI+PEZPTlQgU0laRT0yPjA3LzIyLzIwMDUgMTE6MTAgQVNUPC9GT05UPjxCUj48QlI+ IDxGT05UIFNJWkU9Mj5Ubzo8L0ZPTlQ+IDxGT05UIFNJWkU9Mj4mcXVvdDtNYWNrLUNyYW5lLCBU LiBCZW5qYW1pbiZxdW90OyAmbHQ7QmVuLk1hY2stQ3JhbmVAdGVsbGFicy5jb20mZ3Q7LCAmcXVv dDtBZHJpYW4gRmFycmVsJnF1b3Q7ICZsdDthZHJpYW5Ab2xkZG9nLmNvLnVrJmd0OywgUCZhdW1s O3IgTWF0dHNzb24gJmx0O3BlckBkZWZlcm8uc2UmZ3Q7LCAmcXVvdDtMb2EgQW5kZXJzc29uJnF1 b3Q7ICZsdDtsb2FAcGkuc2UmZ3Q7PC9GT05UPjxCUj4gPEZPTlQgU0laRT0yPmNjOjwvRk9OVD4g PEZPTlQgU0laRT0yPiZsdDtjY2FtcEBvcHMuaWV0Zi5vcmcmZ3Q7PC9GT05UPjxCUj4gPEZPTlQg U0laRT0yPmJjYzo8L0ZPTlQ+IDxCUj4gPEZPTlQgU0laRT0yPlN1YmplY3Q6PC9GT05UPiA8Rk9O VCBTSVpFPTI+UmU6IEZyYW1lZm9ybWF0IGluIGEgbDJjcyBnbXBscyBybnZpcm9ubWVudC48L0ZP TlQ+PEJSPiA8QlI+PEJSPjwvUD48UD5IaSw8QlI+PEJSPkkgdGhpbmsgQmVuIGFuZCBKdWVyZ2Vu IGhhdmUgYSBwb2ludCBoZXJlLiBUaGVyZSBpcyBub3RoaW5nIHRoYXQgY291bGQgYmUgZHluYW1p Y2FsbHkgcHJvdmlzaW9uZWQgdGhhdCBjb3VsZCBub3QgYmUgYWxzbyBwcm92aXNpb25lZCBieSBt YW5hZ2VtZW50LiBJbiBvdGhlciB3b3JkcyBkYXRhIHBsYW5lIGlzIHRoZSBLaW5nIGhlcmUuIFdl IG5lZWQgZGF0YSBwbGFuZSBzdGFuZGFyZChzKSBvbiBob3cgd2UgZW5jb2RlIGEgbGFiZWwsIHdo ZXRoZXIgd2UgbmVlZCBhIGxhYmVsIHN0YWNrIG9yIG5vdCwgaG93IHRoZSBsYWJlbGVkIHRyYWZm aWMgaXMgc3VwcG9zZWQgdG8gYmUmbmJzcDt0cmVhdGVkLCBob3cgbGFiZWxlZCB0cmFmZmljIGNv LWV4aXN0IHdpdGggbm9uLWxhYmVsIHRyYWZmaWMsIGV0Yy4gVGhpcyBpcyBzb21ldGhpbmcgdGhh dCBub3QgZm9yIENDQU1QIHRvIGRlZmluZS4gVGFrZSBmb3IgZXhhbXBsZSBURE0gbmV0d29ya3Mu IEdNUExTIG9ubHkgcHJvdmlkZXMgYSB3YXkgdG8gZHluYW1pY2FsbHkgcHJvdmlzaW9uIEcuNzA3 IG5ldHdvcmtzLiBIZW5jZSB0aGVyZSBpcyBhIG5lZWQgaW4gcGFyYWxsZWwgc3RhbmRhcmRpemF0 aW9uIGFjdGl2aXRpZXMgaW4gSVRVLiBPbmx5IGFmdGVyIHRoYXQgd2UgY2FuIGRpc2N1c3MgaG93 IGFsbCB0aGF0IGNvdWxkIGJlIGR5bmFtaWNhbGx5IHByb3Zpc2lvbmVkLCB0aGF0IGlzIHRoZSBh c3BlY3RzIG9mIGNvbnRyb2wgcGxhbmUuIDxCUj48QlI+SWdvcjxCUj48Rk9OVCBTSVpFPTQ+LS0t LS0gT3JpZ2luYWwgTWVzc2FnZSAtLS0tLSA8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND48Qj5Gcm9t OjwvQj4gPC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOkJl bi5NYWNrLUNyYW5lQHRlbGxhYnMuY29tPk1hY2stQ3JhbmUsIFQuIEJlbmphbWluPC9BPjwvVT48 L0ZPTlQ+PEZPTlQgU0laRT00IENPTE9SPUJMQUNLPiA8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND48 Qj5Ubzo8L0I+IDwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0 bzphZHJpYW5Ab2xkZG9nLmNvLnVrPkFkcmlhbiBGYXJyZWw8L0E+PC9VPjwvRk9OVD48Rk9OVCBT SVpFPTQgQ09MT1I9QkxBQ0s+IDsgPC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1CTFVFPjxVPjxB IEhSRUY9bWFpbHRvOnBlckBkZWZlcm8uc2U+UCZhdW1sO3IgTWF0dHNzb248L0E+PC9VPjwvRk9O VD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxBQ0s+IDsgPC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1C TFVFPjxVPjxBIEhSRUY9bWFpbHRvOmxvYUBwaS5zZT5Mb2EgQW5kZXJzc29uPC9BPjwvVT48L0ZP TlQ+PEZPTlQgU0laRT00IENPTE9SPUJMQUNLPiA8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND48Qj5D Yzo8L0I+IDwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0bzpj Y2FtcEBvcHMuaWV0Zi5vcmc+Y2NhbXBAb3BzLmlldGYub3JnPC9BPjwvVT48L0ZPTlQ+PEZPTlQg U0laRT00IENPTE9SPUJMQUNLPiA8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND48Qj5TZW50OjwvQj4g RnJpZGF5LCBKdWx5IDIyLCAyMDA1IDEwOjIzIEFNPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+PEI+ U3ViamVjdDo8L0I+IFJFOiBGcmFtZWZvcm1hdCBpbiBhIGwyY3MgZ21wbHMgcm52aXJvbm1lbnQu PC9GT05UPjxCUj48QlI+SGksPEJSPjxCUj5JIHRoaW5rIEp1ZXJnZW4gaGFzIHJhaXNlZCBhbiBp bXBvcnRhbnQgcXVlc3Rpb24uJm5ic3A7IEhvdyBmcmFtZXMgYXJlIGxhYmVsZWQgKGFuZCB0aGUg cmVsYXRlZCBkYXRhIHBsYW5lIGZvcndhcmRpbmcgYmVoYXZpb3IpIGlzIG5vdCBkZWZpbmVkIGJ5 IHRoZSBjb250cm9sIHBsYW5lLiZuYnNwOyBUaGUgY29udHJvbCBwbGFuZSBzZXJ2ZXMgdG8gcHJv dmlzaW9uIHRoZSBkYXRhIHBsYW5lLCBub3QgZGVmaW5lIGl0LiZuYnNwOyBJbiB0aGUgZnJhbWV3 b3JrIGRyYWZ0IGl0IGlzIG5vdCBjbGVhciB3aGF0IGRhdGEgcGxhbmUgc3RhbmRhcmRzIGFyZSBj b3ZlcmVkIGJ5IHRoZSBzdGF0ZWQgY29udHJvbCBwbGFuZSByZXF1aXJlbWVudHMuJm5ic3A7IFNv bWUgcmVmZXJlbmNlcyBzaG91bGQgYmUgc3VwcGxpZWQuJm5ic3A7IEluIGFueSBjYXNlLCB0aGUg bGFiZWxpbmcgYW5kIGZvcndhcmRpbmcgYmVoYXZpb3Igc2hvdWxkIGJlIGRlZmluZWQmbmJzcDti eSB0aGVzZSByZWZlcmVuY2VkIHN0YW5kYXJkcywgbm90Jm5ic3A7YnkgR01QTFMuPEJSPjxCUj4o SSBhbSBhc3N1bWluZyBkZWZpbml0aW9uIG9mIG5ldyBkYXRhIHBsYW5lIHN0YW5kYXJkcyBpcyBv dXQgb2Ygc2NvcGUgZm9yIENDQU1QLik8QlI+PEJSPlJlZ2FyZHMsPEJSPkJlbjxCUj48SFIgV0lE VEg9Ijk5JSIgU0laRT0iNCIgQUxJR049ImxlZnQiPjxGT05UIENPTE9SPUJMQUNLPjxCUj48Qj5G cm9tOjwvQj4gb3duZXItY2NhbXBAb3BzLmlldGYub3JnIFttYWlsdG86b3duZXItY2NhbXBAb3Bz LmlldGYub3JnXSA8Qj5PbiBCZWhhbGYgT2YgPC9CPkFkcmlhbiBGYXJyZWw8QlI+PEI+U2VudDo8 L0I+IEZyaWRheSwgSnVseSAyMiwgMjAwNSA4OjM3IEFNPEJSPjxCPlRvOjwvQj4gUCZhdW1sO3Ig TWF0dHNzb247IExvYSBBbmRlcnNzb248QlI+PEI+Q2M6PC9CPiBjY2FtcEBvcHMuaWV0Zi5vcmc8 QlI+PEI+U3ViamVjdDo8L0I+IFJlOiBGcmFtZWZvcm1hdCBpbiBhIGwyY3MgZ21wbHMgcm52aXJv bm1lbnQuPEJSPjxCUj5IaSw8QlI+PEJSPkl0IGlzIGFsd2F5cyByZWZyZXNoaW5nIGhvdyBlbmdp bmVlcnMganVtcCBzdHJhaWdodCB0byBhIGRpc2N1c3Npb24gb2YgdGhlIHNvbHV0aW9uIDotKTxC Uj48QlI+UGVyaGFwcyB3ZSBjYW4gYXNzdW1lIHRoYXQgdGhlIGZyYW1ld29yayBkcmFmdCBpcyBh IGdvb2QgcmVwcmVzZW50YXRpb24gb2YgdGhlIHByb2JsZW0gc3BhY2UsIGFuZCB0aGF0IHdlIGFy ZSByZWFkeSB0byBzdGFydCBkaXNjdXNzaW5nIHRoZSBzb2x1dGlvbnMuPEJSPjxCUj5PbmUgKHBl cmhhcHMgdGhlIG9ubHk/KSBzaWduaWZpY2FudCBxdWVzdGlvbiB0byBhbnN3ZXIgaXMgaG93IHRo ZSBmcmFtZXMgd2lsbCBiZSBsYWJlbGxlZC4gVGhpcyBxdWVzdGlvbiBpcyBvbmUgd2UgbXVzdCBj b21lIHRvIGFzIHNvb24gYXMgd2UgYXJlIGNvbmZpZGVudCB0aGF0IHRoZSByZXF1aXJlbWVudHMg bmVlZCB0byBiZSBhZGRyZXNzZWQgYXQgYWxsLjxCUj48QlI+QXMgaGFzIGJlZW4gcG9pbnRlZCBv dXQsIHRoZXJlIGFyZSBzZXZlcmFsIHBvc3NpYmlsaXRpZXMgYW5kIHRvIHBpY2sgZnJvbSB0aGVz ZSB3ZSBuZWVkIHRvIHVuZGVyc3RhbmQ6PEJSPjEuIERvIHdlIG5lZWQgdG8gc3VwcG9ydCBleHBs aWNpdCBsYWJlbCBzdGFja2luZz88QlI+Jm5ic3A7Jm5ic3A7IE5vdGUgdGhhdCB0aGlzIGlzIG5v dCBzdXBwb3J0ZWQgaW4gVERNLCBMU0Mgb3IgRlNDLjxCUj4yLiBEbyB3ZSBuZWVkIHRvIGJlIGFi bGUgdG8gY29udHJvbCAocGVyaGFwcyB0aHJvdWdoIGFuPEJSPiZuYnNwOyZuYnNwOyZuYnNwO2V4 dGVybmFsIHNpZ25hbGluZyBjb250cm9sbGVyKSBleGlzdGluZyBoYXJkd2FyZSBhbmQgPEJSPiZu YnNwOyZuYnNwOyBpbnN0YWxsIExTUHMgdGhyb3VnaCBleGlzdGluZyBuZXR3b3Jrcz88QlI+My4g RG8gd2UgbmVlZCB0byBzdXBwb3J0IGV4aXN0aW5nIGZ1bmN0aW9uIHNpbXVsdGFuZW91cyB0bzxC Uj4mbmJzcDsmbmJzcDsgdGhlIHN1cHBvcnQgb2YgR01QTFMgTDIgTFNQcz88QlI+PEJSPkkgdGhp bmsgdGhhdCBmcm9tIGEgY2hhaXIncyBwZXJzcGVjdGl2ZSBJIGNhbiBnaXZlIHNvbWUgbGltaXRl ZCBndWlkYW5jZS48QlI+PEJSPkEuIFRoZXNlIHF1ZXN0aW9ucyBtdXN0IGJlIHJhaXNlZCBhbmQg YW5zd2VyZWQgaW4gdGhlIGZyYW1ld29yayBJLUQ8QlI+Qi4gVGhlIGFuc3dlciB0byBxdWVzdGlv biAzIGlzICZxdW90O3llcyZxdW90Oy4gVGhpcyBtZWFucyB0aGF0IHRoZSB1c2Ugb2YgbGFiZWxz PEJSPiZuYnNwOyZuYnNwOyBtdXN0IG5vdCBzaWduaWZpY2FudGx5IGRlcGxldGUgYW55IG5hbWVz cGFjZSB1c2VkIHRvIHN1cHBvcnQgb3RoZXI8QlI+Jm5ic3A7Jm5ic3A7IGZ1bmN0aW9uLjxCUj5D LiBDQ0FNUCBpcyBjaGFydGVyZWQgdG8gbG9vayBhdCB0aGUgY29udHJvbCBvZiB0cmFuc3BvcnQg bmV0d29ya3MuIFRoaXM8QlI+Jm5ic3A7Jm5ic3A7Jm5ic3A7aW5jbHVkZXMgTWV0cm8sIGJ1dCBJ IGFtIHVuc3VyZSBhYm91dCBDYW1wdXMuIEl0IGRvZXMgbm90Jm5ic3A7Y292ZXIgPEJSPiZuYnNw OyZuYnNwOyBzaWduYWxpbmcgdG8gdGhlIGRlc2t0b3AuPEJSPjxCUj5UaGFua3MsPEJSPkFkcmlh bjxCUj48QlI+LS0tLS0gT3JpZ2luYWwgTWVzc2FnZSAtLS0tLSA8QlI+RnJvbTogJnF1b3Q7UCZh dW1sO3IgTWF0dHNzb24mcXVvdDsgJmx0OzwvRk9OVD48Rk9OVCBDT0xPUj1CTFVFPjxVPjxBIEhS RUY9bWFpbHRvOnBlckBkZWZlcm8uc2U+cGVyQGRlZmVyby5zZTwvQT48L1U+PC9GT05UPjxGT05U IENPTE9SPUJMQUNLPiZndDs8QlI+VG86ICZxdW90O0xvYSBBbmRlcnNzb24mcXVvdDsgJmx0Ozwv Rk9OVD48Rk9OVCBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOmxvYUBwaS5zZT5sb2FAcGku c2U8L0E+PC9VPjwvRk9OVD48Rk9OVCBDT0xPUj1CTEFDSz4mZ3Q7PEJSPkNjOiAmbHQ7PC9GT05U PjxGT05UIENPTE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86Y2NhbXBAb3BzLmlldGYub3JnPmNj YW1wQG9wcy5pZXRmLm9yZzwvQT48L1U+PC9GT05UPjxGT05UIENPTE9SPUJMQUNLPiZndDs8QlI+ U2VudDogRnJpZGF5LCBKdWx5IDIyLCAyMDA1IDExOjQxIEFNPEJSPlN1YmplY3Q6IFJlOiBGcmFt ZWZvcm1hdCBpbiBhIGwyY3MgZ21wbHMgcm52aXJvbm1lbnQuPEJSPjxCUj48QlI+Jmd0OyAmZ3Q7 IFBlciBhbmQgRGltaXRyaSw8QlI+Jmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyBJIHdvdWxkIGxpa2Ug dG8gY29tZSBkb3duIHN0cm9uZ2VyIHRoYW4gdGhhdCwgZm9yIG1lIGl0IGlzPEJSPiZndDsgJmd0 OyBhIHZlcnkgc3Ryb25nIHJlcXVpcmVtZW50IHRoYXQgdGhlIHNhbWUgc3dpdGNoIGNhbiBoYW5k bGU8QlI+Jmd0OyAmZ3Q7IGJvdGggVkxBTnMgYW5kIEdNUExzIHRyYWZpYyBjb3JyZWN0bHkuIEkg Y2FuJ3QgZHNlZSBob3cgdGhhdDxCUj4mZ3Q7ICZndDsgY291bGQgYmUgZG9uZSBpZiB1c2luZyB0 aGUgVkxBTiB0cGlkIHRvIGluZGljYXRlIEdNUExTPEJSPiZndDsgJmd0OyB0cmFmZmljLjxCUj4m Z3Q7IDxCUj4mZ3Q7IElmIHlvdSBldmVyIHdhbnQgdGhhdCBzYW1lIHN3aXRjaCB0byBoYW5kbGUg dHJhZmZpYyBmb3IgYSBkaXJlY3RseTxCUj4mZ3Q7IGNvbm5lY3RlZCZuYnNwOyBob3N0IChub3Qg dG8gdW5jb21tZW4pIHlvdSB3b3VsZCB3YW50IHRoYXQgdG8gdXNlIG5vcm1hbDxCUj4mZ3Q7IGV0 aGVybmV0IG1hY2FkZHJlc3Mgc3dpdGNoaW5nLiBTbyBvZiBjb3Vyc2UgeW91IGRvIG5vdCB3YW50 IHRvIGhhdmUgdG88QlI+Jmd0OyBjaG9vc2UgYmV0d2VlbiB2bGFuIGFuZCBnbXBscywgeW91IHdv dWxkIHdhbnQgYm90aCBhdCB0aGUgc2FtZSB0aW1lLjxCUj4mZ3Q7IDxCUj4mZ3Q7IC9wZXI8QlI+ Jmd0OyA8QlI+Jmd0OyA8QlI+Jmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyAvTG9hPEJSPiZndDsgJmd0 OzxCUj4mZ3Q7ICZndDsgUCZhdW1sO3IgTWF0dHNzb24gd3JvdGU6PEJSPiZndDsgJmd0OyZndDsm Z3Q7aGkgcGFyLCBvbmUgb2YgdGhlIHBvc3NpYmlsaXRpZXMgdGhhdCBoYXMgYmVlbiBjb25zaWRl cmVkIHRvIGNvcGUgd2l0aDxCUj4mZ3Q7ICZndDsmZ3Q7Jmd0O3RoaXMgcmVxdWlyZW1lbnQgaXMg dG8gdXNlIGEgZGVkaWNhdGVkIFRQSUQgZm9yIHRoZSBFdGhlcm5ldCBsYWJlbGVkPEJSPiZndDsg Jmd0OyZndDsmZ3Q7ZnJhbWVzOyB0aGlzIHdvdWxkIGFsbG93IGRpZmZlcmVudGlhdGVkIHByb2Nl c3Npbmcgd2l0aCBub24tbGFiZWxlZDxCUj4mZ3Q7ICZndDsmZ3Q7Jmd0O2ZyYW1lc3RoYW5rcy48 QlI+Jmd0OyAmZ3Q7Jmd0OzxCUj4mZ3Q7ICZndDsmZ3Q7PEJSPiZndDsgJmd0OyZndDsgVGhhdCBz ZWVtcyB0byBtYWtlIG1vcmUgc2VuY2UuIElmIHRoYXQgZnJhbWUgaXMgdG8gYmUgc2l6ZWQgbGlr ZSBhPEJSPiZndDsgJmd0OyZndDsgODAyLjFxPEJSPiZndDsgJmd0OyZndDsgZnJhbWUuIFRoZXJl IGlzIG5vdCB0aGF0IG11Y2ggc3BhY2UgbGVmdCB0byBhIGxhYmVsLiBPciBpcyB0aGUgZGVtYW5k IHRvPEJSPiZndDsgJmd0OyZndDsgdXNlIGp1bWJvZnJhbWVzID88QlI+Jmd0OyAmZ3Q7Jmd0OyBI YXMgdGhlcmUgYmVlbiBhbnkgZGlzY3Vzc2lvbiBvbiBsYWJlbHN0YWNraW5nLCBhbmQgbWFpbmx5 IHdoZXJlIHRvPEJSPiZndDsgJmd0OyZndDsgcGxhY2U8QlI+Jmd0OyAmZ3Q7Jmd0OyB0aGUgaW5m b3JtYXRpb24/PEJSPiZndDsgJmd0OyZndDs8QlI+Jmd0OyAmZ3Q7Jmd0OyBSZWdhcmRzLjxCUj4m Z3Q7ICZndDsmZ3Q7IFBlcjxCUj4mZ3Q7ICZndDsmZ3Q7PEJSPiZndDsgJmd0OyZndDs8QlI+Jmd0 OyAmZ3Q7Jmd0OzxCUj4mZ3Q7ICZndDs8QlI+Jmd0OyAmZ3Q7PEJSPiZndDsgJmd0OyAtLTxCUj4m Z3Q7ICZndDsgTG9hIEFuZGVyc3NvbjxCUj4mZ3Q7ICZndDs8QlI+Jmd0OyAmZ3Q7IFByaW5jaXBh bCBOZXR3b3JraW5nIEFyY2hpdGVjdDxCUj4mZ3Q7ICZndDsgQWNyZW8gQUImbmJzcDsmbmJzcDsm bmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJz cDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsm bmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsgcGhvbmU6Jm5ic3A7ICs0NiA4IDYzMiA3NyAx NDxCUj4mZ3Q7ICZndDsgSXNhZmpvcmRzZ2F0YW4gMjImbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsm bmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJzcDsmbmJz cDsmbmJzcDsmbmJzcDsmbmJzcDsgbW9iaWxlOiArNDYgNzM5IDgxIDIxIDY0PEJSPiZndDsgJmd0 OyBLaXN0YSwgU3dlZGVuJm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7 Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5i c3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7IGVtYWlsOiZuYnNwOyA8L0ZPTlQ+PEZPTlQgQ09M T1I9QkxVRT48VT48QSBIUkVGPW1haWx0bzpsb2EuYW5kZXJzc29uQGFjcmVvLnNlPmxvYS5hbmRl cnNzb25AYWNyZW8uc2U8L0E+PC9VPjwvRk9OVD48Rk9OVCBDT0xPUj1CTEFDSz48QlI+Jmd0OyAm Z3Q7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7 Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5i c3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7 Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5i c3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7Jm5ic3A7IDwvRk9OVD48Rk9OVCBDT0xP Uj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOmxvYUBwaS5zZT5sb2FAcGkuc2U8L0E+PC9VPjwvRk9O VD48Rk9OVCBDT0xPUj1CTEFDSz48QlI+Jmd0OyAmZ3Q7PEJSPiZndDsgPEJSPiZndDsgPEJSPiZn dDsgPEJSPiZndDsgPEJSPjxCUj48L0ZPTlQ+PEZPTlQgU0laRT00IEZBQ0U9Ik1vbm9zcGFjZSxD b3VyaWVyIj49PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT08L0ZPTlQ+PEJSPjxGT05UIFNJWkU9NCBGQUNFPSJNb25vc3BhY2UsQ291cmll ciI+VGhlIGluZm9ybWF0aW9uIGNvbnRhaW5lZCBpbiB0aGlzIG1lc3NhZ2UgbWF5IGJlIHByaXZp bGVnZWQ8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9NCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+YW5k IGNvbmZpZGVudGlhbCBhbmQgcHJvdGVjdGVkIGZyb20gZGlzY2xvc3VyZS4gSWYgdGhlIHJlYWRl cjwvRk9OVD48QlI+PEZPTlQgU0laRT00IEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5vZiB0aGlz IG1lc3NhZ2UgaXMgbm90IHRoZSBpbnRlbmRlZCByZWNpcGllbnQsIG9yIGFuIGVtcGxveWVlPC9G T05UPjxCUj48Rk9OVCBTSVpFPTQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPm9yIGFnZW50IHJl c3BvbnNpYmxlIGZvciBkZWxpdmVyaW5nIHRoaXMgbWVzc2FnZSB0byB0aGU8L0ZPTlQ+PEJSPjxG T05UIFNJWkU9NCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+aW50ZW5kZWQgcmVjaXBpZW50LCB5 b3UgYXJlIGhlcmVieSBub3RpZmllZCB0aGF0IGFueSByZXByb2R1Y3Rpb24sPC9GT05UPjxCUj48 Rk9OVCBTSVpFPTQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmRpc3NlbWluYXRpb24gb3IgZGlz dHJpYnV0aW9uIG9mIHRoaXMgY29tbXVuaWNhdGlvbiBpcyBzdHJpY3RseTwvRk9OVD48QlI+PEZP TlQgU0laRT00IEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5wcm9oaWJpdGVkLiBJZiB5b3UgaGF2 ZSByZWNlaXZlZCB0aGlzIGNvbW11bmljYXRpb24gaW4gZXJyb3IsPC9GT05UPjxCUj48Rk9OVCBT SVpFPTQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPnBsZWFzZSBub3RpZnkgdXMgaW1tZWRpYXRl bHkgYnkgcmVwbHlpbmcgdG8gdGhlIG1lc3NhZ2UgYW5kPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQg RkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmRlbGV0aW5nIGl0IGZyb20geW91ciBjb21wdXRlci4g VGhhbmsgeW91LiBUZWxsYWJzPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQgRkFDRT0iTW9ub3NwYWNl LENvdXJpZXIiPj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PTwvRk9OVD48QlI+PC9QPg== Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 15:52:16 +0000 To: "Adrian Farrel" <adrian@olddog.co.uk> Cc: <Dimitri.Papadimitriou@alcatel.be>, <ccamp@ops.ietf.org> MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: Re: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 17:51:08 +0200 Message-ID: <OF4CF8DFD4.FA2E16FA-ONC1257046.005713CB-C1257046.00571440@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmFkcmlhbiAtIG9rIG11Y2ggdGhhbmtz IGZvciB0aGUgY2xhcmlmaWNhdGlvbiAtIDwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3Nw YWNlLENvdXJpZXIiPmkgd2FudGVkIHRvIGJlIHN1cmUgdGhhdCB0aGUgd29ya2luZyBncm91cCBj b25zZW5zdXMgaXMgZ29pbmcgdG8gYmUgdGhlIGVsZW1lbnQgb2YgZGVjaXNpb24gZm9yIHB1dHRp bmcgdGhpcyBpdGVtIGluIHRoZSB3b3JraW5nIGxpc3QgLSBhbmQgd2Ugd29uJ3QgaGF2ZSBvdGhl ciBpbnRlcmZlcmluZyBlbGVtZW50cyBpbiB0aGUgZGVjaXNpb24gcHJvY2VzcyAtIDwvRk9OVD48 QlI+PEJSPjxGT05UIFNJWkU9Mj48Qj4mcXVvdDtBZHJpYW4gRmFycmVsJnF1b3Q7ICZsdDthZHJp YW5Ab2xkZG9nLmNvLnVrJmd0OzwvQj48L0ZPTlQ+PEJSPjxGT05UIFNJWkU9Mj5TZW50IGJ5OiBv d25lci1jY2FtcEBvcHMuaWV0Zi5vcmc8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9Mj4wNy8yMi8yMDA1 IDE2OjA4PC9GT05UPjxCUj48Rk9OVCBTSVpFPTI+UGxlYXNlIHJlc3BvbmQgdG8gJnF1b3Q7QWRy aWFuIEZhcnJlbCZxdW90OzwvRk9OVD48QlI+PEJSPiA8Rk9OVCBTSVpFPTI+VG86PC9GT05UPiA8 Rk9OVCBTSVpFPTI+RGltaXRyaSBQQVBBRElNSVRSSU9VL0JFL0FMQ0FURUxAQUxDQVRFTDwvRk9O VD48QlI+IDxGT05UIFNJWkU9Mj5jYzo8L0ZPTlQ+IDxGT05UIFNJWkU9Mj4mbHQ7Y2NhbXBAb3Bz LmlldGYub3JnJmd0OzwvRk9OVD48QlI+IDxGT05UIFNJWkU9Mj5iY2M6PC9GT05UPiA8QlI+IDxG T05UIFNJWkU9Mj5TdWJqZWN0OjwvRk9OVD4gPEZPTlQgU0laRT0yPlJlOiBGcmFtZWZvcm1hdCBp biBhIGwyY3MgZ21wbHMgcm52aXJvbm1lbnQuPC9GT05UPjxCUj4gPEJSPjxCUj48L1A+PFA+PEZP TlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPkRpbWl0cmksPEJSPjwvRk9OVD48QlI+PEZPTlQg RkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPiZndDsgd2hhdCBkbyB5b3UgRVhBQ1RMWSBtZWFuIHdp dGggdGhlIGZvbGxvd2luZyBzZW50ZW5jZTxCUj4mZ3Q7PEJSPiZndDsgJnF1b3Q7VGhpcyBxdWVz dGlvbiBpcyBvbmUgd2UgbXVzdCBjb21lIHRvIGFzIHNvb24gYXMgd2UgYXJlIGNvbmZpZGVudCB0 aGF0PEJSPnRoZTxCUj4mZ3Q7IHJlcXVpcmVtZW50cyBuZWVkIHRvIGJlIGFkZHJlc3NlZCBhdCBh bGwuJnF1b3Q7PEJSPiZndDs8QlI+Jmd0OyBwbGVhc2UgY2xhcmlmeSAtIGJ0dyB3aG8gaXMgJnF1 b3Q7d2UmcXVvdDsgaW4gdGhpcyBjb250ZXh0ID88QlI+PC9GT05UPjxCUj48Rk9OVCBGQUNFPSJN b25vc3BhY2UsQ291cmllciI+JnF1b3Q7V2UmcXVvdDsgbWVhbnMgJnF1b3Q7dGhlIENDQU1QIHdv cmtpbmcgZ3JvdXAmcXVvdDs8QlI+PC9GT05UPjxCUj48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291 cmllciI+Jmd0OyB0aGlzIHNlbnRlbmNlIHJlYWRzIHRvIG1lIGFzICZxdW90O0NDQU1QIGhhcyBu b3QgeWV0IGNvbW1pdHRlZCB0byB3b3JrIG9uPEJSPnRoaXM8QlI+Jmd0OyB0b3BpYyZxdW90OyBp cyBteSByZWFkaW5nIGNvcnJlY3QgPyBvciBkbyB5b3UgbWVhbiBzb21ldGhpbmcgZWxzZSB3aXRo IHRoaXMgPzxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5Zb3Vy IHJlYWRpbmcgaXMgZXhhY3RseSBjb3JyZWN0IGFzIHN0YXRlZC48QlI+SSByZWZlciB5b3UgdG8g eW91ciBlbWFpbCB0byB0aGUgbGlzdCBvbiAxOXRoIEp1bHkuIFlvdSBhc2tlZC4uLjxCUj4mZ3Q7 IC0gaWYgeWVzLCBpcyB0aGUgaW50ZXJlc3QgaW4gR01QTFMgZm9yIEV0aGVybmV0IGluLS9vdXQt c2lkZSB0aGU8QlI+Jmd0OyAmbmJzcDsgQ0NBTVAgV0cgY29tbXVuaXR5IHN0cm9uZyBlbm91Z2gg dG8gY29uc2lkZXIgdGhlIGNvcnIuIEdNUExTPEJSPiZndDsgJm5ic3A7IHNpZ25hbGluZyBhbmQg cm91dGluZyByZXF1aXJlbWVudHMgKHNlY29uZCBwYXJ0IG9mIHRoZSBkb2N1bWVudCk8QlI+Jmd0 OyAmbmJzcDsgYXMgcGFydCBvZiB0aGUgQ0NBTVAgV0cgZWZmb3J0ID88QlI+PC9GT05UPjxCUj48 Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+VGh1cywgdGhlIHB1cnBvc2Ugb2YgdGhlIEkt RCB3YXMgdG8gY3J5c3RhbGl6ZSB0aGUgYXV0aG9ycycgdGhvdWdodHMgb248QlI+R01QTFMgY29u dHJvbCBvZiBFdGhlcm5ldCBzd2l0Y2hpbmcgYW5kIHRvIGluaXRpYXRlIGRpc2N1c3Npb24gb24g d2hldGhlcjxCUj50aGUgcHJvcG9zZWQgZGVwbG95bWVudCBtb2RlbHMgYXJlIHJlYWxpc3RpYyBh bmQgZGVzaXJhYmxlLiBGb2xsb3dpbmcgYTxCUj5wb3NpdGl2ZSBhbnN3ZXIgYW5kIGFzc3VtaW5n IHN1cHBvcnQgKGJ5IENDQU1QKSBmb3IgdGhlIGRvaW5nIHRoZSB3b3JrIChpbjxCUj5DQ0FNUCkg YWxsIHRoYXQgcmVtYWlucyBpcyB0byBhZGQgdGhlIHdvcmsgdG8gdGhlIGNoYXJ0ZXIgKHNlZSBw cmV2aW91czxCUj5jaGFydGVyIGRpc2N1c3Npb25zKSBhbmQgZ2V0IG9uIHdpdGggaXQuPEJSPjwv Rk9OVD48QlI+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPlRoYW5rcyw8QlI+QWRyaWFu PEJSPjwvRk9OVD48L1A+ Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 15:35:37 +0000 Content-Class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 16:33:59 +0100 Message-ID: <B5E87B043D4C514389141E2661D255EC0A835AE5@i2km41-ukdy.domain1.systemhost.net> Thread-Topic: Frameformat in a l2cs gmpls rnvironment. Thread-Index: AcWOv97Qnu365RxCQfqdrvWhC7nQtAAAIniQ From: <richard.spencer@bt.com> To: <juergen.heiles@siemens.com>, <loa@pi.se>, <per@defero.se> Cc: <ccamp@ops.ietf.org> Juergen, Before you can use GMPLS to set up VLANs a data plane must first exist = to carry the GMPLS control plane traffic. Yes you can use the standard = Ethernet data plane for this purpose (and in fact must do so for the = solution to be viable), but its not as simple as that. You cannot assume = that just because two GMPLS peers are directly connected that they can = communicate with each other. To exchange GMPLS control plane packets between a pair of GMPLS peers, = the peers will need to agree to use either untagged control packets or = to tag them with a common VLAN ID. The use of a separate VLAN for GMPLS = control traffic is analogous to using a separate VLAN for management = traffic today. If two switches belong to the same management VLAN then = you can use ICMP ping to verify connectivity between them. However, if = one switch has been put into the wrong management VLAN then the ping = will obviously fail. If you want to use a separate VLAN for GMPLS control plane traffic, then = you either need to (i) define a reserved GMPLS signalling VLAN ID, or = (ii) leave it up to the operator to choose a VLAN ID. If you leave it up = to the operator to decide what VLAN ID to use, then you need to consider = what the implications in the inter-operator case will be, i.e. what VLAN = ID range should be used for control plane traffic and what VLAN ID range = should be used for user plane traffic. I'm not saying that defining how GMPLS control plane packets are = forwarded in the Ethernet data plane will be an issue, simply that this = is something that must be addressed to develop interoperable solutions. Regards, Richard > -----Original Message----- > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On > Behalf Of Heiles Juergen > Sent: 22 July 2005 14:17 > To: Loa Andersson; P=E4r Mattsson > Cc: ccamp@ops.ietf.org > Subject: RE: Frameformat in a l2cs gmpls rnvironment. >=20 >=20 > GMPLS for SONET/SDH has no impact on the SONET/SDH data=20 > plane. It introduces a control plane for connection setup=20 > instead of a network management based connection setup. GMPLS=20 > for Ethernet can work in the same way. Ethernet as such is=20 > cl, but VLANs are setup and GMPLS can provide a control plane=20 > for VLAN setup. The Ethernet data plane is not changed and=20 > MAC address learning and forwarding is still be done within a=20 > VLAN. So no change to the Ethernet data plane. No GMPLS=20 > traffic is introduced, it is still VLAN traffic. > Introducing a dedicated GMPLS label for Ethernet (GMPLS=20 > traffic) goes beyond GMPLS as a control plane technique. >=20 > Regards >=20 > Juergen >=20 >=20 > =20 >=20 > > -----Original Message----- > > From: owner-ccamp@ops.ietf.org=20 > > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson > > Sent: Friday, July 22, 2005 12:05 PM > > To: P=E4r Mattsson > > Cc: ccamp@ops.ietf.org > > Subject: Re: Frameformat in a l2cs gmpls rnvironment. > >=20 > > Per and Dimitri, > >=20 > > I would like to come down stronger than that, for me it is > > a very strong requirement that the same switch can handle > > both VLANs and GMPLs trafic correctly. I can't dsee how that > > could be done if using the VLAN tpid to indicate GMPLS > > traffic. > >=20 > > /Loa > >=20 > > P=E4r Mattsson wrote: > > >>hi par, one of the possibilities that has been considered=20 > > to cope with > > >>this requirement is to use a dedicated TPID for the=20 > Ethernet labeled > > >>frames; this would allow differentiated processing with=20 > non-labeled > > >>framesthanks. > > >=20 > > >=20 > > > That seems to make more sence. If that frame is to be sized=20 > > like a 802.1q > > > frame. There is not that much space left to a label. Or is=20 > > the demand to > > > use jumboframes ? > > > Has there been any discussion on labelstacking, and mainly=20 > > where to place > > > the information? > > >=20 > > > Regards. > > > Per > > >=20 > > >=20 > > >=20 > >=20 > >=20 > > --=20 > > Loa Andersson > >=20 > > Principal Networking Architect > > Acreo AB phone: +46 8 632 77 14 > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > Kista, Sweden email: loa.andersson@acreo.se > > loa@pi.se > >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 15:18:06 +0000 Message-ID: <1a5a01c58ed0$c2ed71d0$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <Dimitri.Papadimitriou@alcatel.be> Cc: <ccamp@ops.ietf.org> Subject: Re: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 16:08:34 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Dimitri, > what do you EXACTLY mean with the following sentence > > "This question is one we must come to as soon as we are confident that the > requirements need to be addressed at all." > > please clarify - btw who is "we" in this context ? "We" means "the CCAMP working group" > this sentence reads to me as "CCAMP has not yet committed to work on this > topic" is my reading correct ? or do you mean something else with this ? Your reading is exactly correct as stated. I refer you to your email to the list on 19th July. You asked... > - if yes, is the interest in GMPLS for Ethernet in-/out-side the > CCAMP WG community strong enough to consider the corr. GMPLS > signaling and routing requirements (second part of the document) > as part of the CCAMP WG effort ? Thus, the purpose of the I-D was to crystalize the authors' thoughts on GMPLS control of Ethernet switching and to initiate discussion on whether the proposed deployment models are realistic and desirable. Following a positive answer and assuming support (by CCAMP) for the doing the work (in CCAMP) all that remains is to add the work to the charter (see previous charter discussions) and get on with it. Thanks, Adrian Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 15:17:58 +0000 Message-ID: <1a5b01c58ed0$c42ccfa0$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: "Mack-Crane, T. Benjamin" <Ben.Mack-Crane@tellabs.com> Cc: <ccamp@ops.ietf.org> Subject: Re: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 16:18:30 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Ben, You raised some questions that I can answer with my chair hat comfortably on. > I think Juergen has raised an important question. Yes, it is an important question, even if we are getting ahead of ourselves. I take the question and the excitement as good evidence that we are rapidly moving beyond the issue of whether requirements exist and are now discussing how to implement. > How frames are labeled (and the related data plane forwarding behavior) is > not defined by the control plane. It is true to say that this is not defined by the control plane. Nevertheless, the control plane passes labels around and it is important to define how those labels will be interpreted. I refer you to all of the non-packet labeling mechanisms already defined: you will see that they do not define how the data is labeled, but they do define how the signaled label maps to the labeled data. > The control plane serves to provision the data plane, not define it. Yes. > In the framework draft it is not clear what data plane standards are > covered by the stated control plane requirements. Some references > should be supplied. Such text would not be welcomed by the chairs who specifically instructed it to be left out. You may want to see a complete solution presented in one go, but the chairs want to see the deployment scenarios separated from the solutions work. The debate about how to label packets is distinctly unhelpful in determining what we want to be able to achieve with GMPLS control of Ethernet networks. Further, the requirements may drive us towards one solution rather than another, so placing the solution within the requirements/framework I-D would be dangerous. > In any case, the labeling and forwarding behavior should be defined by these > referenced standards, not by GMPLS. I think you will find that there are adequate standards for forwarding data in Ethernet, without any new standards being defined. The choice which will ultimately be made (FOR EXAMPLE between MAC addresses, VLAN TAGs, or new TPIDs) does not touch existing forwarding paradigms, but might change how those forwarding paradigms are installed. > (I am assuming definition of new data plane standards is out of scope for > CCAMP.) Are you? I would be surprised if this work introduced new data plane standards. Thanks, Adrian Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 15:11:55 +0000 Message-ID: <014201c58ecf$81096c20$6501a8c0@movaz.com> From: "Igor Bryskin" <ibryskin@movaz.com> To: "Mack-Crane, T. Benjamin" <Ben.Mack-Crane@tellabs.com>, "Adrian Farrel" <adrian@olddog.co.uk>, =?iso-8859-1?Q?P=E4r_Mattsson?= <per@defero.se>, "Loa Andersson" <loa@pi.se> Cc: <ccamp@ops.ietf.org> Subject: Re: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 11:10:31 -0400 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_013F_01C58EAD.F99BA260" This is a multi-part message in MIME format. ------=_NextPart_000_013F_01C58EAD.F99BA260 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Hi, I think Ben and Juergen have a point here. There is nothing that could = be dynamically provisioned that could not be also provisioned by = management. In other words data plane is the King here. We need data = plane standard(s) on how we encode a label, whether we need a label = stack or not, how the labeled traffic is supposed to be treated, how = labeled traffic co-exist with non-label traffic, etc. This is something = that not for CCAMP to define. Take for example TDM networks. GMPLS only = provides a way to dynamically provision G.707 networks. Hence there is a = need in parallel standardization activities in ITU. Only after that we = can discuss how all that could be dynamically provisioned, that is the = aspects of control plane.=20 Igor ----- Original Message -----=20 From: Mack-Crane, T. Benjamin=20 To: Adrian Farrel ; P=E4r Mattsson ; Loa Andersson=20 Cc: ccamp@ops.ietf.org=20 Sent: Friday, July 22, 2005 10:23 AM Subject: RE: Frameformat in a l2cs gmpls rnvironment. Hi, I think Juergen has raised an important question. How frames are = labeled (and the related data plane forwarding behavior) is not defined = by the control plane. The control plane serves to provision the data = plane, not define it. In the framework draft it is not clear what data = plane standards are covered by the stated control plane requirements. = Some references should be supplied. In any case, the labeling and = forwarding behavior should be defined by these referenced standards, not = by GMPLS. (I am assuming definition of new data plane standards is out of scope = for CCAMP.) Regards, Ben -------------------------------------------------------------------------= --- From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On = Behalf Of Adrian Farrel Sent: Friday, July 22, 2005 8:37 AM To: P=E4r Mattsson; Loa Andersson Cc: ccamp@ops.ietf.org Subject: Re: Frameformat in a l2cs gmpls rnvironment. Hi, It is always refreshing how engineers jump straight to a discussion = of the solution :-) Perhaps we can assume that the framework draft is a good = representation of the problem space, and that we are ready to start = discussing the solutions. One (perhaps the only?) significant question to answer is how the = frames will be labelled. This question is one we must come to as soon as = we are confident that the requirements need to be addressed at all. As has been pointed out, there are several possibilities and to pick = from these we need to understand: 1. Do we need to support explicit label stacking? Note that this is not supported in TDM, LSC or FSC. 2. Do we need to be able to control (perhaps through an external signaling controller) existing hardware and=20 install LSPs through existing networks? 3. Do we need to support existing function simultaneous to the support of GMPLS L2 LSPs? I think that from a chair's perspective I can give some limited = guidance. A. These questions must be raised and answered in the framework I-D B. The answer to question 3 is "yes". This means that the use of = labels must not significantly deplete any namespace used to support = other function. C. CCAMP is chartered to look at the control of transport networks. = This includes Metro, but I am unsure about Campus. It does not cover=20 signaling to the desktop. Thanks, Adrian ----- Original Message -----=20 From: "P=E4r Mattsson" <per@defero.se> To: "Loa Andersson" <loa@pi.se> Cc: <ccamp@ops.ietf.org> Sent: Friday, July 22, 2005 11:41 AM Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > Per and Dimitri, > > > > I would like to come down stronger than that, for me it is > > a very strong requirement that the same switch can handle > > both VLANs and GMPLs trafic correctly. I can't dsee how that > > could be done if using the VLAN tpid to indicate GMPLS > > traffic. >=20 > If you ever want that same switch to handle traffic for a directly > connected host (not to uncommen) you would want that to use = normal > ethernet macaddress switching. So of course you do not want to = have to > choose between vlan and gmpls, you would want both at the same = time. >=20 > /per >=20 >=20 > > > > /Loa > > > > P=E4r Mattsson wrote: > >>>hi par, one of the possibilities that has been considered to = cope with > >>>this requirement is to use a dedicated TPID for the Ethernet = labeled > >>>frames; this would allow differentiated processing with = non-labeled > >>>framesthanks. > >> > >> > >> That seems to make more sence. If that frame is to be sized = like a > >> 802.1q > >> frame. There is not that much space left to a label. Or is the = demand to > >> use jumboframes ? > >> Has there been any discussion on labelstacking, and mainly = where to > >> place > >> the information? > >> > >> Regards. > >> Per > >> > >> > >> > > > > > > -- > > Loa Andersson > > > > Principal Networking Architect > > Acreo AB phone: +46 8 632 77 14 > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > Kista, Sweden email: = loa.andersson@acreo.se > > loa@pi.se > > >=20 >=20 >=20 >=20 =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any reproduction, dissemination or distribution of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. Tellabs =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D ------=_NextPart_000_013F_01C58EAD.F99BA260 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META http-equiv=3DContent-Type content=3D"text/html; = charset=3Diso-8859-1"> <META content=3D"MSHTML 6.00.2800.1400" name=3DGENERATOR> <STYLE></STYLE> </HEAD> <BODY bgColor=3D#ffffff> <DIV><FONT face=3DArial size=3D2>Hi,</FONT></DIV> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <DIV><FONT face=3DArial size=3D2>I think Ben and Juergen have a point = here. There is=20 nothing that could be dynamically provisioned that could not be also = provisioned=20 by management. In other words data plane is the King here. We need data = plane=20 standard(s) on how we encode a label, whether we need a label stack or = not, how=20 the labeled traffic is supposed to be treated, how labeled traffic = co-exist=20 with non-label traffic, etc. This is something that not for CCAMP to = define.=20 Take for example TDM networks. GMPLS only provides a way to dynamically=20 provision G.707 networks. Hence there is a need in parallel = standardization=20 activities in ITU. Only after that we can discuss how all that could be=20 dynamically provisioned, that is the aspects of control plane. = </FONT></DIV> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <DIV><FONT face=3DArial size=3D2>Igor</FONT></DIV> <BLOCKQUOTE dir=3Dltr=20 style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; = BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px"> <DIV style=3D"FONT: 10pt arial">----- Original Message ----- </DIV> <DIV=20 style=3D"BACKGROUND: #e4e4e4; FONT: 10pt arial; font-color: = black"><B>From:</B>=20 <A title=3DBen.Mack-Crane@tellabs.com=20 href=3D"mailto:Ben.Mack-Crane@tellabs.com">Mack-Crane, T. Benjamin</A> = </DIV> <DIV style=3D"FONT: 10pt arial"><B>To:</B> <A = title=3Dadrian@olddog.co.uk=20 href=3D"mailto:adrian@olddog.co.uk">Adrian Farrel</A> ; <A = title=3Dper@defero.se=20 href=3D"mailto:per@defero.se">P=E4r Mattsson</A> ; <A = title=3Dloa@pi.se=20 href=3D"mailto:loa@pi.se">Loa Andersson</A> </DIV> <DIV style=3D"FONT: 10pt arial"><B>Cc:</B> <A = title=3Dccamp@ops.ietf.org=20 href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A> </DIV> <DIV style=3D"FONT: 10pt arial"><B>Sent:</B> Friday, July 22, 2005 = 10:23=20 AM</DIV> <DIV style=3D"FONT: 10pt arial"><B>Subject:</B> RE: Frameformat in a = l2cs gmpls=20 rnvironment.</DIV> <DIV><FONT face=3DArial size=3D2></FONT><BR></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>Hi,</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2></FONT></SPAN> </DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>I think Juergen has raised an important = question. =20 How frames are labeled (and the related data plane forwarding = behavior) is not=20 defined by the control plane. The control plane serves to = provision the=20 data plane, not define it. In the framework draft it is not = clear what=20 data plane standards are covered by the stated control plane=20 requirements. Some references should be supplied. In any = case, the=20 labeling and forwarding behavior should be defined by these = referenced=20 standards, not by GMPLS.</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2></FONT></SPAN> </DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>(I am assuming definition of new data plane = standards is=20 out of scope for CCAMP.)</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2></FONT></SPAN> </DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>Regards,</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT = face=3DArial=20 color=3D#0000ff size=3D2>Ben</FONT></SPAN></DIV><BR> <BLOCKQUOTE=20 style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #0000ff 2px = solid; MARGIN-RIGHT: 0px"> <DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr = align=3Dleft> <HR tabIndex=3D-1> <FONT face=3DTahoma size=3D2><B>From:</B> owner-ccamp@ops.ietf.org=20 [mailto:owner-ccamp@ops.ietf.org] <B>On Behalf Of </B>Adrian=20 Farrel<BR><B>Sent:</B> Friday, July 22, 2005 8:37 AM<BR><B>To:</B> = P=E4r=20 Mattsson; Loa Andersson<BR><B>Cc:</B> = ccamp@ops.ietf.org<BR><B>Subject:</B>=20 Re: Frameformat in a l2cs gmpls rnvironment.<BR></FONT><BR></DIV> <DIV></DIV> <DIV><FONT face=3DCourier size=3D2>Hi,</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>It is always refreshing how = engineers jump=20 straight to a discussion of the solution :-)</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Perhaps we can assume that the = framework=20 draft is a good representation of the problem space, and that we are = ready=20 to start discussing the solutions.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>One (perhaps the only?) = significant question=20 to answer is how the frames will be labelled. This question is one = we must=20 come to as soon as we are confident that the requirements need to be = addressed at all.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>As has been pointed out, there = are several=20 possibilities and to pick from these we need to = understand:</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>1. Do we need to support explicit = label=20 stacking?</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> Note that this is = not supported=20 in TDM, LSC or FSC.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>2. Do we need to be able to = control (perhaps=20 through an</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> external = signaling=20 controller) existing hardware and </FONT></DIV> <DIV><FONT face=3DCourier size=3D2> install LSPs through = existing=20 networks?</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>3. Do we need to support existing = function=20 simultaneous to</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> the support of GMPLS = L2=20 LSPs?</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>I think that from a chair's = perspective I can=20 give some limited guidance.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>A. These questions must be raised = and=20 answered in the framework I-D</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>B. The answer to question 3 is = "yes". This=20 means that the use of labels</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> must not = significantly deplete=20 any namespace used to support other</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> = function.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>C. CCAMP is chartered to look at = the control=20 of transport networks. This</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> includes Metro, = but I am=20 unsure about Campus. It does not cover </FONT></DIV> <DIV><FONT face=3DCourier size=3D2> signaling to the=20 desktop.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Thanks,</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Adrian</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>----- Original Message ----- = </FONT> <DIV><FONT face=3DCourier size=3D2>From: "P=E4r Mattsson" = <</FONT><A=20 href=3D"mailto:per@defero.se"><FONT face=3DCourier=20 size=3D2>per@defero.se</FONT></A><FONT face=3DCourier = size=3D2>></FONT></DIV> <DIV><FONT face=3DCourier size=3D2>To: "Loa Andersson" <</FONT><A = href=3D"mailto:loa@pi.se"><FONT face=3DCourier = size=3D2>loa@pi.se</FONT></A><FONT=20 face=3DCourier size=3D2>></FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Cc: <</FONT><A=20 href=3D"mailto:ccamp@ops.ietf.org"><FONT face=3DCourier=20 size=3D2>ccamp@ops.ietf.org</FONT></A><FONT face=3DCourier=20 size=3D2>></FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Sent: Friday, July 22, 2005 11:41 = AM</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Subject: Re: Frameformat in a = l2cs gmpls=20 rnvironment.</FONT></DIV></DIV> <DIV><FONT face=3DCourier><BR><FONT = size=3D2></FONT></FONT></DIV><FONT=20 face=3DCourier size=3D2>> > Per and Dimitri,<BR>> = ><BR>> > I=20 would like to come down stronger than that, for me it is<BR>> = > a very=20 strong requirement that the same switch can handle<BR>> > both = VLANs=20 and GMPLs trafic correctly. I can't dsee how that<BR>> > could = be done=20 if using the VLAN tpid to indicate GMPLS<BR>> > = traffic.<BR>>=20 <BR>> If you ever want that same switch to handle traffic for a=20 directly<BR>> connected host (not to uncommen) you would = want that=20 to use normal<BR>> ethernet macaddress switching. So of course = you do not=20 want to have to<BR>> choose between vlan and gmpls, you would = want both=20 at the same time.<BR>> <BR>> /per<BR>> <BR>> <BR>>=20 ><BR>> > /Loa<BR>> ><BR>> > P=E4r Mattsson = wrote:<BR>>=20 >>>hi par, one of the possibilities that has been = considered to=20 cope with<BR>> >>>this requirement is to use a dedicated = TPID=20 for the Ethernet labeled<BR>> >>>frames; this would = allow=20 differentiated processing with non-labeled<BR>>=20 >>>framesthanks.<BR>> >><BR>> >><BR>> = >>=20 That seems to make more sence. If that frame is to be sized like = a<BR>>=20 >> 802.1q<BR>> >> frame. There is not that much space = left to=20 a label. Or is the demand to<BR>> >> use jumboframes = ?<BR>>=20 >> Has there been any discussion on labelstacking, and mainly = where=20 to<BR>> >> place<BR>> >> the information?<BR>>=20 >><BR>> >> Regards.<BR>> >> Per<BR>>=20 >><BR>> >><BR>> >><BR>> ><BR>> = ><BR>>=20 > --<BR>> > Loa Andersson<BR>> ><BR>> > = Principal=20 Networking Architect<BR>> > Acreo=20 = AB  = ; = =20 phone: +46 8 632 77 14<BR>> > Isafjordsgatan=20 = 22  = ; =20 mobile: +46 739 81 21 64<BR>> > Kista,=20 = Sweden &= nbsp; =20 email: </FONT><A href=3D"mailto:loa.andersson@acreo.se"><FONT=20 face=3DCourier size=3D2>loa.andersson@acreo.se</FONT></A><BR><FONT = face=3DCourier=20 size=3D2>>=20 = > &nb= sp; &nbs= p;  = ; =20 </FONT><A href=3D"mailto:loa@pi.se"><FONT face=3DCourier=20 size=3D2>loa@pi.se</FONT></A><BR><FONT face=3DCourier size=3D2>> = ><BR>>=20 <BR>> <BR>> <BR>> = </FONT></BLOCKQUOTE><PRE>=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any reproduction, dissemination or distribution of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. Tellabs =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D </PRE></BLOCKQUOTE></BODY></HTML> ------=_NextPart_000_013F_01C58EAD.F99BA260-- Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 15:01:10 +0000 Message-ID: <42E10945.2030904@pi.se> Date: Fri, 22 Jul 2005 16:57:09 +0200 From: Loa Andersson <loa@pi.se> Organization: Acreo AB User-Agent: Mozilla Thunderbird 1.0.5 (Windows/20050711) MIME-Version: 1.0 To: Heiles Juergen <juergen.heiles@siemens.com> CC: richard.spencer@bt.com, per@defero.se, ccamp@ops.ietf.org Subject: Re: Frameformat in a l2cs gmpls rnvironment. Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 8bit Juergen, that is only because we were told to avoid "solutions" in the framework, consequently we who route studd that is not necessarily based on vlan did not discuss how we wanted to "label" the frame. The proposal to use the vlan (tpid 8100) did include the solution in the document, which if we listen the wg chairs it should not have. What I belive is that when we start use GMPLS labled frames in an Ethernet network the behavior will be different from those that are vlan-tagged, hence it won't be possible to use the vlan-tag as gmpls-label, the switch can distinguish betwen them. Also there is a clear requirement in the section 6.2 that labels should be locally significant. /Loa Heiles Juergen wrote: > Loa, > > I interpret the ID as a proposal to use GMPLS for VLAN setup. So GMPLS and VLAN to not compete. The VLAN is at the data plane and GMPLS at the control plane. The question is how and should different control plane techniques like GMPLS and (GVRP and STB) work together? > > Regards > > Juergen > > > >>-----Original Message----- >>From: owner-ccamp@ops.ietf.org >>[mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson >>Sent: Friday, July 22, 2005 2:53 PM >>To: richard.spencer@bt.com >>Cc: per@defero.se; ccamp@ops.ietf.org >>Subject: Re: Frameformat in a l2cs gmpls rnvironment. >> >>Richard, >> >>I agree to most of this. Since we have doubts about the viability >>of taking GMPLS all the way to end-user or enterprise I think it >>would be good, from a wg perspective, if we agreed to solve the >>core network problems first. >> >>Do you have any comment on the requirement to run both VLANs and >>GMPLS on the same switch? >> >>/Loa >> >>richard.spencer@bt.com wrote: >> >>>Regardless of whether or not a switch is directly connected >> >> > to hosts, it must be able to forward packets using the >>connectionless >> > Ethernet data plane. This is due to the fundamental >>requirement that >> > for GMPLS switches to be able to exchange control information with >> > each other, a data plane for control traffic must be present. >> > This is akin to using the IP data plane for MPLS signalling in an >> > IP/MPLS network. An alternative would be to use a static reserved >> > L2-LSP for control traffic in the same way that reserved VPI/VCIs >> > are used for PNNI signalling in ATM. >> >>>Regarding connecting hosts to GMPLS switches, I personally don't >> >> > think extending L2-LSPs into the enterprise/home network is >> > commercially viable. However, if you do want to use GMPLS switches >> > in the home/enterprise network and for some reason don't want to >> > extend L2-LSPs down to the host then you will not be performing >> > normal Ethernet Mac address switching anyway. Instead you will >> > need some kind of policy on the switch that maps connectionless >> > Ethernet packets (e.g. based on MAC src/dest, 802.1p, VLAN) to a >> > L2-LSP. This is because multiple L2-LSPs to the same destination >> > (e.g. a gateway router) may exist for different services/flows >> > (e.g. video download, VoIP call, etc.). >> >>>Regards, >>>Richard >>> >>> >>> >>>>-----Original Message----- >>>>From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On >>>>Behalf Of Pär Mattsson >>>>Sent: 22 July 2005 11:42 >>>>To: Loa Andersson >>>>Cc: ccamp@ops.ietf.org >>>>Subject: Re: Frameformat in a l2cs gmpls rnvironment. >>>> >>>> >>>> >>>> >>>>>Per and Dimitri, >>>>> >>>>>I would like to come down stronger than that, for me it is >>>>>a very strong requirement that the same switch can handle >>>>>both VLANs and GMPLs trafic correctly. I can't dsee how that >>>>>could be done if using the VLAN tpid to indicate GMPLS >>>>>traffic. >>>> >>>>If you ever want that same switch to handle traffic for a directly >>>>connected host (not to uncommen) you would want that to use normal >>>>ethernet macaddress switching. So of course you do not want >> >>to have to >> >>>>choose between vlan and gmpls, you would want both at the same time. >>>> >>>>/per >>>> >>>> >>>> >>>> >>>>>/Loa >>>>> >>>>>Pär Mattsson wrote: >>>>> >>>>> >>>>>>>hi par, one of the possibilities that has been considered >>>> >>>>to cope with >>>> >>>> >>>>>>>this requirement is to use a dedicated TPID for the >>>> >>>>Ethernet labeled >>>> >>>> >>>>>>>frames; this would allow differentiated processing with >> >>non-labeled >> >>>>>>>framesthanks. >>>>>> >>>>>> >>>>>>That seems to make more sence. If that frame is to be sized like a >>>>>>802.1q >>>>>>frame. There is not that much space left to a label. Or is >>>> >>>>the demand to >>>> >>>> >>>>>>use jumboframes ? >>>>>>Has there been any discussion on labelstacking, and >> >>mainly where to >> >>>>>>place >>>>>>the information? >>>>>> >>>>>>Regards. >>>>>>Per >>>>>> >>>>>> >>>>>> >>>>> >>>>> >>>>>-- >>>>>Loa Andersson >>>>> >>>>>Principal Networking Architect >>>>>Acreo AB phone: +46 8 632 77 14 >>>>>Isafjordsgatan 22 mobile: +46 739 81 21 64 >>>>>Kista, Sweden email: loa.andersson@acreo.se >>>>> loa@pi.se >>>>> >>>> >>>> >>>> >> >>-- >>Loa Andersson >> >>Principal Networking Architect >>Acreo AB phone: +46 8 632 77 14 >>Isafjordsgatan 22 mobile: +46 739 81 21 64 >>Kista, Sweden email: loa.andersson@acreo.se >> loa@pi.se >> > > > -- Loa Andersson Principal Networking Architect Acreo AB phone: +46 8 632 77 14 Isafjordsgatan 22 mobile: +46 739 81 21 64 Kista, Sweden email: loa.andersson@acreo.se loa@pi.se Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 14:26:03 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01C58EC9.2CDB7192" Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 09:23:59 -0500 Message-ID: <A1A52203CA93634BA1748887B9993AEA010468EA@USNVEX1.tellabs-west.tellabsinc.net> Thread-Topic: Frameformat in a l2cs gmpls rnvironment. Thread-Index: AcWOwsSnwxhnILPySYOqhzXS5kDp+wABSG7w From: "Mack-Crane, T. Benjamin" <Ben.Mack-Crane@tellabs.com> To: "Adrian Farrel" <adrian@olddog.co.uk>, =?iso-8859-1?Q?P=E4r_Mattsson?= <per@defero.se>, "Loa Andersson" <loa@pi.se> Cc: <ccamp@ops.ietf.org> This is a multi-part message in MIME format. ------_=_NextPart_001_01C58EC9.2CDB7192 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Hi, =20 I think Juergen has raised an important question. How frames are labeled (= and the related data plane forwarding behavior) is not defined by the contr= ol plane. The control plane serves to provision the data plane, not define= it. In the framework draft it is not clear what data plane standards are = covered by the stated control plane requirements. Some references should b= e supplied. In any case, the labeling and forwarding behavior should be de= fined by these referenced standards, not by GMPLS. =20 (I am assuming definition of new data plane standards is out of scope for C= CAMP.) =20 Regards, Ben ________________________________ From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On Behalf= Of Adrian Farrel Sent: Friday, July 22, 2005 8:37 AM To: P=E4r Mattsson; Loa Andersson Cc: ccamp@ops.ietf.org Subject: Re: Frameformat in a l2cs gmpls rnvironment. =09 =09 Hi, =20 It is always refreshing how engineers jump straight to a discussion of the= solution :-) =20 Perhaps we can assume that the framework draft is a good representation of= the problem space, and that we are ready to start discussing the solutions. =20 One (perhaps the only?) significant question to answer is how the frames w= ill be labelled. This question is one we must come to as soon as we are con= fident that the requirements need to be addressed at all. =20 As has been pointed out, there are several possibilities and to pick from = these we need to understand: 1. Do we need to support explicit label stacking? Note that this is not supported in TDM, LSC or FSC. 2. Do we need to be able to control (perhaps through an external signaling controller) existing hardware and=20 install LSPs through existing networks? 3. Do we need to support existing function simultaneous to the support of GMPLS L2 LSPs? =20 I think that from a chair's perspective I can give some limited guidance. =20 A. These questions must be raised and answered in the framework I-D B. The answer to question 3 is "yes". This means that the use of labels must not significantly deplete any namespace used to support other function. C. CCAMP is chartered to look at the control of transport networks. This includes Metro, but I am unsure about Campus. It does not cover=20 signaling to the desktop. =20 Thanks, Adrian =20 ----- Original Message -----=20 From: "P=E4r Mattsson" <per@defero.se <mailto:per@defero.se> > To: "Loa Andersson" <loa@pi.se <mailto:loa@pi.se> > Cc: <ccamp@ops.ietf.org <mailto:ccamp@ops.ietf.org> > Sent: Friday, July 22, 2005 11:41 AM Subject: Re: Frameformat in a l2cs gmpls rnvironment. =09 =09 > > Per and Dimitri, > > > > I would like to come down stronger than that, for me it is > > a very strong requirement that the same switch can handle > > both VLANs and GMPLs trafic correctly. I can't dsee how that > > could be done if using the VLAN tpid to indicate GMPLS > > traffic. >=20 > If you ever want that same switch to handle traffic for a directly > connected host (not to uncommen) you would want that to use normal > ethernet macaddress switching. So of course you do not want to have to > choose between vlan and gmpls, you would want both at the same time. >=20 > /per >=20 >=20 > > > > /Loa > > > > P=E4r Mattsson wrote: > >>>hi par, one of the possibilities that has been considered to cope with > >>>this requirement is to use a dedicated TPID for the Ethernet labeled > >>>frames; this would allow differentiated processing with non-labeled > >>>framesthanks. > >> > >> > >> That seems to make more sence. If that frame is to be sized like a > >> 802.1q > >> frame. There is not that much space left to a label. Or is the demand= to > >> use jumboframes ? > >> Has there been any discussion on labelstacking, and mainly where to > >> place > >> the information? > >> > >> Regards. > >> Per > >> > >> > >> > > > > > > -- > > Loa Andersson > > > > Principal Networking Architect > > Acreo AB phone: +46 8 632 77 14 > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > Kista, Sweden email: loa.andersson@acreo.se <mai= lto:loa.andersson@acreo.se>=20 > > loa@pi.se <mailto:loa@pi.s= e>=20 > > >=20 >=20 >=20 >=20 =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any reproduction, dissemination or distribution of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. Tellabs =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D ------_=_NextPart_001_01C58EC9.2CDB7192 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META http-equiv=3DContent-Type content=3D"text/html; charset=3Diso-8859-1"> <META content=3D"MSHTML 6.00.2800.1106" name=3DGENERATOR> <STYLE></STYLE> </HEAD> <BODY> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT face=3DA= rial=20 color=3D#0000ff size=3D2>Hi,</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT face=3DA= rial=20 color=3D#0000ff size=3D2></FONT></SPAN> </DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT face=3DA= rial=20 color=3D#0000ff size=3D2>I think Juergen has raised an important question.&= nbsp; How=20 frames are labeled (and the related data plane forwarding behavior) is not=20 defined by the control plane. The control plane serves to provision t= he=20 data plane, not define it. In the framework draft it is not clear wha= t=20 data plane standards are covered by the stated control plane requirements.&= nbsp;=20 Some references should be supplied. In any case, the labeling and=20 forwarding behavior should be defined by these referenced standards,=20 not by GMPLS.</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT face=3DA= rial=20 color=3D#0000ff size=3D2></FONT></SPAN> </DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT face=3DA= rial=20 color=3D#0000ff size=3D2>(I am assuming definition of new data plane standa= rds is=20 out of scope for CCAMP.)</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT face=3DA= rial=20 color=3D#0000ff size=3D2></FONT></SPAN> </DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT face=3DA= rial=20 color=3D#0000ff size=3D2>Regards,</FONT></SPAN></DIV> <DIV dir=3Dltr align=3Dleft><SPAN class=3D481061614-22072005><FONT face=3DA= rial=20 color=3D#0000ff size=3D2>Ben</FONT></SPAN></DIV><BR> <BLOCKQUOTE=20 style=3D"PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #0000ff 2px soli= d; MARGIN-RIGHT: 0px"> <DIV class=3DOutlookMessageHeader lang=3Den-us dir=3Dltr align=3Dleft> <HR tabIndex=3D-1> <FONT face=3DTahoma size=3D2><B>From:</B> owner-ccamp@ops.ietf.org=20 [mailto:owner-ccamp@ops.ietf.org] <B>On Behalf Of </B>Adrian=20 Farrel<BR><B>Sent:</B> Friday, July 22, 2005 8:37 AM<BR><B>To:</B> P=E4r=20 Mattsson; Loa Andersson<BR><B>Cc:</B> ccamp@ops.ietf.org<BR><B>Subject:</= B>=20 Re: Frameformat in a l2cs gmpls rnvironment.<BR></FONT><BR></DIV> <DIV></DIV> <DIV><FONT face=3DCourier size=3D2>Hi,</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>It is always refreshing how engineers = jump=20 straight to a discussion of the solution :-)</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Perhaps we can assume that the framewo= rk draft=20 is a good representation of the problem space, and that we are ready to s= tart=20 discussing the solutions.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>One (perhaps the only?) significant qu= estion to=20 answer is how the frames will be labelled. This question is one we must c= ome=20 to as soon as we are confident that the requirements need to be addressed= at=20 all.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>As has been pointed out, there are sev= eral=20 possibilities and to pick from these we need to understand:</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>1. Do we need to support explicit labe= l=20 stacking?</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> Note that this is not sup= ported in=20 TDM, LSC or FSC.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>2. Do we need to be able to control (p= erhaps=20 through an</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> external signaling=20 controller) existing hardware and </FONT></DIV> <DIV><FONT face=3DCourier size=3D2> install LSPs through exis= ting=20 networks?</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>3. Do we need to support existing func= tion=20 simultaneous to</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> the support of GMPLS L2=20 LSPs?</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>I think that from a chair's perspectiv= e I can=20 give some limited guidance.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>A. These questions must be raised and = answered=20 in the framework I-D</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>B. The answer to question 3 is "yes". = This=20 means that the use of labels</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> must not significantly de= plete any=20 namespace used to support other</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> function.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>C. CCAMP is chartered to look at the c= ontrol of=20 transport networks. This</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> includes Metro, but = I am=20 unsure about Campus. It does not cover </FONT></DIV> <DIV><FONT face=3DCourier size=3D2> signaling to the=20 desktop.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Thanks,</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Adrian</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>----- Original Message ----- </FONT> <DIV><FONT face=3DCourier size=3D2>From: "P=E4r Mattsson" <</FONT><A=20 href=3D"mailto:per@defero.se"><FONT face=3DCourier=20 size=3D2>per@defero.se</FONT></A><FONT face=3DCourier size=3D2>></FONT= ></DIV> <DIV><FONT face=3DCourier size=3D2>To: "Loa Andersson" <</FONT><A=20 href=3D"mailto:loa@pi.se"><FONT face=3DCourier size=3D2>loa@pi.se</FONT><= /A><FONT=20 face=3DCourier size=3D2>></FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Cc: <</FONT><A=20 href=3D"mailto:ccamp@ops.ietf.org"><FONT face=3DCourier=20 size=3D2>ccamp@ops.ietf.org</FONT></A><FONT face=3DCourier=20 size=3D2>></FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Sent: Friday, July 22, 2005 11:41=20 AM</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Subject: Re: Frameformat in a l2cs gmp= ls=20 rnvironment.</FONT></DIV></DIV> <DIV><FONT face=3DCourier><BR><FONT size=3D2></FONT></FONT></DIV><FONT=20 face=3DCourier size=3D2>> > Per and Dimitri,<BR>> ><BR>> &= gt; I=20 would like to come down stronger than that, for me it is<BR>> > a v= ery=20 strong requirement that the same switch can handle<BR>> > both VLAN= s and=20 GMPLs trafic correctly. I can't dsee how that<BR>> > could be done = if=20 using the VLAN tpid to indicate GMPLS<BR>> > traffic.<BR>> <BR>&= gt;=20 If you ever want that same switch to handle traffic for a directly<BR>>= ;=20 connected host (not to uncommen) you would want that to use=20 normal<BR>> ethernet macaddress switching. So of course you do not wan= t to=20 have to<BR>> choose between vlan and gmpls, you would want both at the= same=20 time.<BR>> <BR>> /per<BR>> <BR>> <BR>> ><BR>> >=20 /Loa<BR>> ><BR>> > P=E4r Mattsson wrote:<BR>> >>>= hi par,=20 one of the possibilities that has been considered to cope with<BR>>=20 >>>this requirement is to use a dedicated TPID for the Ethernet=20 labeled<BR>> >>>frames; this would allow differentiated proce= ssing=20 with non-labeled<BR>> >>>framesthanks.<BR>> >><BR>&g= t;=20 >><BR>> >> That seems to make more sence. If that frame is= to=20 be sized like a<BR>> >> 802.1q<BR>> >> frame. There is = not=20 that much space left to a label. Or is the demand to<BR>> >> use=20 jumboframes ?<BR>> >> Has there been any discussion on labelstac= king,=20 and mainly where to<BR>> >> place<BR>> >> the=20 information?<BR>> >><BR>> >> Regards.<BR>> >>=20 Per<BR>> >><BR>> >><BR>> >><BR>> ><BR>&g= t;=20 ><BR>> > --<BR>> > Loa Andersson<BR>> ><BR>> >=20 Principal Networking Architect<BR>> > Acreo=20 AB  = ; &n= bsp; =20 phone: +46 8 632 77 14<BR>> > Isafjordsgatan=20 22  = ; =20 mobile: +46 739 81 21 64<BR>> > Kista,=20 Sweden &= nbsp; =20 email: </FONT><A href=3D"mailto:loa.andersson@acreo.se"><FONT face= =3DCourier=20 size=3D2>loa.andersson@acreo.se</FONT></A><BR><FONT face=3DCourier size= =3D2>>=20 > &nb= sp; = &nb= sp; =20 </FONT><A href=3D"mailto:loa@pi.se"><FONT face=3DCourier=20 size=3D2>loa@pi.se</FONT></A><BR><FONT face=3DCourier size=3D2>> ><= BR>>=20 <BR>> <BR>> <BR>> </FONT></BLOCKQUOTE><pre>=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any reproduction, dissemination or distribution of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. Tellabs =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D </pre></BODY></HTML> ------_=_NextPart_001_01C58EC9.2CDB7192-- Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 14:20:46 +0000 To: "Adrian Farrel" <adrian@olddog.co.uk> Cc: =?iso-8859-1?Q?P=E4r_Mattsson?= <per@defero.se>, "Loa Andersson" <loa@pi.se>, <ccamp@ops.ietf.org> From: Dimitri.Papadimitriou@alcatel.be Subject: Re: Frameformat in a l2cs gmpls rnvironment. MIME-Version: 1.0 Date: Fri, 22 Jul 2005 16:19:35 +0200 Message-ID: <OF29F9F78D.85672E04-ONC1257046.004EB250-C1257046.004EB2AF@netfr.alcatel.fr> Content-type: text/plain; charset=us-ascii adrian - what do you EXACTLY mean with the following sentence "This question is one we must come to as soon as we are confident that the requirements need to be addressed at all." please clarify - btw who is "we" in this context ? this sentence reads to me as "CCAMP has not yet committed to work on this topic" is my reading correct ? or do you mean something else with this ? thanks, - dimitri. Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 14:05:03 +0000 To: Heiles Juergen <juergen.heiles@siemens.com> Cc: Loa Andersson <loa@pi.se>, =?iso-8859-1?Q?P=E4r_Mattsson?= <per@defero.se>, ccamp@ops.ietf.org MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 16:03:13 +0200 Message-ID: <OF14688868.09D9C2E2-ONC1257046.004D3278-C1257046.004D3310@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmhpIGp1ZXJnZW4sPC9GT05UPjwvUD48 UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+aWYgd2VsbCB1bmRlcnN0YW5kIHlvdXIg cG9pbnQgb2YgdmlldywgeW91IHdvdWxkIGJlIGluIGZhdm9yIG9mIHNldHRpbmcgdXAgcGF0aCB3 aXRob3V0IGludHJvZHVjaW5nIGFueSBzcGVjaWZpYyBpbmRpY2F0aW9uIGluIHRoZSBkYXRhIHBs YW5lIEV0aGVybmV0IGZyYW1lIGZsb3c8L0ZPTlQ+PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFj ZSxDb3VyaWVyIj5tb3JlIHByZWNpc2VseSwgb25jZSB0aGUgJnF1b3Q7bG9naWNhbCBwYXRoJnF1 b3Q7IGlzIHNldHVwIHVzaW5nIHRoZSBjb250cm9sIHBsYW5lLCB0aGUgZm9yd2FyZGluZyBiZWhh dmlvdXIgd291bGQgcmVtYWluIHVuY2hhbmdlZCBmcm9tIGxlZ2FjeSBFdGhlcm5ldCBiZWhhdmlv dXIgKGkuZSBNQUMgYWRkcmVzcyBsZWFybmluZyBhbmQgZm9yd2FyZGluZyBiYXNlZCBvbiBNQUMg REEpIGluIHRoaXMgY2FzZSwgdGhlIGNvbnRyb2wgcGxhbmUgZnVuY3Rpb24gcmVwbGFjZXMgdGhl IFZMQU4gaWQgY29uZmlndXJhdGlvbiBwZXJmb3JtZWQgdG9kYXkgbWFudWFsbHkgb24gYSBwZXIg aW50ZXJmYWNlIGJhc2lzID8gLSBwbGVhc2UgY29uZmlybTwvRk9OVD48L1A+PFA+PEZPTlQgRkFD RT0iTW9ub3NwYWNlLENvdXJpZXIiPm5vdywgaSB3b3VsZCBpbnRlcmVzdGVkIHRvIGtub3cgd2hh dCBpbiB5b3VyIHZpZXcgd291bGQgYmUgdGhlIHJvbGUgb2YgODAyLjFkIGFuZCByZWxhdGVkIGlu IHRoaXMgY29udGV4dCAoYWxzbyB0aGVyZSBhcmUgc2V2ZXJhbCBsZXZlbHMgb2YgaW50ZXJhY3Rp b24gcG9zc2libGUgaGVyZSk8L0ZPTlQ+PC9QPjxQPjxCUj48Rk9OVCBTSVpFPTI+PEI+IEhlaWxl cyBKdWVyZ2VuICZsdDtqdWVyZ2VuLmhlaWxlc0BzaWVtZW5zLmNvbSZndDs8L0I+PC9GT05UPjxC Uj48Rk9OVCBTSVpFPTI+U2VudCBieTogb3duZXItY2NhbXBAb3BzLmlldGYub3JnPC9GT05UPjxC Uj48Rk9OVCBTSVpFPTI+MDcvMjIvMjAwNSAxNToxNyBaRTI8L0ZPTlQ+PEJSPjxCUj4gPEZPTlQg U0laRT0yPlRvOjwvRk9OVD4gPEZPTlQgU0laRT0yPkxvYSBBbmRlcnNzb24gJmx0O2xvYUBwaS5z ZSZndDssIFAmYXVtbDtyIE1hdHRzc29uICZsdDtwZXJAZGVmZXJvLnNlJmd0OzwvRk9OVD48QlI+ IDxGT05UIFNJWkU9Mj5jYzo8L0ZPTlQ+IDxGT05UIFNJWkU9Mj5jY2FtcEBvcHMuaWV0Zi5vcmc8 L0ZPTlQ+PEJSPiA8Rk9OVCBTSVpFPTI+YmNjOjwvRk9OVD4gPEJSPiA8Rk9OVCBTSVpFPTI+U3Vi amVjdDo8L0ZPTlQ+IDxGT05UIFNJWkU9Mj5SRTogRnJhbWVmb3JtYXQgaW4gYSBsMmNzIGdtcGxz IHJudmlyb25tZW50LjwvRk9OVD48QlI+IDxCUj48QlI+PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9z cGFjZSxDb3VyaWVyIj5HTVBMUyBmb3IgU09ORVQvU0RIIGhhcyBubyBpbXBhY3Qgb24gdGhlIFNP TkVUL1NESCBkYXRhIHBsYW5lLiBJdCBpbnRyb2R1Y2VzIGEgY29udHJvbCBwbGFuZSBmb3IgY29u bmVjdGlvbiBzZXR1cCBpbnN0ZWFkIG9mIGEgbmV0d29yayBtYW5hZ2VtZW50IGJhc2VkIGNvbm5l Y3Rpb24gc2V0dXAuIEdNUExTIGZvciBFdGhlcm5ldCBjYW4gd29yayBpbiB0aGUgc2FtZSB3YXku IEV0aGVybmV0IGFzIHN1Y2ggaXMgY2wsIGJ1dCBWTEFOcyBhcmUgc2V0dXAgYW5kIEdNUExTIGNh biBwcm92aWRlIGEgY29udHJvbCBwbGFuZSBmb3IgVkxBTiBzZXR1cC4gVGhlIEV0aGVybmV0IGRh dGEgcGxhbmUgaXMgbm90IGNoYW5nZWQgYW5kIE1BQyBhZGRyZXNzIGxlYXJuaW5nIGFuZCBmb3J3 YXJkaW5nIGlzIHN0aWxsIGJlIGRvbmUgd2l0aGluIGEgVkxBTi4gU28gbm8gY2hhbmdlIHRvIHRo ZSBFdGhlcm5ldCBkYXRhIHBsYW5lLiBObyBHTVBMUyB0cmFmZmljIGlzIGludHJvZHVjZWQsIGl0 IGlzIHN0aWxsIFZMQU4gdHJhZmZpYy48QlI+SW50cm9kdWNpbmcgYSBkZWRpY2F0ZWQgR01QTFMg bGFiZWwgZm9yIEV0aGVybmV0IChHTVBMUyB0cmFmZmljKSBnb2VzIGJleW9uZCBHTVBMUyBhcyBh IGNvbnRyb2wgcGxhbmUgdGVjaG5pcXVlLjxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9z cGFjZSxDb3VyaWVyIj5SZWdhcmRzPEJSPjwvRk9OVD48QlI+PEZPTlQgRkFDRT0iTW9ub3NwYWNl LENvdXJpZXIiPkp1ZXJnZW48QlI+PC9GT05UPjxCUj48QlI+PEJSPjxCUj48Rk9OVCBGQUNFPSJN b25vc3BhY2UsQ291cmllciI+Jmd0OyAtLS0tLU9yaWdpbmFsIE1lc3NhZ2UtLS0tLTxCUj4mZ3Q7 IEZyb206IG93bmVyLWNjYW1wQG9wcy5pZXRmLm9yZzxCUj4mZ3Q7IFttYWlsdG86b3duZXItY2Nh bXBAb3BzLmlldGYub3JnXSBPbiBCZWhhbGYgT2YgTG9hIEFuZGVyc3NvbjxCUj4mZ3Q7IFNlbnQ6 IEZyaWRheSwgSnVseSAyMiwgMjAwNSAxMjowNSBQTTxCUj4mZ3Q7IFRvOiBQJmF1bWw7ciBNYXR0 c3NvbjxCUj4mZ3Q7IENjOiBjY2FtcEBvcHMuaWV0Zi5vcmc8QlI+Jmd0OyBTdWJqZWN0OiBSZTog RnJhbWVmb3JtYXQgaW4gYSBsMmNzIGdtcGxzIHJudmlyb25tZW50LjxCUj4mZ3Q7PEJSPiZndDsg UGVyIGFuZCBEaW1pdHJpLDxCUj4mZ3Q7PEJSPiZndDsgSSB3b3VsZCBsaWtlIHRvIGNvbWUgZG93 biBzdHJvbmdlciB0aGFuIHRoYXQsIGZvciBtZSBpdCBpczxCUj4mZ3Q7IGEgdmVyeSBzdHJvbmcg cmVxdWlyZW1lbnQgdGhhdCB0aGUgc2FtZSBzd2l0Y2ggY2FuIGhhbmRsZTxCUj4mZ3Q7IGJvdGgg VkxBTnMgYW5kIEdNUExzIHRyYWZpYyBjb3JyZWN0bHkuIEkgY2FuJ3QgZHNlZSBob3cgdGhhdDxC Uj4mZ3Q7IGNvdWxkIGJlIGRvbmUgaWYgdXNpbmcgdGhlIFZMQU4gdHBpZCB0byBpbmRpY2F0ZSBH TVBMUzxCUj4mZ3Q7IHRyYWZmaWMuPEJSPiZndDs8QlI+Jmd0OyAvTG9hPEJSPiZndDs8QlI+Jmd0 OyBQJmF1bWw7ciBNYXR0c3NvbiB3cm90ZTo8QlI+Jmd0OyAmZ3Q7Jmd0O2hpIHBhciwgb25lIG9m IHRoZSBwb3NzaWJpbGl0aWVzIHRoYXQgaGFzIGJlZW4gY29uc2lkZXJlZDxCUj4mZ3Q7IHRvIGNv cGUgd2l0aDxCUj4mZ3Q7ICZndDsmZ3Q7dGhpcyByZXF1aXJlbWVudCBpcyB0byB1c2UgYSBkZWRp Y2F0ZWQgVFBJRCBmb3IgdGhlIEV0aGVybmV0IGxhYmVsZWQ8QlI+Jmd0OyAmZ3Q7Jmd0O2ZyYW1l czsgdGhpcyB3b3VsZCBhbGxvdyBkaWZmZXJlbnRpYXRlZCBwcm9jZXNzaW5nIHdpdGggbm9uLWxh YmVsZWQ8QlI+Jmd0OyAmZ3Q7Jmd0O2ZyYW1lc3RoYW5rcy48QlI+Jmd0OyAmZ3Q7PEJSPiZndDsg Jmd0OzxCUj4mZ3Q7ICZndDsgVGhhdCBzZWVtcyB0byBtYWtlIG1vcmUgc2VuY2UuIElmIHRoYXQg ZnJhbWUgaXMgdG8gYmUgc2l6ZWQ8QlI+Jmd0OyBsaWtlIGEgODAyLjFxPEJSPiZndDsgJmd0OyBm cmFtZS4gVGhlcmUgaXMgbm90IHRoYXQgbXVjaCBzcGFjZSBsZWZ0IHRvIGEgbGFiZWwuIE9yIGlz PEJSPiZndDsgdGhlIGRlbWFuZCB0bzxCUj4mZ3Q7ICZndDsgdXNlIGp1bWJvZnJhbWVzID88QlI+ Jmd0OyAmZ3Q7IEhhcyB0aGVyZSBiZWVuIGFueSBkaXNjdXNzaW9uIG9uIGxhYmVsc3RhY2tpbmcs IGFuZCBtYWlubHk8QlI+Jmd0OyB3aGVyZSB0byBwbGFjZTxCUj4mZ3Q7ICZndDsgdGhlIGluZm9y bWF0aW9uPzxCUj4mZ3Q7ICZndDs8QlI+Jmd0OyAmZ3Q7IFJlZ2FyZHMuPEJSPiZndDsgJmd0OyBQ ZXI8QlI+Jmd0OyAmZ3Q7PEJSPiZndDsgJmd0OzxCUj4mZ3Q7ICZndDs8QlI+Jmd0OzxCUj4mZ3Q7 PEJSPiZndDsgLS08QlI+Jmd0OyBMb2EgQW5kZXJzc29uPEJSPiZndDs8QlI+Jmd0OyBQcmluY2lw YWwgTmV0d29ya2luZyBBcmNoaXRlY3Q8QlI+Jmd0OyBBY3JlbyBBQiAmbmJzcDsgJm5ic3A7ICZu YnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5i c3A7ICZuYnNwOyAmbmJzcDsgcGhvbmU6ICZuYnNwOys0NiA4IDYzMiA3NyAxNDxCUj4mZ3Q7IElz YWZqb3Jkc2dhdGFuIDIyICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7 ICZuYnNwOyAmbmJzcDsgJm5ic3A7bW9iaWxlOiArNDYgNzM5IDgxIDIxIDY0PEJSPiZndDsgS2lz dGEsIFN3ZWRlbiAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJz cDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ZW1haWw6ICZuYnNwO2xvYS5hbmRlcnNzb25A YWNyZW8uc2U8QlI+Jmd0OyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNw OyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7 ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsg bG9hQHBpLnNlPEJSPiZndDs8L0ZPTlQ+PC9QPg== Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 13:38:59 +0000 Message-ID: <193401c58ec2$de6863b0$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: =?iso-8859-1?Q?P=E4r_Mattsson?= <per@defero.se>, "Loa Andersson" <loa@pi.se> Cc: <ccamp@ops.ietf.org> Subject: Re: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 14:37:03 +0100 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_191C_01C58ECA.D361FF00" This is a multi-part message in MIME format. ------=_NextPart_000_191C_01C58ECA.D361FF00 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Hi, It is always refreshing how engineers jump straight to a discussion of = the solution :-) Perhaps we can assume that the framework draft is a good representation = of the problem space, and that we are ready to start discussing the = solutions. One (perhaps the only?) significant question to answer is how the frames = will be labelled. This question is one we must come to as soon as we are = confident that the requirements need to be addressed at all. As has been pointed out, there are several possibilities and to pick = from these we need to understand: 1. Do we need to support explicit label stacking? Note that this is not supported in TDM, LSC or FSC. 2. Do we need to be able to control (perhaps through an external signaling controller) existing hardware and=20 install LSPs through existing networks? 3. Do we need to support existing function simultaneous to the support of GMPLS L2 LSPs? I think that from a chair's perspective I can give some limited = guidance. A. These questions must be raised and answered in the framework I-D B. The answer to question 3 is "yes". This means that the use of labels must not significantly deplete any namespace used to support other function. C. CCAMP is chartered to look at the control of transport networks. This includes Metro, but I am unsure about Campus. It does not cover=20 signaling to the desktop. Thanks, Adrian ----- Original Message -----=20 From: "P=E4r Mattsson" <per@defero.se> To: "Loa Andersson" <loa@pi.se> Cc: <ccamp@ops.ietf.org> Sent: Friday, July 22, 2005 11:41 AM Subject: Re: Frameformat in a l2cs gmpls rnvironment. > > Per and Dimitri, > > > > I would like to come down stronger than that, for me it is > > a very strong requirement that the same switch can handle > > both VLANs and GMPLs trafic correctly. I can't dsee how that > > could be done if using the VLAN tpid to indicate GMPLS > > traffic. >=20 > If you ever want that same switch to handle traffic for a directly > connected host (not to uncommen) you would want that to use normal > ethernet macaddress switching. So of course you do not want to have to > choose between vlan and gmpls, you would want both at the same time. >=20 > /per >=20 >=20 > > > > /Loa > > > > P=E4r Mattsson wrote: > >>>hi par, one of the possibilities that has been considered to cope = with > >>>this requirement is to use a dedicated TPID for the Ethernet = labeled > >>>frames; this would allow differentiated processing with non-labeled > >>>framesthanks. > >> > >> > >> That seems to make more sence. If that frame is to be sized like a > >> 802.1q > >> frame. There is not that much space left to a label. Or is the = demand to > >> use jumboframes ? > >> Has there been any discussion on labelstacking, and mainly where to > >> place > >> the information? > >> > >> Regards. > >> Per > >> > >> > >> > > > > > > -- > > Loa Andersson > > > > Principal Networking Architect > > Acreo AB phone: +46 8 632 77 14 > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > Kista, Sweden email: loa.andersson@acreo.se > > loa@pi.se > > >=20 >=20 >=20 > ------=_NextPart_000_191C_01C58ECA.D361FF00 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META http-equiv=3DContent-Type content=3D"text/html; = charset=3Diso-8859-1"> <META content=3D"MSHTML 6.00.2800.1106" name=3DGENERATOR> <STYLE></STYLE> </HEAD> <BODY> <DIV><FONT face=3DCourier size=3D2>Hi,</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>It is always refreshing how engineers = jump=20 straight to a discussion of the solution :-)</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Perhaps we can assume that the = framework draft is=20 a good representation of the problem space, and that we are ready to = start=20 discussing the solutions.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>One (perhaps the only?) significant = question to=20 answer is how the frames will be labelled. This question is one we must = come to=20 as soon as we are confident that the requirements need to be addressed = at=20 all.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>As has been pointed out, there are = several=20 possibilities and to pick from these we need to understand:</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>1. Do we need to support explicit = label=20 stacking?</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> Note that this is not = supported in=20 TDM, LSC or FSC.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>2. Do we need to be able to control = (perhaps=20 through an</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> external signaling = controller)=20 existing hardware and </FONT></DIV> <DIV><FONT face=3DCourier size=3D2> install LSPs through = existing=20 networks?</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>3. Do we need to support existing = function=20 simultaneous to</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> the support of GMPLS L2=20 LSPs?</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>I think that from a chair's = perspective I can=20 give some limited guidance.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>A. These questions must be raised and = answered in=20 the framework I-D</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>B. The answer to question 3 is "yes". = This means=20 that the use of labels</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> must not significantly = deplete any=20 namespace used to support other</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> function.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>C. CCAMP is chartered to look at the = control of=20 transport networks. This</FONT></DIV> <DIV><FONT face=3DCourier size=3D2> includes Metro, but = I am unsure=20 about Campus. It does not cover </FONT></DIV> <DIV><FONT face=3DCourier size=3D2> signaling to the=20 desktop.</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>Thanks,</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Adrian</FONT></DIV> <DIV><FONT face=3DCourier size=3D2></FONT> </DIV> <DIV><FONT face=3DCourier size=3D2>----- Original Message ----- </FONT> <DIV><FONT face=3DCourier size=3D2>From: "P=E4r Mattsson" <</FONT><A=20 href=3D"mailto:per@defero.se"><FONT face=3DCourier=20 size=3D2>per@defero.se</FONT></A><FONT face=3DCourier = size=3D2>></FONT></DIV> <DIV><FONT face=3DCourier size=3D2>To: "Loa Andersson" <</FONT><A=20 href=3D"mailto:loa@pi.se"><FONT face=3DCourier = size=3D2>loa@pi.se</FONT></A><FONT=20 face=3DCourier size=3D2>></FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Cc: <</FONT><A=20 href=3D"mailto:ccamp@ops.ietf.org"><FONT face=3DCourier=20 size=3D2>ccamp@ops.ietf.org</FONT></A><FONT face=3DCourier = size=3D2>></FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Sent: Friday, July 22, 2005 11:41 = AM</FONT></DIV> <DIV><FONT face=3DCourier size=3D2>Subject: Re: Frameformat in a l2cs = gmpls=20 rnvironment.</FONT></DIV></DIV> <DIV><FONT face=3DCourier><BR><FONT size=3D2></FONT></FONT></DIV><FONT = face=3DCourier=20 size=3D2>> > Per and Dimitri,<BR>> ><BR>> > I would = like to come=20 down stronger than that, for me it is<BR>> > a very strong = requirement=20 that the same switch can handle<BR>> > both VLANs and GMPLs trafic = correctly. I can't dsee how that<BR>> > could be done if using the = VLAN=20 tpid to indicate GMPLS<BR>> > traffic.<BR>> <BR>> If you = ever want=20 that same switch to handle traffic for a directly<BR>> = connected host=20 (not to uncommen) you would want that to use normal<BR>> ethernet = macaddress=20 switching. So of course you do not want to have to<BR>> choose = between vlan=20 and gmpls, you would want both at the same time.<BR>> <BR>> = /per<BR>>=20 <BR>> <BR>> ><BR>> > /Loa<BR>> ><BR>> > P=E4r = Mattsson=20 wrote:<BR>> >>>hi par, one of the possibilities that has = been=20 considered to cope with<BR>> >>>this requirement is to use a = dedicated TPID for the Ethernet labeled<BR>> >>>frames; this = would=20 allow differentiated processing with non-labeled<BR>>=20 >>>framesthanks.<BR>> >><BR>> >><BR>> = >>=20 That seems to make more sence. If that frame is to be sized like = a<BR>>=20 >> 802.1q<BR>> >> frame. There is not that much space = left to a=20 label. Or is the demand to<BR>> >> use jumboframes ?<BR>> = >>=20 Has there been any discussion on labelstacking, and mainly where = to<BR>>=20 >> place<BR>> >> the information?<BR>> = >><BR>>=20 >> Regards.<BR>> >> Per<BR>> >><BR>> = >><BR>>=20 >><BR>> ><BR>> ><BR>> > --<BR>> > Loa=20 Andersson<BR>> ><BR>> > Principal Networking = Architect<BR>> >=20 Acreo=20 AB  = ; = =20 phone: +46 8 632 77 14<BR>> > Isafjordsgatan=20 22  = ; =20 mobile: +46 739 81 21 64<BR>> > Kista,=20 Sweden &= nbsp; =20 email: </FONT><A href=3D"mailto:loa.andersson@acreo.se"><FONT = face=3DCourier=20 size=3D2>loa.andersson@acreo.se</FONT></A><BR><FONT face=3DCourier = size=3D2>>=20 > &nb= sp; &nbs= p;  = ; =20 </FONT><A href=3D"mailto:loa@pi.se"><FONT face=3DCourier=20 size=3D2>loa@pi.se</FONT></A><BR><FONT face=3DCourier size=3D2>> = ><BR>>=20 <BR>> <BR>> <BR>> </FONT></BODY></HTML> ------=_NextPart_000_191C_01C58ECA.D361FF00-- Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 13:24:42 +0000 Message-ID: <FF8AC5030873D6118BCB0002A58EDA990B9D2947@mchh2a7e.mchh.siemens.de> From: Heiles Juergen <juergen.heiles@siemens.com> To: Loa Andersson <loa@pi.se>, richard.spencer@bt.com Cc: per@defero.se, ccamp@ops.ietf.org Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 15:23:43 +0200 MIME-Version: 1.0 Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: quoted-printable Loa, I interpret the ID as a proposal to use GMPLS for VLAN setup. So GMPLS = and VLAN to not compete. The VLAN is at the data plane and GMPLS at the = control plane. The question is how and should different control plane = techniques like GMPLS and (GVRP and STB) work together? Regards Juergen =20 > -----Original Message----- > From: owner-ccamp@ops.ietf.org=20 > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson > Sent: Friday, July 22, 2005 2:53 PM > To: richard.spencer@bt.com > Cc: per@defero.se; ccamp@ops.ietf.org > Subject: Re: Frameformat in a l2cs gmpls rnvironment. >=20 > Richard, >=20 > I agree to most of this. Since we have doubts about the viability > of taking GMPLS all the way to end-user or enterprise I think it > would be good, from a wg perspective, if we agreed to solve the > core network problems first. >=20 > Do you have any comment on the requirement to run both VLANs and > GMPLS on the same switch? >=20 > /Loa >=20 > richard.spencer@bt.com wrote: > > Regardless of whether or not a switch is directly connected > > to hosts, it must be able to forward packets using the=20 > connectionless > > Ethernet data plane. This is due to the fundamental=20 > requirement that > > for GMPLS switches to be able to exchange control information with > > each other, a data plane for control traffic must be present. > > This is akin to using the IP data plane for MPLS signalling in an > > IP/MPLS network. An alternative would be to use a static reserved > > L2-LSP for control traffic in the same way that reserved VPI/VCIs > > are used for PNNI signalling in ATM. > >=20 > > Regarding connecting hosts to GMPLS switches, I personally don't=20 > > think extending L2-LSPs into the enterprise/home network is > > commercially viable. However, if you do want to use GMPLS switches > > in the home/enterprise network and for some reason don't want to > > extend L2-LSPs down to the host then you will not be performing > > normal Ethernet Mac address switching anyway. Instead you will > > need some kind of policy on the switch that maps connectionless > > Ethernet packets (e.g. based on MAC src/dest, 802.1p, VLAN) to a > > L2-LSP. This is because multiple L2-LSPs to the same destination > > (e.g. a gateway router) may exist for different services/flows > > (e.g. video download, VoIP call, etc.). > >=20 > > Regards, > > Richard > >=20 > >=20 > >>-----Original Message----- > >>From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On > >>Behalf Of P=E4r Mattsson > >>Sent: 22 July 2005 11:42 > >>To: Loa Andersson > >>Cc: ccamp@ops.ietf.org > >>Subject: Re: Frameformat in a l2cs gmpls rnvironment. > >> > >> > >> > >>>Per and Dimitri, > >>> > >>>I would like to come down stronger than that, for me it is > >>>a very strong requirement that the same switch can handle > >>>both VLANs and GMPLs trafic correctly. I can't dsee how that > >>>could be done if using the VLAN tpid to indicate GMPLS > >>>traffic. > >> > >>If you ever want that same switch to handle traffic for a directly > >>connected host (not to uncommen) you would want that to use normal > >>ethernet macaddress switching. So of course you do not want=20 > to have to > >>choose between vlan and gmpls, you would want both at the same = time. > >> > >>/per > >> > >> > >> > >>>/Loa > >>> > >>>P=E4r Mattsson wrote: > >>> > >>>>>hi par, one of the possibilities that has been considered=20 > >> > >>to cope with > >> > >>>>>this requirement is to use a dedicated TPID for the=20 > >> > >>Ethernet labeled > >> > >>>>>frames; this would allow differentiated processing with=20 > non-labeled > >>>>>framesthanks. > >>>> > >>>> > >>>>That seems to make more sence. If that frame is to be sized like = a > >>>>802.1q > >>>>frame. There is not that much space left to a label. Or is > >> > >>the demand to > >> > >>>>use jumboframes ? > >>>>Has there been any discussion on labelstacking, and=20 > mainly where to > >>>>place > >>>>the information? > >>>> > >>>>Regards. > >>>>Per > >>>> > >>>> > >>>> > >>> > >>> > >>>-- > >>>Loa Andersson > >>> > >>>Principal Networking Architect > >>>Acreo AB phone: +46 8 632 77 14 > >>>Isafjordsgatan 22 mobile: +46 739 81 21 64 > >>>Kista, Sweden email: loa.andersson@acreo.se > >>> loa@pi.se > >>> > >> > >> > >> > >=20 >=20 >=20 > --=20 > Loa Andersson >=20 > Principal Networking Architect > Acreo AB phone: +46 8 632 77 14 > Isafjordsgatan 22 mobile: +46 739 81 21 64 > Kista, Sweden email: loa.andersson@acreo.se > loa@pi.se >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 13:24:34 +0000 To: Loa Andersson <loa@pi.se> Cc: =?iso-8859-1?Q?P=E4r_Mattsson?= <per@defero.se>, ccamp@ops.ietf.org MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: Re: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 15:24:10 +0200 Message-ID: <OFDD93A200.6D8D14C9-ONC1257046.00499F58-C1257046.00499FDC@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmxvYSwgPC9GT05UPjwvUD48UD48Rk9O VCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+ZHVyaW5nIHRoZSBkaXNjdXNzaW9uIGxvb3Agb25l IG9mIHRoZSBwb3NzaWJpbGl0aWVzIHRoYXQgY2FtZSBvdXQgY29uc2lzdGVkIGluIHVzaW5nIGEg ZGVkaWNhdGVkIFRQSUQgdmFsdWUgKGRpZmZlcmVudCBmcm9tIHRoZSB2YWx1ZSAweDgxMDAgdXNl ZCBpbiAuMXEpLCB0aGF0IHdvdWxkIHByZWNlZGUgdGhlIGVuY29kZWQgdmFsdWUgYXMgcGFydCBv ZiB0aGUgVENJIDwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPndv dWxkIHlvdSBwbGVhc2UgY2xhcmlmeSB5b3VyIHNwZWNpZmljIGNvbmNlcm4gPyBpIHdvdWxkIGxp a2UgYWxzbyB0byBwb2ludCBvdXQgdGhhdCB0aGVyZSBhcmUgc2V2ZXJhbCBjYXNlcyBoZXJlPC9G T05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+MS4gYm90aCB0eXBlIG9m IHRyYWZmaWMgb24gdGhlIHNhbWUgc3dpdGNoIGJ1dCBub3QgbmVjZXNzYXJpbHkgb24gdGhlIHNh bWUgaW50ZXJmYWNlPC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+ Mi4gYm90aCB0eXBlIG9mIHRyYWZmaWMgb24gdGhlIHNhbWUgaW50ZXJmYWNlIDwvRk9OVD48L1A+ PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPndpdGggdGhlIGZvbGxvd2luZyB2YXJp YW50PC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+MmEuIGludGVy ZmFjZSBtdXN0IGFjY29tbW9kYXRlIGFueSB0eXBlIG9mIHRyYWZmaWMgKHVuLS90YWdnZWQgYW5k IGxhYmVsZWQpPC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+MmIu IGludGVyZmFjZSBtdXN0IGFjY29tbW9kYXRlIHRhZ2dlZCBhbmQgbGFiZWxlZCB0eXBlIG9mIHRy YWZmaWMgb25seTwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPnRo YW5rcyw8L0ZPTlQ+PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj4tIGRpbWl0 cmkuPC9GT05UPjxCUj48QlI+PEZPTlQgU0laRT0yPjxCPkxvYSBBbmRlcnNzb24gJmx0O2xvYUBw aS5zZSZndDs8L0I+PC9GT05UPjxCUj48Rk9OVCBTSVpFPTI+U2VudCBieTogb3duZXItY2NhbXBA b3BzLmlldGYub3JnPC9GT05UPjxCUj48Rk9OVCBTSVpFPTI+MDcvMjIvMjAwNSAxMjowNCBaRTI8 L0ZPTlQ+PEJSPjxCUj4gPEZPTlQgU0laRT0yPlRvOjwvRk9OVD4gPEZPTlQgU0laRT0yPlAmYXVt bDtyIE1hdHRzc29uICZsdDtwZXJAZGVmZXJvLnNlJmd0OzwvRk9OVD48QlI+IDxGT05UIFNJWkU9 Mj5jYzo8L0ZPTlQ+IDxGT05UIFNJWkU9Mj5jY2FtcEBvcHMuaWV0Zi5vcmc8L0ZPTlQ+PEJSPiA8 Rk9OVCBTSVpFPTI+YmNjOjwvRk9OVD4gPEJSPiA8Rk9OVCBTSVpFPTI+U3ViamVjdDo8L0ZPTlQ+ IDxGT05UIFNJWkU9Mj5SZTogRnJhbWVmb3JtYXQgaW4gYSBsMmNzIGdtcGxzIHJudmlyb25tZW50 LjwvRk9OVD48QlI+IDxCUj48QlI+PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVy Ij5QZXIgYW5kIERpbWl0cmksPEJSPjwvRk9OVD48QlI+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENv dXJpZXIiPkkgd291bGQgbGlrZSB0byBjb21lIGRvd24gc3Ryb25nZXIgdGhhbiB0aGF0LCBmb3Ig bWUgaXQgaXM8QlI+YSB2ZXJ5IHN0cm9uZyByZXF1aXJlbWVudCB0aGF0IHRoZSBzYW1lIHN3aXRj aCBjYW4gaGFuZGxlPEJSPmJvdGggVkxBTnMgYW5kIEdNUExzIHRyYWZpYyBjb3JyZWN0bHkuIEkg Y2FuJ3QgZHNlZSBob3cgdGhhdDxCUj5jb3VsZCBiZSBkb25lIGlmIHVzaW5nIHRoZSBWTEFOIHRw aWQgdG8gaW5kaWNhdGUgR01QTFM8QlI+dHJhZmZpYy48QlI+PC9GT05UPjxCUj48Rk9OVCBGQUNF PSJNb25vc3BhY2UsQ291cmllciI+L0xvYTxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9z cGFjZSxDb3VyaWVyIj5QJmF1bWw7ciBNYXR0c3NvbiB3cm90ZTo8QlI+Jmd0OyZndDtoaSBwYXIs IG9uZSBvZiB0aGUgcG9zc2liaWxpdGllcyB0aGF0IGhhcyBiZWVuIGNvbnNpZGVyZWQgdG8gY29w ZSB3aXRoPEJSPiZndDsmZ3Q7dGhpcyByZXF1aXJlbWVudCBpcyB0byB1c2UgYSBkZWRpY2F0ZWQg VFBJRCBmb3IgdGhlIEV0aGVybmV0IGxhYmVsZWQ8QlI+Jmd0OyZndDtmcmFtZXM7IHRoaXMgd291 bGQgYWxsb3cgZGlmZmVyZW50aWF0ZWQgcHJvY2Vzc2luZyB3aXRoIG5vbi1sYWJlbGVkPEJSPiZn dDsmZ3Q7ZnJhbWVzdGhhbmtzLjxCUj4mZ3Q7PEJSPiZndDs8QlI+Jmd0OyBUaGF0IHNlZW1zIHRv IG1ha2UgbW9yZSBzZW5jZS4gSWYgdGhhdCBmcmFtZSBpcyB0byBiZSBzaXplZCBsaWtlIGEgODAy LjFxPEJSPiZndDsgZnJhbWUuIFRoZXJlIGlzIG5vdCB0aGF0IG11Y2ggc3BhY2UgbGVmdCB0byBh IGxhYmVsLiBPciBpcyB0aGUgZGVtYW5kIHRvPEJSPiZndDsgdXNlIGp1bWJvZnJhbWVzID88QlI+ Jmd0OyBIYXMgdGhlcmUgYmVlbiBhbnkgZGlzY3Vzc2lvbiBvbiBsYWJlbHN0YWNraW5nLCBhbmQg bWFpbmx5IHdoZXJlIHRvIHBsYWNlPEJSPiZndDsgdGhlIGluZm9ybWF0aW9uPzxCUj4mZ3Q7PEJS PiZndDsgUmVnYXJkcy48QlI+Jmd0OyBQZXI8QlI+Jmd0OzxCUj4mZ3Q7PEJSPiZndDs8QlI+PC9G T05UPjxCUj48QlI+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPi0tPEJSPkxvYSBBbmRl cnNzb248QlI+PC9GT05UPjxCUj48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+UHJpbmNp cGFsIE5ldHdvcmtpbmcgQXJjaGl0ZWN0PEJSPkFjcmVvIEFCICZuYnNwOyAmbmJzcDsgJm5ic3A7 ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsg Jm5ic3A7ICZuYnNwOyBwaG9uZTogJm5ic3A7KzQ2IDggNjMyIDc3IDE0PEJSPklzYWZqb3Jkc2dh dGFuIDIyICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAm bmJzcDsgJm5ic3A7bW9iaWxlOiArNDYgNzM5IDgxIDIxIDY0PEJSPktpc3RhLCBTd2VkZW4gJm5i c3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJzcDsgJm5ic3A7ICZuYnNwOyAmbmJz cDsgJm5ic3A7ICZuYnNwO2VtYWlsOiAmbmJzcDtsb2EuYW5kZXJzc29uQGFjcmVvLnNlPC9GT05U PjwvUD48VUw+PFVMPjxVTD48VUw+PFVMPjxVTD48VUw+PFVMPjxVTD48VUw+PFVMPjxVTD48VUw+ PFVMPjxVTD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+bG9hQHBpLnNlPEJSPjwvRk9O VD48L1VMPg== Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 13:18:21 +0000 Message-ID: <FF8AC5030873D6118BCB0002A58EDA990B9D2941@mchh2a7e.mchh.siemens.de> From: Heiles Juergen <juergen.heiles@siemens.com> To: Loa Andersson <loa@pi.se>, =?iso-8859-1?Q?P=E4r_Mattsson?= <per@defero.se> Cc: ccamp@ops.ietf.org Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 15:17:19 +0200 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable GMPLS for SONET/SDH has no impact on the SONET/SDH data plane. It = introduces a control plane for connection setup instead of a network = management based connection setup. GMPLS for Ethernet can work in the = same way. Ethernet as such is cl, but VLANs are setup and GMPLS can = provide a control plane for VLAN setup. The Ethernet data plane is not = changed and MAC address learning and forwarding is still be done within = a VLAN. So no change to the Ethernet data plane. No GMPLS traffic is = introduced, it is still VLAN traffic. Introducing a dedicated GMPLS label for Ethernet (GMPLS traffic) goes = beyond GMPLS as a control plane technique. Regards Juergen =20 > -----Original Message----- > From: owner-ccamp@ops.ietf.org=20 > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Loa Andersson > Sent: Friday, July 22, 2005 12:05 PM > To: P=E4r Mattsson > Cc: ccamp@ops.ietf.org > Subject: Re: Frameformat in a l2cs gmpls rnvironment. >=20 > Per and Dimitri, >=20 > I would like to come down stronger than that, for me it is > a very strong requirement that the same switch can handle > both VLANs and GMPLs trafic correctly. I can't dsee how that > could be done if using the VLAN tpid to indicate GMPLS > traffic. >=20 > /Loa >=20 > P=E4r Mattsson wrote: > >>hi par, one of the possibilities that has been considered=20 > to cope with > >>this requirement is to use a dedicated TPID for the Ethernet = labeled > >>frames; this would allow differentiated processing with non-labeled > >>framesthanks. > >=20 > >=20 > > That seems to make more sence. If that frame is to be sized=20 > like a 802.1q > > frame. There is not that much space left to a label. Or is=20 > the demand to > > use jumboframes ? > > Has there been any discussion on labelstacking, and mainly=20 > where to place > > the information? > >=20 > > Regards. > > Per > >=20 > >=20 > >=20 >=20 >=20 > --=20 > Loa Andersson >=20 > Principal Networking Architect > Acreo AB phone: +46 8 632 77 14 > Isafjordsgatan 22 mobile: +46 739 81 21 64 > Kista, Sweden email: loa.andersson@acreo.se > loa@pi.se >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 12:57:45 +0000 Message-ID: <42E0EC48.2010005@pi.se> Date: Fri, 22 Jul 2005 14:53:28 +0200 From: Loa Andersson <loa@pi.se> Organization: Acreo AB User-Agent: Mozilla Thunderbird 1.0.5 (Windows/20050711) MIME-Version: 1.0 To: richard.spencer@bt.com CC: per@defero.se, ccamp@ops.ietf.org Subject: Re: Frameformat in a l2cs gmpls rnvironment. Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 8bit Richard, I agree to most of this. Since we have doubts about the viability of taking GMPLS all the way to end-user or enterprise I think it would be good, from a wg perspective, if we agreed to solve the core network problems first. Do you have any comment on the requirement to run both VLANs and GMPLS on the same switch? /Loa richard.spencer@bt.com wrote: > Regardless of whether or not a switch is directly connected > to hosts, it must be able to forward packets using the connectionless > Ethernet data plane. This is due to the fundamental requirement that > for GMPLS switches to be able to exchange control information with > each other, a data plane for control traffic must be present. > This is akin to using the IP data plane for MPLS signalling in an > IP/MPLS network. An alternative would be to use a static reserved > L2-LSP for control traffic in the same way that reserved VPI/VCIs > are used for PNNI signalling in ATM. > > Regarding connecting hosts to GMPLS switches, I personally don't > think extending L2-LSPs into the enterprise/home network is > commercially viable. However, if you do want to use GMPLS switches > in the home/enterprise network and for some reason don't want to > extend L2-LSPs down to the host then you will not be performing > normal Ethernet Mac address switching anyway. Instead you will > need some kind of policy on the switch that maps connectionless > Ethernet packets (e.g. based on MAC src/dest, 802.1p, VLAN) to a > L2-LSP. This is because multiple L2-LSPs to the same destination > (e.g. a gateway router) may exist for different services/flows > (e.g. video download, VoIP call, etc.). > > Regards, > Richard > > >>-----Original Message----- >>From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On >>Behalf Of Pär Mattsson >>Sent: 22 July 2005 11:42 >>To: Loa Andersson >>Cc: ccamp@ops.ietf.org >>Subject: Re: Frameformat in a l2cs gmpls rnvironment. >> >> >> >>>Per and Dimitri, >>> >>>I would like to come down stronger than that, for me it is >>>a very strong requirement that the same switch can handle >>>both VLANs and GMPLs trafic correctly. I can't dsee how that >>>could be done if using the VLAN tpid to indicate GMPLS >>>traffic. >> >>If you ever want that same switch to handle traffic for a directly >>connected host (not to uncommen) you would want that to use normal >>ethernet macaddress switching. So of course you do not want to have to >>choose between vlan and gmpls, you would want both at the same time. >> >>/per >> >> >> >>>/Loa >>> >>>Pär Mattsson wrote: >>> >>>>>hi par, one of the possibilities that has been considered >> >>to cope with >> >>>>>this requirement is to use a dedicated TPID for the >> >>Ethernet labeled >> >>>>>frames; this would allow differentiated processing with non-labeled >>>>>framesthanks. >>>> >>>> >>>>That seems to make more sence. If that frame is to be sized like a >>>>802.1q >>>>frame. There is not that much space left to a label. Or is >> >>the demand to >> >>>>use jumboframes ? >>>>Has there been any discussion on labelstacking, and mainly where to >>>>place >>>>the information? >>>> >>>>Regards. >>>>Per >>>> >>>> >>>> >>> >>> >>>-- >>>Loa Andersson >>> >>>Principal Networking Architect >>>Acreo AB phone: +46 8 632 77 14 >>>Isafjordsgatan 22 mobile: +46 739 81 21 64 >>>Kista, Sweden email: loa.andersson@acreo.se >>> loa@pi.se >>> >> >> >> > -- Loa Andersson Principal Networking Architect Acreo AB phone: +46 8 632 77 14 Isafjordsgatan 22 mobile: +46 739 81 21 64 Kista, Sweden email: loa.andersson@acreo.se loa@pi.se Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 11:46:19 +0000 Content-Class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 12:44:08 +0100 Message-ID: <B5E87B043D4C514389141E2661D255EC0A835AE2@i2km41-ukdy.domain1.systemhost.net> Thread-Topic: Frameformat in a l2cs gmpls rnvironment. Thread-Index: AcWOqjL1VwqIq2haR4mlmgcOF+B70gAA1pkg From: <richard.spencer@bt.com> To: <per@defero.se>, <loa@pi.se> Cc: <ccamp@ops.ietf.org> Regardless of whether or not a switch is directly connected to hosts, it = must be able to forward packets using the connectionless Ethernet data = plane. This is due to the fundamental requirement that for GMPLS = switches to be able to exchange control information with each other, a = data plane for control traffic must be present. This is akin to using = the IP data plane for MPLS signalling in an IP/MPLS network. An = alternative would be to use a static reserved L2-LSP for control traffic = in the same way that reserved VPI/VCIs are used for PNNI signalling in = ATM. Regarding connecting hosts to GMPLS switches, I personally don't think = extending L2-LSPs into the enterprise/home network is commercially = viable. However, if you do want to use GMPLS switches in the = home/enterprise network and for some reason don't want to extend L2-LSPs = down to the host then you will not be performing normal Ethernet Mac = address switching anyway. Instead you will need some kind of policy on = the switch that maps connectionless Ethernet packets (e.g. based on MAC = src/dest, 802.1p, VLAN) to a L2-LSP. This is because multiple L2-LSPs to = the same destination (e.g. a gateway router) may exist for different = services/flows (e.g. video download, VoIP call, etc.). Regards, Richard > -----Original Message----- > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On > Behalf Of P=E4r Mattsson > Sent: 22 July 2005 11:42 > To: Loa Andersson > Cc: ccamp@ops.ietf.org > Subject: Re: Frameformat in a l2cs gmpls rnvironment. >=20 >=20 > > Per and Dimitri, > > > > I would like to come down stronger than that, for me it is > > a very strong requirement that the same switch can handle > > both VLANs and GMPLs trafic correctly. I can't dsee how that > > could be done if using the VLAN tpid to indicate GMPLS > > traffic. >=20 > If you ever want that same switch to handle traffic for a directly > connected host (not to uncommen) you would want that to use normal > ethernet macaddress switching. So of course you do not want to have to > choose between vlan and gmpls, you would want both at the same time. >=20 > /per >=20 >=20 > > > > /Loa > > > > P=E4r Mattsson wrote: > >>>hi par, one of the possibilities that has been considered=20 > to cope with > >>>this requirement is to use a dedicated TPID for the=20 > Ethernet labeled > >>>frames; this would allow differentiated processing with non-labeled > >>>framesthanks. > >> > >> > >> That seems to make more sence. If that frame is to be sized like a > >> 802.1q > >> frame. There is not that much space left to a label. Or is=20 > the demand to > >> use jumboframes ? > >> Has there been any discussion on labelstacking, and mainly where to > >> place > >> the information? > >> > >> Regards. > >> Per > >> > >> > >> > > > > > > -- > > Loa Andersson > > > > Principal Networking Architect > > Acreo AB phone: +46 8 632 77 14 > > Isafjordsgatan 22 mobile: +46 739 81 21 64 > > Kista, Sweden email: loa.andersson@acreo.se > > loa@pi.se > > >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 10:43:18 +0000 Message-ID: <36243.212.247.39.209.1122028915.squirrel@webmail.defero.se> Date: Fri, 22 Jul 2005 12:41:55 +0200 (CEST) Subject: Re: Frameformat in a l2cs gmpls rnvironment. From: P=?iso-8859-1?Q?=E4r_Mattsson?= <per@defero.se> To: "Loa Andersson" <loa@pi.se> Cc: ccamp@ops.ietf.org User-Agent: SquirrelMail/1.4.0 MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-1 > Per and Dimitri, > > I would like to come down stronger than that, for me it is > a very strong requirement that the same switch can handle > both VLANs and GMPLs trafic correctly. I can't dsee how that > could be done if using the VLAN tpid to indicate GMPLS > traffic. If you ever want that same switch to handle traffic for a directly connected host (not to uncommen) you would want that to use normal ethernet macaddress switching. So of course you do not want to have to choose between vlan and gmpls, you would want both at the same time. /per > > /Loa > > Pär Mattsson wrote: >>>hi par, one of the possibilities that has been considered to cope with >>>this requirement is to use a dedicated TPID for the Ethernet labeled >>>frames; this would allow differentiated processing with non-labeled >>>framesthanks. >> >> >> That seems to make more sence. If that frame is to be sized like a >> 802.1q >> frame. There is not that much space left to a label. Or is the demand to >> use jumboframes ? >> Has there been any discussion on labelstacking, and mainly where to >> place >> the information? >> >> Regards. >> Per >> >> >> > > > -- > Loa Andersson > > Principal Networking Architect > Acreo AB phone: +46 8 632 77 14 > Isafjordsgatan 22 mobile: +46 739 81 21 64 > Kista, Sweden email: loa.andersson@acreo.se > loa@pi.se > Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 10:28:24 +0000 To: =?iso-8859-1?Q?P=E4r_Mattsson?= <per@defero.se> Cc: ccamp@ops.ietf.org MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: Re:Re: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 12:27:37 +0200 Message-ID: <OFFBE0C522.71FFD430-ONC1257046.0039758A-C1257046.00397624@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+Jm5ic3A7PEJSPjxCUj48Rk9OVCBTSVpFPTI+PEI+UCZhdW1sO3IgTWF0dHNzb24gJmx0O3Bl ckBkZWZlcm8uc2UmZ3Q7PC9CPjwvRk9OVD48QlI+PEZPTlQgU0laRT0yPlNlbnQgYnk6IG93bmVy LWNjYW1wQG9wcy5pZXRmLm9yZzwvRk9OVD48QlI+PEZPTlQgU0laRT0yPjA3LzIyLzIwMDUgMTE6 NTQgWkUyPC9GT05UPjxCUj48QlI+IDxGT05UIFNJWkU9Mj5Ubzo8L0ZPTlQ+IDxGT05UIFNJWkU9 Mj5jY2FtcEBvcHMuaWV0Zi5vcmc8L0ZPTlQ+PEJSPiA8Rk9OVCBTSVpFPTI+Y2M6PC9GT05UPiA8 QlI+IDxGT05UIFNJWkU9Mj5iY2M6PC9GT05UPiA8QlI+IDxGT05UIFNJWkU9Mj5TdWJqZWN0Ojwv Rk9OVD4gPEZPTlQgU0laRT0yPlJlOlJlOiBGcmFtZWZvcm1hdCBpbiBhIGwyY3MgZ21wbHMgcm52 aXJvbm1lbnQuPC9GT05UPjxCUj4gPEJSPjxCUj48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNl LENvdXJpZXIiPiZndDsmZ3Q7IGhpIHBhciwgb25lIG9mIHRoZSBwb3NzaWJpbGl0aWVzIHRoYXQg aGFzIGJlZW4gY29uc2lkZXJlZCB0byBjb3BlIHdpdGg8QlI+Jmd0OyZndDsgdGhpcyByZXF1aXJl bWVudCBpcyB0byB1c2UgYSBkZWRpY2F0ZWQgVFBJRCBmb3IgdGhlIEV0aGVybmV0IGxhYmVsZWQ8 QlI+Jmd0OyZndDsgZnJhbWVzOyB0aGlzIHdvdWxkIGFsbG93IGRpZmZlcmVudGlhdGVkIHByb2Nl c3Npbmcgd2l0aCBub24tbGFiZWxlZDxCUj4mZ3Q7Jmd0OyBmcmFtZXN0aGFua3MuPEJSPjwvRk9O VD48QlI+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPiZndDsgVGhhdCBzZWVtcyB0byBt YWtlIG1vcmUgc2VuY2UuIElmIHRoYXQgZnJhbWUgaXMgdG8gYmUgc2l6ZWQgbGlrZSBhIDgwMi4x cTxCUj4mZ3Q7IGZyYW1lLiBUaGVyZSBpcyBub3QgdGhhdCBtdWNoIHNwYWNlIGxlZnQgdG8gYSBs YWJlbC4gT3IgaXMgdGhlIGRlbWFuZCB0bzxCUj4mZ3Q7IHVzZSBqdW1ib2ZyYW1lcyA/PC9GT05U PjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+Jmd0OyBIYXMgdGhlcmUgYmVl biBhbnkgZGlzY3Vzc2lvbiBvbiBsYWJlbHN0YWNraW5nLCBhbmQgbWFpbmx5IHdoZXJlIHRvIHBs YWNlPEJSPiZndDsgdGhlIGluZm9ybWF0aW9uPzxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1v bm9zcGFjZSxDb3VyaWVyIj48QlI+dGhlcmUgYXJlIHNldmVyYWwgcG9zc2liaWxpdGllcyBoZXJl IC0gb25lIG9mIHRoZW0gY29uc2lzdHMgaW4gZW5jb2RpbmcgdGhlIGxhYmVsIHZhbHVlIGFzIHBh cnQgb2YgdGhlIFRDSSAtPC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmll ciI+YWxzbyB0aGUgY3VycmVudCBkb2N1bWVudCBkb2VzIG5vdCBkaXNjdXNzIGFueSBzcGVjaWZp YyAmcXVvdDtzdGFja2luZyZxdW90OyBvZiBzdWNoIHZhbHVlcyAod2Ugc2hvdWxkIHB1dCB0aGlz IG9uIHRoZSBvcGVuIGRpc2N1c3Npb24gcG9pbnQpPC9GT05UPjwvUD4= Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 10:25:22 +0000 Message-ID: <42E0C4C5.6050501@pi.se> Date: Fri, 22 Jul 2005 12:04:53 +0200 From: Loa Andersson <loa@pi.se> Organization: Acreo AB User-Agent: Mozilla Thunderbird 1.0.5 (Windows/20050711) MIME-Version: 1.0 To: =?ISO-8859-1?Q?P=E4r_Mattsson?= <per@defero.se> CC: ccamp@ops.ietf.org Subject: Re: Frameformat in a l2cs gmpls rnvironment. Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 8bit Per and Dimitri, I would like to come down stronger than that, for me it is a very strong requirement that the same switch can handle both VLANs and GMPLs trafic correctly. I can't dsee how that could be done if using the VLAN tpid to indicate GMPLS traffic. /Loa Pär Mattsson wrote: >>hi par, one of the possibilities that has been considered to cope with >>this requirement is to use a dedicated TPID for the Ethernet labeled >>frames; this would allow differentiated processing with non-labeled >>framesthanks. > > > That seems to make more sence. If that frame is to be sized like a 802.1q > frame. There is not that much space left to a label. Or is the demand to > use jumboframes ? > Has there been any discussion on labelstacking, and mainly where to place > the information? > > Regards. > Per > > > -- Loa Andersson Principal Networking Architect Acreo AB phone: +46 8 632 77 14 Isafjordsgatan 22 mobile: +46 739 81 21 64 Kista, Sweden email: loa.andersson@acreo.se loa@pi.se Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 09:55:38 +0000 Message-ID: <35655.212.247.39.209.1122026080.squirrel@webmail.defero.se> Date: Fri, 22 Jul 2005 11:54:40 +0200 (CEST) Subject: Re:Re: Frameformat in a l2cs gmpls rnvironment. From: P=?iso-8859-1?Q?=E4r_Mattsson?= <per@defero.se> To: ccamp@ops.ietf.org User-Agent: SquirrelMail/1.4.0 MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-1 > hi par, one of the possibilities that has been considered to cope with > this requirement is to use a dedicated TPID for the Ethernet labeled > frames; this would allow differentiated processing with non-labeled > framesthanks. That seems to make more sence. If that frame is to be sized like a 802.1q frame. There is not that much space left to a label. Or is the demand to use jumboframes ? Has there been any discussion on labelstacking, and mainly where to place the information? Regards. Per Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 09:46:35 +0000 To: =?iso-8859-1?Q?P=E4r_Mattsson?= <per@defero.se> Cc: ccamp@ops.ietf.org MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: Re: Frameformat in a l2cs gmpls rnvironment. Date: Fri, 22 Jul 2005 11:45:34 +0200 Message-ID: <OFDFF9FE10.41F8470E-ONC1257046.00359BF2-C1257046.00359C93@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmhpIHBhciwgPC9GT05UPjwvUD48UD48 Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+b25lIG9mIHRoZSBwb3NzaWJpbGl0aWVzIHRo YXQgaGFzIGJlZW4gY29uc2lkZXJlZCB0byBjb3BlIHdpdGggdGhpcyByZXF1aXJlbWVudCBpcyB0 byB1c2UgYSBkZWRpY2F0ZWQgVFBJRCBmb3IgdGhlIEV0aGVybmV0IGxhYmVsZWQgZnJhbWVzOyB0 aGlzIHdvdWxkIGFsbG93IGRpZmZlcmVudGlhdGVkIHByb2Nlc3Npbmcgd2l0aCBub24tbGFiZWxl ZCBmcmFtZXM8L0ZPTlQ+PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj50aGFu a3MsPC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+LSBkaW1pdHJp LjwvRk9OVD48QlI+PEJSPjxGT05UIFNJWkU9Mj48Qj5QJmF1bWw7ciBNYXR0c3NvbiAmbHQ7cGVy QGRlZmVyby5zZSZndDs8L0I+PC9GT05UPjxCUj48Rk9OVCBTSVpFPTI+U2VudCBieTogb3duZXIt Y2NhbXBAb3BzLmlldGYub3JnPC9GT05UPjxCUj48Rk9OVCBTSVpFPTI+MDcvMjIvMjAwNSAxMToz NiBaRTI8L0ZPTlQ+PEJSPjxCUj4gPEZPTlQgU0laRT0yPlRvOjwvRk9OVD4gPEZPTlQgU0laRT0y PmNjYW1wQG9wcy5pZXRmLm9yZzwvRk9OVD48QlI+IDxGT05UIFNJWkU9Mj5jYzo8L0ZPTlQ+IDxC Uj4gPEZPTlQgU0laRT0yPmJjYzo8L0ZPTlQ+IDxCUj4gPEZPTlQgU0laRT0yPlN1YmplY3Q6PC9G T05UPiA8Rk9OVCBTSVpFPTI+RnJhbWVmb3JtYXQgaW4gYSBsMmNzIGdtcGxzIHJudmlyb25tZW50 LjwvRk9OVD48QlI+IDxCUj48QlI+PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVy Ij5IaSBhbGwuPEJSPkFmdGVyIHJlYWRpbmcgdGhlIGRyYWZ0PEJSPihkcmFmdC1wYXBhZGltaXRy aW91LWNjYW1wLWdtcGxzLWV0aGVybmV0LWZyYW1ld29yay0wMC50eHQpIGFuZCB0aGlua2luZzxC Uj5mb3IgYXdoaWxlIG9uIGRpZmZlcmVudCB3YXlzIHRvIGltcGxlbWVudGluZyB0aGlzLCBpIHdh cyB3b25kZXJpbmcgd2hhdDxCUj5hbHRlcm5hdGl2ZXMgdGhlcmUgaXMgZm9yIHRoZSBmcmFtZm9y bWF0IGZvciB0aGlzLiBJIHVuZGVyc3RhbmQgdGhhdCBvbmU8QlI+b3B0aW9uIGlzIHVzaW5nIFRQ SUQ9ODEwMCwgYnV0IHdoYXQgYmVjYW1lIG9mIHRoZSBsb2thbCBzaWduaWZpY2FudCBsYWJlbCw8 QlI+YW5kIHRoYXQgYWxsIHN3aXRjaGVzIHNob3VsZCBiZSBhYmxlIHRvIHN3aXRjaCBjdXJyZW50 IGV0aGVybmV0IGZvcm1hdCBhdDxCUj50aGUgc2FtZSB0aW1lLiBIb3cgdG8gZGlzdGluZ3Vpc2gg YmV0d2VlbiBHTVBMUyBhbmQgVkxBTiA/PEJSPkRvIGFueWJvZHkgaGF2ZSBhbnkgY29tbWVudCBv biB0aGlzID88QlI+PC9GT05UPjxCUj48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+UmVn YXJkcy48QlI+UGVyPC9GT05UPjwvUD4= Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 09:39:00 +0000 Message-ID: <35391.212.247.39.209.1122025004.squirrel@webmail.defero.se> Date: Fri, 22 Jul 2005 11:36:44 +0200 (CEST) Subject: Frameformat in a l2cs gmpls rnvironment. From: P=?iso-8859-1?Q?=E4r_Mattsson?= <per@defero.se> To: ccamp@ops.ietf.org User-Agent: SquirrelMail/1.4.0 MIME-Version: 1.0 Content-Type: text/plain;charset=iso-8859-1 Hi all. After reading the draft (draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt) and thinking for awhile on different ways to implementing this, i was wondering what alternatives there is for the framformat for this. I understand that one option is using TPID=8100, but what became of the lokal significant label, and that all switches should be able to switch current ethernet format at the same time. How to distinguish between GMPLS and VLAN ? Do anybody have any comment on this ? Regards. Per Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 04:25:02 +0000 Message-ID: <42E074A1.1060703@kddilabs.jp> Date: Fri, 22 Jul 2005 13:22:57 +0900 From: Tomohiro Otani <otani@kddilabs.jp> User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; ja-JP; rv:1.7.2) Gecko/20040803 MIME-Version: 1.0 To: Dimitri.Papadimitriou@alcatel.be Cc: ccamp@ops.ietf.org Subject: Re: http://www.ietf.org/internet-drafts/draft-otani-ccamp-gmpls-cspf-constraints-00 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi Dimitri and everyone, We uploaded 01 version of this draft. Your comments and feedback are highly appreciated. With best regards, tomo ----------------------------------------------------------------------------- A New Internet-Draft is available from the on-line Internet-Drafts directories. Title : GMPLS constraints consideration for CSPF path computation Author(s) : T. Otani, et al. Filename : draft-otani-ccamp-gmpls-cspf-constraints-01.txt Pages : 0 Date : 2005-7-21 This draft provides the guideline to consider generalized multi- protocol label switching (GMPLS) traffic-engineering (TE) attributes for constraint-based shortest path first (CSPF) path computation at a source node in a GMPLS network environment. This draft summarizes most possible cases of GMPLS constraint TE attributes at an ingress link, transit links and an egress link to establish a GMPLS label switched path (LSP) appropriately. A URL for this Internet-Draft is: http://www.ietf.org/internet-drafts/draft-otani-ccamp-gmpls-cspf-constraints-01.txt Dimitri.Papadimitriou@alcatel.be wrote: >hi, > >i would say that 1) should be considered iff there have been several identified >misbehaviour detected and inconsistencies observed such that a set of >prescriptive guidelines have to be put together (but this would translate some >"common" mis-interpretation of the set of documents we currently have) > >if this is not the case, then beside some base description/clarification, i >don't see how we could turn this into 2) or 3) without obtaining a larger sample >of CSPF/routing practices - the difference between 2) and 3) being that in the >former case, information gathered is processed to describe practices and devise >operational guidelines while in the latter the document appears as a collection >of practices from individual practices > >in any case, and in order to progress - we need more feedback - > > > >thanks, > >-dimitri. > > >*Tomohiro Otani <otani@kddilabs.jp>* >Sent by: owner-ccamp@ops.ietf.org >07/19/2005 09:48 ZE9 > >To: Dimitri PAPADIMITRIOU/BE/ALCATEL@ALCATEL >cc: ccamp@ops.ietf.org >bcc: >Subject: Re: >http://www.ietf.org/internet-drafts/draft-otani-ccamp-gmpls-cspf-constraints-00 > > >Hi Dimitri, > >Per our initial discussion, we would like to formulate >the unified view on CSPF consideration, and we simply >thought it is BCP. > >Considering more deeply, our draft does not define the >protocol itself but the protocol related specification > >In that sense, you are right, it is prescriptive more >than descriptive for guiding operations. Moreover, I came >up my mind that it may be more standard track than informational. >But I am not sure. What's your thought on this ? > >With best regards, > >tomo > >PS: We posted 01 version. > > > > >Dimitri.Papadimitriou@alcatel.be wrote: > > >authors, is the purpose of the above mentioned draft: > > > > 1. prescriptive (even if informational) > > > > 2. descriptive for guiding operations (BCP) > > > > 3. informative (survey-like) > > > >from what i read, it is more 1 like - is my understanding correct ? > > > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 01:25:02 +0000 Date: Fri, 22 Jul 2005 10:23:38 +0900 From: Kenji Kumaki <ke-kumaki@kddi.com> To: ccamp@ops.ietf.org Subject: draft-kumaki-ccamp-mpls-gmpls-interworking-01.txt Message-Id: <20050722102333.7745.KE-KUMAKI@kddi.com> MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit Hi, We updated the following draft. "Operational, Deployment and Interworking Considerations for GMPLS " http://www.ietf.org/internet-drafts/draft-kumaki-ccamp-mpls-gmpls-interworking-01.txt Since last draft,we mainly updated the following items. 1. merged "GMPLS Deployment in Existing IP/MPLS networks" draft http://www.ietf.org/internet-drafts/draft-ali-ccamp-gmpls-deployment-augmented-model-00.txt 2. changed the title of I-D and section due to merge of "GMPLS Deployment in Existing IP/MPLS networks" draft 3. added MPLS/GMPLS deployment,operational and interworking requirements as a section 4. added MPLS/GMPLS LSP Resource Affinity Mapping 5. removed section 3.1 OSPF extensions for Link Priority Identification section 3.2 ISIS extensions for Link Priority Identification 6. added Terminology 7. category change: BCP 8. corrected some typos We welcome your comments and feedback on this list or during Paris meeting. Thanks, Kenji -- Kenji Kumaki <ke-kumaki@kddi.com> Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 22 Jul 2005 00:30:05 +0000 From: k-sagara@crl.hitachi.co.jp Message-Type: Multiple Part MIME-Version: 1.0 Message-ID: <"MNXA011275l1181022U*/C=JP/ADMD=gmax000/PRMD=gmax000/O=KOI/OU=gmax000/S=Sagara/G=Kazuhiko-crl/"@MHS> Content-Type: text/plain; charset="us-ascii" To: <ccamp@ops.ietf.org> Date: Fri, 22 Jul 2005 09:26:45 +0900 Subject: Re:draft-otani-ccamp-interas-gmpls-te-03.txt (support) Content-Transfer-Encoding: 7bit Hi Session Chair and everyone; We think that "draft-otani-ccamp-interas-gmpls-te-03.txt" opens a very important scheme in the GMPLS inter-domain TE requirement for several vendors. We support his proposal and hope that the draft would pass through the charter process and be a WG document as soon as possible. Regards, Kazuhiko Sagara Hitachi ----------------------------------------------------------------------------- A New Internet-Draft is available from the on-line Internet-Drafts directories. Title : GMPLS Inter-domain Traffic Engineering Requirements Author(s) : T. Otani, et al. Filename : draft-otani-ccamp-interas-gmpls-te-03.txt Pages : 15 Date : 2005-7-20 This draft provides requirements for the support of generalized multi-protocol label switching (GMPLS) inter-domain traffic engineering (TE). Its main objective is to present the differences between MPLS inter-domain TE and GMPLS inter-domain TE. This draft covers not only GMPLS Inter-domain architecture but also functional requirements in terms of GMPLS signaling and routing in order to specify these in a GMPLS Inter-domain environment. Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 21 Jul 2005 14:59:17 +0000 Thread-Topic: End-to-end L2-LSP thread-index: AcWOBL6WoZ1LxoiyQ2CfXbLdUjivyA== Reply-To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> From: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> To: <richard.spencer@bt.com>, <ccamp@ops.ietf.org> Cc: Bcc: Subject: RE: End-to-end L2-LSP Date: Thu, 21 Jul 2005 23:59:07 +0900 Comment: GQ19@|@ZEk=E?,18?x, BcN1b<z:P<.F@, 4c4g Message-ID: <377fd01c58e04$be9b0150$8310fe81@email1> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Class: urn:content-classes:message Hi, Richard, Thank you for your tons of questions. It is a bit hard to answer in single mail. So, I divided it in two pieces. Here I send you the first part of answers to questions 1~5. >I agree that extending signalling down to the end-user is one >possible solution to address charging/billing issues, another possible >solution would be to use session based billing. I also agree that using >L2-LSPs could be a viable solution for providing traffic separation >and bandwidth guarantees in a service providers network, another Indeed, that's one important message I wanted to cast in this WG. Not only this is technically interesting, but the behind goal is about improving income structure of Internet service providers. I hope this can be an answer to the issue of QoS&Pricing that Neil Harrison once noted in last mail of this thread. Customers may basically use cl-ps service in fixed rate base, and when they need real-time service such as IP-TV or video phone, they may request simple CBR-like co-ps SVC-based service (i.e. L2-LSP) over cl-ps network using GMPLS signal. This way, operators may improve their service structure and invest to upgrade their access network. [snip] >1. If you want to eliminate the possibility of congestion in the >enterprise/home network using L2-LSPs for the reasons you >give below, then on every node in the network you will need: > >- A CPU: for control plane protocol processing >- Routing: At least static, an IGP for large networks and >OSPF-TE if TE is required >- Signalling: RSVP-TE >- Policing: Per connection policing, e.g. per VLAN > >I do not know how much it would cost to upgrade the $20 five >port 10/100 switch sitting on my desk to support all this, >but I do not believe this would be a trivial amount, particularly >at the volumes demanded by enterprises. I guess your five port $20 switch will be connected to machines owned only by you. If you can control resource usage of your machines, you don't need GMPLS implemented switch. However, if you need to share a switch with some of your colleague, I will recommend spending more that $20 on switch in order to enjoy nice video communication while your colleague abusing your network with P2P file transfer. I think this is worth while to spend money considering importance of your personal communication. As I noted in previous mail, GMPLS implemented switch will be necessary in nodes where automated resource control is crucial. >In addition, a hierarchical >architecture would be required for the solution to scale and >therefore aggregation switches within the enterprise network >would also need to support hierarchical CAC and policing. This is beyond issue we can discuss at the moment. I'd like further technical discussion, however, I think we need to wait until this idea of L2-LSP is officially chartered in CCAMP. I hope enough number of people support the idea that we can continue work on this issue. >2. Unless you want to reinvent LANE, then you will need to >continue to support CL Ethernet broadcasts and multicasts on >GMPLS enabled switches. To protect the L2-LSP traffic from >CL broadcast traffic you could rate limit it to say 1% of >the bandwidth available. However, how would you capacity >plan for and rate limit (dynamic, receiver driven) multicast >traffic for which usage patterns are difficult to predict? >In my opinion you will need to use over provisioning along with CoS, >which would raise the question, why is over provisioning >and CoS OK for multicast traffic >(which could be critical, e.g. trading floor information), >but not OK for unicast traffic? There will be a technical solution for multicast and broadcast. A member of DT once showed his interest on multicast solution for Ethernet using RSVP instead of IGMP. I also have an idea of extending point-to-point L2-LSP to multipoint L2-LSP as similar way that learning bridges learn backward path from source address. However, I have some doubt on necessity of multicast service. Do we really need multicast for streaming service in local/access ? I think even IP-TV will be better serviced using point-to-point connection in access network. There's very low chance of two subscribers receiving same channel at the same time if the receiver group is not that great. Multicasting is necessary in core network, however I doubt the same needs in local/access network, except some rare case. >3. Routers are relatively stable systems that are switched on >permanently and therefore RSVP-TE sessions are stable >and only go down when there is a failure. End user systems >however such as PCs, set top boxes, games consoles, etc. >are not always switched on and are not as stable. >What will the effect on the network be when large number >of RSVP-TE sessions are going up/down, e.g. as a result >of game console resets or PC crashes/freezes? >These effects will ripple through the enterprise/home >network into the operators network, and therefore mechanisms >must be put in place to protect against them. Control overhead of RSVP was discussed a lot in conjunction with IntServ. There have been proposals to reduce the overhead in core network such as RFC3175 or [HIER]. In particular, proposed architecture for LSP aggregation [HIER] will greatly help to solve the issue. % [HIER] draft-ieft-mpls-lsp-hierarchy-08.txt >4. When an IP device wants to forward a packet it uses ARP >to discover the MAC address of the IP next hop >(unless its already in the MAC table) and adds the Ethernet >header before forwarding the packet on. Now, if you want to >extend connections down to the end user device, then the >device will need to map IP traffic to different L2-LSPs. >A simple 1-to-1 mapping between an IP next hop address >and a MAC address will not be sufficient as there may >be multiple connections (e.g. using different VLAN IDs) >associated with an IP next hop. So, policies will need >to be configured on end user devices to map IP traffic >(e.g. based on src/dest IP address, port numbers, etc.) >to a specific L2-LSP (e.g. based on src/dest MAC address and/or VLAN ID). You pointed out important issue. Currently, Ethernet layer only maps IP and MAC in 1-to-1. Unless Ethernet-IP interface is extended as you suggested, users must be satisfied with at most single L2-LSP established between machine to machine. If they want another connection, they have to disconnect previous connection. This scenario will be OK in most real-life applications. However in future, extension to layer interface will be necessary and the solution should be consulted with IEEE. >5. End user devices such as PCs and VoIP phones are >likely to only need to support a relatively low number >of L2-LSPs. However, devices such as content servers >and call servers/gateways will need to support >hundreds/thousands of L2-LSPs. Connection setup and >state maintenance for large numbers of L2-LSPs will >consume a significant amount of processing resources, >as will policies to map IP traffic to the different L2-LSPs. >Therefore, I would expect that the system software >(and possibly hardware) will need to be designed from >the ground up to be able to support large numbers of L2-LSPs. Some high-performance servers are capable of handling millions of TCP connections. I believe the added complexity for LSP connection will not be far greater that the work-load normally assumed in contemporary servers. It may be a problem to local gateways if the signaling load is not bypassed. However, even in this case, there can be some solutions we can devise. ok, for rest of questions, I'll reply it in next mail. thanks Jaihyung -----?? ???----- ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com> ?? ??: 2005-07-21 ?? 12:09:56 ?? ??: "jaihyung@etri.re.kr" <jaihyung@etri.re.kr>, "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> ??: ??: RE: End-to-end L2-LSP Jaihyung, I agree that extending signalling down to the end-user is one possible solution to address charging/billing issues, another possible solution would be to use session based billing. I also agree that using L2-LSPs could be a viable solution for providing traffic separation and bandwidth guarantees in a service providers network, another possible solution would be to extended MPLS into the access network. However, if you want to extend L2-LSPs (not just signalling) down to the end user, then there are a number of complicated technical and commercial issues that must be overcome. Please see below for some of the key issues: 1. If you want to eliminate the possibility of congestion in the enterprise/home network using L2-LSPs for the reasons you give below, then on every node in the network you will need: - A CPU: for control plane protocol processing - Routing: At least static, an IGP for large networks and OSPF-TE if TE is required - Signalling: RSVP-TE - Policing: Per connection policing, e.g. per VLAN I do not know how much it would cost to upgrade the $20 five port 10/100 switch sitting on my desk to support all this, but I do not believe this would be a trivial amount, particularly at the volumes demanded by enterprises. In addition, a hierarchical architecture would be required for the solution to scale and therefore aggregation switches within the enterprise network would also need to support hierarchical CAC and policing. 2. Unless you want to reinvent LANE, then you will need to continue to support CL Ethernet broadcasts and multicasts on GMPLS enabled switches. To protect the L2-LSP traffic from CL broadcast traffic you could rate limit it to say 1% of the bandwidth available. However, how would you capacity plan for and rate limit (dynamic, receiver driven) multicast traffic for which usage patterns are difficult to predict? In my opinion you will need to use over provisioning along with CoS, which would raise the question, why is over provisioning and CoS OK for multicast traffic (which could be critical, e.g. trading floor information), but not OK for unicast traffic? 3. Routers are relatively stable systems that are switched on permanently and therefore RSVP-TE sessions are stable and only go down when there is a failure. End user systems however such as PCs, set top boxes, games consoles, etc. are not always switched on and are not as stable. What will the effect on the network be when large number of RSVP-TE sessions are going up/down, e.g. as a result of game console resets or PC crashes/freezes? These effects will ripple through the enterprise/home network into the operators network, and therefore mechanisms must be put in place to protect against them. 4. When an IP device wants to forward a packet it uses ARP to discover the MAC address of the IP next hop (unless its already in the MAC table) and adds the Ethernet header before forwarding the packet on. Now, if you want to extend connections down to the end user device, then the device will need to map IP traffic to different L2-LSPs. A simple 1-to-1 mapping between an IP next hop address and a MAC address will not be sufficient as there may be multiple connections (e.g. using different VLAN IDs) associated with an IP next hop. So, policies will need to be configured on end user devices to map IP traffic (e.g. based on src/dest IP address, port numbers, etc.) to a specific L2-LSP (e.g. based on src/dest MAC address and/or VLAN ID). 5. End user devices such as PCs and VoIP phones are likely to only need to support a relatively low number of L2-LSPs. However, devices such as content servers and call servers/gateways will need to support hundreds/thousands of L2-LSPs. Connection setup and state maintenance for large numbers of L2-LSPs will consume a significant amount of processing resources, as will policies to map IP traffic to the different L2-LSPs. Therefore, I would expect that the system software (and possibly hardware) will need to be designed from the ground up to be able to support large numbers of L2-LSPs. 6. Enterprise networks are built using not just switches but routers as well. Firstly, for L2-LSPs to be set up between switches and routers, they must belong to the same address space to communicate with each other. However, this address space must be separate from the IPv4 forwarding address space. To use an example of why this is true, routers connected via Ethernet segments are considered to be directly connected, however, if the IPv4 and L2-LSP address spaces were the same, then IP routers separated by GMPLS switches would be multiple hops away from each other. In addition to address space separation, routing protocol instances must also be separated. This need for this separation is obvious when considering using a link state routing protocol which relies on a topological view of the network, i.e. the IPv4 and L2-LSP topologies are different. Secondly, routers break up Ethernet segments and must therefore must always be the headend or tailend for a L2-LSPs, i.e. they cannot be midpoints. This means that from a control plane perspective, on receiving a L2-LSP path message a router must terminate the LSP and send back a resv message whether it is the intended destination/sink or not. It must then originate a path message for a new L2-LSP and send it onto the next hop GMPLS switch. This is not how RSVP-TE works today and therefore will require modifications. In addition, traffic will be blackholed if transmitted by the source before all the separate L2-LSPs have been setup end-to-end to the sink/destination. This means that some kind of end-to-end signalling message will need to be sent by the sink/destination once the final L2-LSP has been established to inform the source that the end-to-end connectivity has been provided. This functionality is not supported by RSVP-TE today and will therefore require protocol modifications. Thirdly, routers forward packets based on information in the IP header. So, when a router receives a packet via a L2-LSP it will strip of the L2 header and perform an IP lookup. Now, for the packet to be forwarded using the correct L2-LSP the router must have a policy that tells it which L2-LSP to use as there may be multiple L2-LSPs to the same destination, e.g. for different voice calls or video downloads. However, before a policy can even be configured, the router needs to know what traffic (e.g. based on src/dest address, port number, DSCP, etc.) to map to which L2-LSP, and this information needs to come from the traffic source using signalling, i.e. RSVP-TE. I do not believe this is supported today and will therefore require modifications. Regards, Richard > -----Original Message----- > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > Sent: 14 July 2005 04:24 > To: Spencer,R,Richard,XDE73 R; ccamp@ops.ietf.org > Subject: RE: End-to-end L2-LSP > > Dear Richard, > > See in-line please, > > >Thank you for taking the time to respond to my questions. I > anticipated > >that your response would include "over-provisioning is the > best answer" > >to decrease the chance of congestion on non-GMPLS switches. > >This is what is done in the LAN environment today in > conjunction with CoS, > >and this is why I believe there will be two opposing views on how > >far L2-LSP connections should be extended down to the user. One view > >is that over provisioning and CoS work fine in the > enterprise network today, > >and therefore the L2-LSP should only be extended down to the point > >at which the enterprise network connects to the SP network, > >e.g. Ethernet DSLAM. I think the other view will be that > over provisioning > >and CoS in the enterprise network don't work and therefore L2-LSPs > >should be provided end-to-end using only GMPLS capable switches. > >I wouldn't have thought there would be much call for mixed > >GMPLS/non-GMPLS enterprise networks (except for during migration) > >due to the opposing views, e.g. if over provisioning is good > enough for > >one switch, why isn't it good enough for the other? > > Over-provisioned network is non-guaranteed network. It can't prevent > occasional congestion at any rate. Further, network condition > in private > network is very diverse. You can't predict minimum service level > inside every private network. > > Suppose that an ISP want to provide high-quality, video-phone service > for enterprise users. Nothing can be guaranteed inside the > over-provisioned > customer network. If the service level is not satisfactory, > customers will complain > to service provider even if the main cause is in his network > or destination network. > We need some structure that coordinate service quality in > both private sector, > as well as in public sector consistently. > An end-to-end signal is an effective means to communicate between > user terminal, private switches and provider switches, and > concert service > quality at each different operations network. This is a goal > that RSVP (RFC2205) > pursued, however ineffective in local network because > Ethernet industry > didn't support RSVP over Ethernet (RFC2814, 2815, 2816). > Now we have another opportunity to realize the goal using L2-LSP. > Therefore, I believe end-to-end LSP is necessary even in > enterprise network > for the purpose of providing guaranteed commercial service. > > >I think where most people will agree is that using connections > >in the SPs access network does make sense in order to > provide efficient > >link utilisation and per connection policing to ensure one customers > >traffic can not affect another's. Regarding the ability to > set up new > >connections based on a request from the user (rather than > constraining > >RSVP signalling to inside the SP network), I think this will come > >down to enterprise customer demand for improved service performance. > >If a customer wasn't getting the performance they needed from their > >SP and was offered a new service that allowed them to setup > new L2-LSP > >connections on demand that would offer better performance > (e.g. for voice > >calls or video conferences), then upgrading the user devices > to support > >the new service could well be an attractive proposition. If > on the other > > Yes, this will be one of prospective service newly enabled > using L2-LSP. > > >hand a customer buys a best effort service and considers the > performance > >to be 'good enough', then there would be no reason to go to > the expense > >of upgrading the user devices to support the set up of L2-LSPs. > > Ever increasing besteffort bandwidth will eventually not satisfy > both providers and customers for many reasons. Service providers > will have difficulty in collecting expenses enough to invest > on increased > bandwidth. This will in turn gradually downgrade overall > communication quality, > and applications requiring real-time service will never > satisfy customers. > We already see such symptom in some low-price service providers. > I believe new service structure that can provide difference > of service > quality will offer benefit to customers, providers as well as > vendors in the end. > > I also look forward to see you at Paris and continue helpful > discussion. > > Sincerely > > Jaihyung > > > > > Dr. Jaihyung Cho > ETRI, Korea > phone : 042) 860-5514 > oversea: +82-42-860-5514 > fax: +82-42-861-5550 > > > > > -----?? ???----- > ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com> > ?? ??: 2005-07-14 ?? 1:10:30 > ?? ??: "jaihyung@etri.re.kr" <jaihyung@etri.re.kr>, > "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> > ??: > ??: RE: End-to-end L2-LSP > > > > Jaihyung, > > Thank you for taking the time to respond to my questions. I > anticipated that your response would include > "over-provisioning is the best answer" to decrease the chance > of congestion on non-GMPLS switches. This is what is done in > the LAN environment today in conjunction with CoS, and this > is why I believe there will be two opposing views on how far > L2-LSP connections should be extended down to the user. One > view is that over provisioning and CoS work fine in the > enterprise network today, and therefore the L2-LSP should > only be extended down to the point at which the enterprise > network connects to the SP network, e.g. Ethernet DSLAM. I > think the other view will be that over provisioning and CoS > in the enterprise network don't work and therefore L2-LSPs > should be provided end-to-end using only GMPLS capable > switches. I wouldn't have thought there would be much call > for mixed GMPLS/non-GMPLS enterprise networks (except for > during migration) due to the opposing views, e.g. if over > provisioning is good enough for one switch, why isn't it good > enough for the other? > > I think where most people will agree is that using > connections in the SPs access network does make sense in > order to provide efficient link utilisation and per > connection policing to ensure one customers traffic can not > affect another's. Regarding the ability to set up new > connections based on a request from the user (rather than > constraining RSVP signalling to inside the SP network), I > think this will come down to enterprise customer demand for > improved service performance. If a customer wasn't getting > the performance they needed from their SP and was offered a > new service that allowed them to setup new L2-LSP connections > on demand that would offer better performance (e.g. for voice > calls or video conferences), then upgrading the user devices > to support the new service could well be an attractive > proposition. If on the other hand a customer buys a best > effort service and considers the performance to be 'good > enough', then there would be no reason to go to the expense > of upgrading the user devices to support the set up of L2-LSPs. > > I look forward to the discussions in Paris. > > Regards, > Richard > > > -----Original Message----- > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > > Sent: 13 July 2005 15:14 > > To: Spencer,R,Richard,XDE73 R; ccamp@ops.ietf.org > > Subject: RE: End-to-end L2-LSP > > > > > > > > > > Dear Richard Spencer > > > > see in-line > > > > >Jaihyung, > > > > > >Perhaps I'm becoming too much of a pessimist;-) > > > > oh, not at all, > > you are reasonable enough to dig up unclear knowledge > > please go ahead until you finally get satisfactory answer. > > > > > > >Ignoring the commercial aspects, please can you help clarify > > a few technical > > >questions I have in order to help me gain a better > > understanding of the proposal? > > > > > >1. You say that one of the key reasons for using GMPLS in > > the enterprise > > >network is to provide end-to-end QoS control. However, you > > also say that not > > >all the switches will need to support GMPLS, only those at > > bottle necks. > > >Please can you expand on this to explain how you can have > end-to-end > > >QoS control when the L2-LSP is not end-to-end? > > > > > > The contextual meaning of 'End-to-end LSP' is that GMPLS label is > > assigned from source terminal and removed at destination terminal, > > because RSVP-TE message is exchanged from host to host. > > It may not perfectly guarantee QoS when the GMPLS deployment > > is not complete, however it will at least be good enough > > for end-user applications to tell those GMPLS implemented > L2 switches > > what frame must be given priority service when congestion occur. > > > > > > >As being discussed > > >on the PWE3 mailing list, the performance a service receives > > is only as > > >good as the worst performing server layer. Therefore, if > there is a > > >standard 802.1 workgroup switch between the user and the > > L2-LSP that is > > >experiencing congestion (e.g. due to a large amount of > > peer-to-peer traffic), > > >then the QoS that the user receives will only be as good as > > that provided > > >by the congested workgroup switch. > > > > > > of course, ETH-GMPLS does not prevent congestion inevitably > occurring > > in those non-GMPLS deployed part. > > > > > > >The only way I can see a network using a mixture of GMPLS > > and non-GMPLS > > >switches guaranteeing end-to-end QoS would be to design the > > network in > > >such a way that the non-GMPLS switches could never become > congested. > > > > > > Yes, that will be the way wise network designer improve > their network. > > > > > > >To do this would require that RSVP requests be used for ALL > > traffic flows > > >to ensure that there was enough bandwidth available, along > > with strict > > >per L2-LSP policing on the user devices (e.g. PC). > > >However, the obvious flaw in this approach is that someone could > > >just come along and connect > > >an end device to a non-GMPLS switch and start transmitting traffic > > >(without performing CAC/policing) causing the switch to > > become congested. > > >Therefore, please can you explain how end-to-end QoS can be > > guaranteed > > >in a mixed non-GMPLS and GMPLS switching environment? > > > > > > Of course traffic came from non controlled ingress can not be > > prevented > > in non-GMPLS node. There's no magic. Only the > > over-provisioning is the > > best answer. However, once the aggregated traffic pass 'ANY' one of > > GMPLS enabled node, those uncontrolled traffic will suffer > > queuing discipline. > > > > > > >2. Section 5.1 currently states "When the customer initiates > > data transmission, > > >the access switch maps the user flow into the L2 LSP. > > Mapping procedure is > > >subject to implementation choices, and is out of the scope of this > > >document." In order to be able to map any traffic, first of > > all the switch needs > > >to know what type of traffic (e.g. based on source/dest MAC, > > VLAN ID, 802.1p > > >bits, etc) should be mapped to the L2-LSP. In Figure 1, if > > an L2-LSP is set up > > >from the DSLAM to the edge router based on a request from > > the user (e.g. via > > >a PC or VoIP phone), then won't the user also need to signal > > what type of > > >traffic should to be mapped to the L2-LSP? > > > > > > Thank you for pointing out important thing. > > That's the most unfortunate part I reluctantly agreed because the > > work condition in the framework DT was not to span L2-LSP > > beyond provider network. > > As you noticed, mapping user traffic at provider access is awkward. > > A smooth method will be for user terminal aware of the L2 label > > to use by receiving RSVP-TE message. > > However, there was strong argument that RSVP-TE must not > > go beyond provider network because the DT is not chartered > > to work in customer network. > > That's why I requested amendment of CCAMP charter > > to include user terminal explicitly in CCAMP work scope, so that > > avoid such silly argument in future. > > > > > > >3. In section 5.1, the text states that the aggregation > > switch between the > > >DSLAM and edge router may be a 802.1 switch, i.e. it doesn't > > have to support > > >GMPLS. To set up a connection, first of all there needs to > > be a data plane > > >capable of carrying control plane traffic between GMPLS > > peers. Secondly, > > >the GMPLS peers need reachability information in order to > > forward control > > >messages onto the next peer (i.e. a static or IGP route). If > > the GMPLS peers > > >(e.g. the DSLAM and edge router in Figure 1) are not > > directly connected, > > >then the GMPLS control packets will need to be forwarded by > > the 802.1 switch. > > >Would the idea be to have say a "control traffic VLAN" > > configured between the > > >DSLAM and edge router via the transit 802.1 switch so that > > control packets > > >could be forwarded transparently by the 802.1 switch? > > > > > > Although it doesn't preclude use of VLAN, I think such > > special configuration > > for RSVP-TE control message is not necessary in legacy switches. > > Legacy switches only need to pass RSVP-TE messages as they > > pass normal data frames. Only the GMPLS implemented switches > > need to distinguish frames containing RSVP-TE messages and > > snoop them. > > Detail of implementation is, of course, for future discussion. > > > > > > >4. If we assume that control messages can be transparently > forwarded > > >by the intermediate 802.1 aggregation switch then how are > > user frames > > >forwarded in the data plane by the 802.1 switch? > > > > > > Legacy switches do not distinguish control/data frames. > > > > > > >If the DSLAM sets up a L2-LSP > > >to the edge router using say VLAN ID 99, how does the switch > > know that any > > >packets coming in with a VLAN ID of 99 should be forwarded > > out of the port > > >connecting to the edge router? > > > > > > oh, you misunderstood the L2-LSP. > > The question you are asking depends on how L2 label is encoded. > > there was hot debate in the DT about the issue how L2 label > > is encoded. > > My proposal is to use part of MAC address for label and the > > other people > > argued to use VLAN ID for label. The debate was not concluded > > because the goal of the L2SC framework document is not to select > > implementation method but to draw necessary work scope. > > I believe this issue will be dealt in depth after Paris meeting. > > So, please reserve your question until then and ask to those > > VLAN proponents. > > > > > > >5. During connection setup, I take it that the GMPLS peers > > assume that > > >resources are available within the intermediate 802.1 switch > > to support > > >the connection (as its transparent). However, as there is no > > >CAC/policing being performed by the 802.1 switch > > >(the L2-LSPs are transparent), and due to the any-to-any and > > broadcast > > >based nature of Ethernet (i.e. its CL properties), then > how can this > > >assumption be made? This comes back to my original question, how > > >can end-to-end QoS be guaranteed in an environment where there is > > >a mixture of 802.1 and GMPLS enabled switches? > > > > > > yes, you came back to the question I answered in above ;-) > > so, please see above. > > GMPLS enabled nodes must be positioned at points congestion > > is likely to be expected. > > > > Thanks > > > > Jaihyung > > > > > > ============================================ > > > > > > > > -----?? ???----- > > ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com> > > ?? ??: 2005-07-13 ?? 8:53:13 > > ?? ??: "jaihyung@etri.re.kr" <jaihyung@etri.re.kr>, > > "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> > > ??: > > ??: RE: End-to-end L2-LSP > > > > > > > > > > Jaihyung, > > > > Perhaps I'm becoming too much of a pessimist;-) Ignoring the > > commercial aspects, please can you help clarify a few > > technical questions I have in order to help me gain a better > > understanding of the proposal? > > > > 1. You say that one of the key reasons for using GMPLS in the > > enterprise network is to provide end-to-end QoS control. > > However, you also say that not all the switches will need to > > support GMPLS, only those at bottle necks. Please can you > > expand on this to explain how you can have end-to-end QoS > > control when the L2-LSP is not end-to-end? As being discussed > > on the PWE3 mailing list, the performance a service receives > > is only as good as the worst performing server layer. > > Therefore, if there is a standard 802.1 workgroup switch > > between the user and the L2-LSP that is experiencing > > congestion (e.g. due to a large amount of peer-to-peer > > traffic), then the QoS that the user receives will only be as > > good as that provided by the congested workgroup switch. > > > > The only way I can see a network using a mixture of GMPLS and > > non-GMPLS switches guaranteeing end-to-end QoS would be to > > design the network in such a way that the non-GMPLS switches > > could never become congested. To do this would require that > > RSVP requests be used for ALL traffic flows to ensure that > > there was enough bandwidth available, along with strict per > > L2-LSP policing on the user devices (e.g. PC). However, the > > obvious flaw in this approach is that someone could just come > > along and connect an end device to a non-GMPLS switch and > > start transmitting traffic (without performing CAC/policing) > > causing the switch to become congested. Therefore, please can > > you explain how end-to-end QoS can be guaranteed in a mixed > > non-GMPLS and GMPLS switching environment? > > > > 2. Section 5.1 currently states "When the customer initiates > > data transmission, the access switch maps the user flow into > > the L2 LSP. Mapping procedure is subject to implementation > > choices, and is out of the scope of this document." In order > > to be able to map any traffic, first of all the switch needs > > to know what type of traffic (e.g. based on source/dest MAC, > > VLAN ID, 802.1p bits, etc) should be mapped to the L2-LSP. In > > Figure 1, if an L2-LSP is set up from the DSLAM to the edge > > router based on a request from the user (e.g. via a PC or > > VoIP phone), then won't the user also need to signal what > > type of traffic should to be mapped to the L2-LSP? > > > > 3. In section 5.1, the text states that the aggregation > > switch between the DSLAM and edge router may be a 802.1 > > switch, i.e. it doesn't have to support GMPLS. To set up a > > connection, first of all there needs to be a data plane > > capable of carrying control plane traffic between GMPLS > > peers. Secondly, the GMPLS peers need reachability > > information in order to forward control messages onto the > > next peer (i.e. a static or IGP route). If the GMPLS peers > > (e.g. the DSLAM and edge router in Figure 1) are not directly > > connected, then the GMPLS control packets will need to be > > forwarded by the 802.1 switch. Would the idea be to have say > > a "control traffic VLAN" configured between the DSLAM and > > edge router via the transit 802.1 switch so that control > > packets could be forwarded transparently by the 802.1 switch? > > > > 4. If we assume that control messages can be transparently > > forwarded by the intermediate 802.1 aggregation switch then > > how are user frames forwarded in the data plane by the 802.1 > > switch? If the DSLAM sets up a L2-LSP to the edge router > > using say VLAN ID 99, how does the switch know that any > > packets coming in with a VLAN ID of 99 should be forwarded > > out of the port connecting to the edge router? > > > > 5. During connection setup, I take it that the GMPLS peers > > assume that resources are available within the intermediate > > 802.1 switch to support the connection (as its transparent). > > However, as there is no CAC/policing being performed by the > > 802.1 switch (the L2-LSPs are transparent), and due to the > > any-to-any and broadcast based nature of Ethernet (i.e. its > > CL properties), then how can this assumption be made? This > > comes back to my original question, how can end-to-end QoS be > > guaranteed in an environment where there is a mixture of > > 802.1 and GMPLS enabled switches? > > > > Thanks, > > Richard > > > > > -----Original Message----- > > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > > > Sent: 12 July 2005 09:21 > > > To: Spencer,R,Richard,XDE73 R; ccamp@ops.ietf.org > > > Subject: RE: End-to-end L2-LSP > > > > > > > > > > > > > > > Dear Richard, > > > > > > Yes, I believe it is not only commercially profitable but it > > > may also grow total > > > revenue of service operators by introducing new services. > > > > > > For example, think the example of IP phone you referred. > > > How do you think you can provide reliable VoIP service and > > > safely authenticate phone users and charge fees as counted by > > > packet usage > > > using simple Ethernet based xDSL network ? > > > > > > There's hardly new service that providers may offer using > > > Ethernet based > > > architecture, and collect sufficient profit to investment. > > > > > > A sophisticated ATM switch might do the work if it is not > > expensive. > > > However, my intension is modifiying existing design of > > > Ethernet hardware > > > with a little bit additional cost, and provide the level of > > > service that > > > ATM switches might offer. > > > > > > We have initial design of the Ethernet-MPLS hardware. I > don't think > > > GMPLS enabled Ethernet switch will be so complex. The > market price > > > of GMPLS enabled switch will not be so expensive than normal > > > Ethernet switches. > > > > > > Further, it only requires upgrade of switches in some > > > bottleneck points of network. > > > Read the framework documents of L2SC. It doesn't require > > > entire replacement > > > of switches to introduce new service. > > > > > > Every year, ISPs purchase new access switches and replaces > > > old machines. > > > With a bit additional cost, ISPs may use the GMPLS enabled > > switch for > > > normal Ethernet switching in initial deployment. > > > > > > When the GMPLS function is enabled in some upgraded part of > > network, > > > ISPs may provide high quality, reliable VoIP service to > > xDSL customers > > > using RSVP-TE enabled VoIP phone. > > > The high quality video-phone service enabled by GMPLS will be > > > competable to legacy voice only phone service. > > > > > > > > > There are lots of things I can argue about commercial aspects > > > of this technique. > > > However I don't think it is not appropriate to discuss this > > > issue in this mailing list. > > > All I can say is, > > > yes, I believe that it may be commercially viable. > > > > > > > > > Thanks > > > > > > Jaihyung > > > > > > ================================================ > > > > > > Do you really think the benefits of extended GMPLS down to > > > the end user could ever justify the costs involved in > > > upgrading LAN switches and things like IP phones (simple low > > > cost commodity items) to support GMPLS signalling/routing? > > > > > > Your proposal may be technically possible (though it would > > > require new hardware and significant software upgrades), but > > > commercially I would consider this to be a non-starter. > > > > > > Regards, > > > Richard > > > > > > > -----Original Message----- > > > > From: owner-ccamp@ops.ietf.org > > [mailto:owner-ccamp@ops.ietf.org]On > > > > Behalf Of CHO, JAI HYUNG > > > > Sent: 12 July 2005 01:10 > > > > To: ccamp@ops.ietf.org > > > > Subject: End-to-end L2-LSP > > > > > > > > > > > > > > > > Dear all, > > > > > > > > Greetings. > > > > > > > > I am a co-editor of the Ethernet-GMPLS framework document. > > > > (draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt) > > > > > > > > I am writing this to explain the motivation behind the > > > > proposal of scenario-1, and to raise the issue of > > > > end-to-end LSP to be included in CCAMP charter. > > > > > > > > An important goal intended in scenario-1 is > > > > QoS control over end-to-end packet delivery using LSP. > > > > In other words, the purpose of scenario-1 is extending coverage > > > > of GMPLS control including customer network and > end-user device, > > > > such as PC, VoIP phone, IP-TV, or even future 4G mobile phone. > > > > End-user applications will have a means to request resources > > > > to network using GMPLS signal. > > > > > > > > Unfortunately, this view is not included in the > > > > framework because there was strong opinion that > > > > L2-LSP must not span beyond provider network because > > > > CCAMP is not chartered to work on customer > > > > network. I am wondering if this is right argument. > > > > Perhaps the charter can be amended to include this work. > > > > > > > > I believe there is little technical reason to preclude > > > > LSP being used in customer network. It is a reasonable > > > > assumption that private company may also use GMPLS > > > > implemented Ethernet switch, once the L2SC work is successful. > > > > Currently, CCAMP is the only place discussing the matter of > > > > Ethernet and LSP. There is no other WG dealing similar issue. > > > > If CCAMP wants to work on Ethernet, > > > > I think the coverage of common GMPLS control must be > > > > extended to include all area where Ethernet is deployed. > > > > Currently, Ethernet is dominantly deployed in first-mile > > > > network where access network as well as customer network > > > > and user hosts are usually included. Therefore, user host must > > > > be considered as initiation and termination point of L2-LSP. > > > > > > > > Another important reason to include user host in CCAMP > > > > work scope is to achieve consistent end-to-end control of QoS. > > > > End-to-end QoS has long been desired goal since RSVP was > > > > first designed. This goal can be best achieved using > > > > end-to-end GMPLS signal and L2-LSP. > > > > Currently, RSVP (RFC2205) is available in many user hosts. > > > > There is only a little difference between RSVP and RSVP-TE. > > > > If the RSVP-TE signal of provider network can reach user host > > > > across Ethernet based private network, an LSP can be > established > > > > from application to application via the provider backbone. > > > > Service providers will be able to control application flows > > > > in L2-LSP level as well as in aggregated LSP pipe of similar > > > > application type. A variety of new services will be > enabled using > > > > the enhanced capability of distinguishing and controlling each > > > > individual application flow. > > > > > > > > I personally believe that end-to-end L2-LSP may offer > > > > innovative solutions for servicing IP-TV as well as fast > > mobility. > > > > There have been some experimental approaches combining > > > > MPLS and IP mobility. However such potential of the new > > > > application can only be explored when end-user host is > > > > included in the scope of CCAMP work area. > > > > > > > > Therefore, I hereby sincerely request people's attention > > > > and support on this perspective of end-to-end LSP. > > > > If there is enough number of people agree on this view, > > > > the WG can request the necessary update of the charter. > > > > > > > > Thank you. > > > > > > > > Jaihyung Cho > > > > > > > > > > > > > > > > ETRI, Korea > > > > phone : 042) 860-5514 > > > > oversea: +82-42-860-5514 > > > > fax: +82-42-861-5550 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 21 Jul 2005 11:35:13 +0000 Message-ID: <164701c58de8$60c5ec00$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: "Tom Petch" <nwnetworks@dial.pipex.com>, <ccamp@ops.ietf.org> Subject: Re: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON Lexicography Date: Thu, 21 Jul 2005 12:27:01 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit OK Understood. A ----- Original Message ----- From: "Tom Petch" <nwnetworks@dial.pipex.com> To: "Adrian Farrel" <adrian@olddog.co.uk>; <ccamp@ops.ietf.org> Sent: Wednesday, July 20, 2005 8:41 PM Subject: Re: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON Lexicography > <inline> > Tom Petch > ----- Original Message ----- > From: "Adrian Farrel" <adrian@olddog.co.uk> > To: "Tom Petch" <nwnetworks@dial.pipex.com>; <ccamp@ops.ietf.org> > Sent: Wednesday, July 20, 2005 11:54 AM > Subject: Re: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON > Lexicography > > > > Tom, > > > > I don't think it is CCAMP's role to correct the English of another SDO :-) > > > > Adrian > > OK, but then I do not think you should reply "Correct" because IMHO it is not > correct; the response should be something about what, primarily, CCAMP is > responsible for:-) > > > ----- Original Message ----- > > From: "Tom Petch" <nwnetworks@dial.pipex.com> > > To: "Adrian Farrel" <adrian@olddog.co.uk>; <ccamp@ops.ietf.org> > > Sent: Tuesday, July 19, 2005 7:18 PM > > Subject: Re: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON > > Lexicography > > > > > > > Adrian > > > > > > I think it should be > > > " > > > - CCAMP has primary responsibility for the descriptions of GMPLS terms > > > > > > Correct " > > > > > > even if that is not what SG15 said. > > > > > > Primarily (sic) CCAMP has responsibility for the specification of GMPLS, > > not > > > just its terms:-) > > > > > > Tom Petch > > > > > > ----- Original Message ----- > > > From: "Adrian Farrel" <adrian@olddog.co.uk> > > > To: <ccamp@ops.ietf.org> > > > Cc: "'Kireeti Kompella'" <kireeti@juniper.net>; <zinin@psg.com>; "Bill > > Fenner" > > > <fenner@research.att.com>; "Scott Bradner" <sob@harvard.edu> > > > Sent: Tuesday, July 19, 2005 1:59 PM > > > Subject: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON > > > Lexicography > > > > > > > > > > Hi, > > > > > > > > Please comment on this draft response liaison which I intend to send > > on > > > > Monday 25th July. > > > > You can see the original incoming liaison at > > > > http://www.olddog.co.uk/incoming.htm > > > > > > > > Thanks, > > > > Adrian > > > > ===== > > > > To: ITU-T SG15 Questions 12 and 14 > > > > From: IETF CCAMP Working Group > > > > Subject: Response to your liaison on GMPLS/ASON Lexicography > > > > For: Action > > > > Deadline: August 31st 2005 > > > > > > > > The CCAMP Working Group would like to thank SG15 and especially > > Questions > > > > 12 and 14 for the time and effort they put into providing input and > > > > feedback on the GMPLS/ASON lexicography Internet-Draft that is being > > > > developed by CCAMP. Much of the material supplied as direct comments > > or as > > > > mark-ups to the Internet-Draft has been incorporated into the latest > > > > revision which is available at > > > > > > > > > > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-gmpls-ason-lexicography-03. > > > txt > > > > > > > > In your liaison you state some assumptions which we would like to > > comment > > > > on as follows. > > > > > > > > - The purpose of the document is to enable those familiar with ASON > > > > to understand GMPLS terminology in an ASON context > > > > > > > > Correct. > > > > > > > > - GMPLS terms are described informally in the document in a way that > > > > is generally consistent with GMPLS (existing RFCs and work in > > progress) > > > > > > > > While the definition of GMPLS terms would be helpful for understanding > > > > GMPLS terms in a GMPLS context, that is not the purpose of this > > document. > > > > > > > > - Based on these descriptions, interpretations of these GMPLS terms > > > > in ASON terminology are provided > > > > > > > > It is intended that the interpretation of the GMPLS terms is not based > > on > > > > these descriptions, but based on the full meaning of the GMPLS terms. > > The > > > > descriptions of the GMPLS terms are provided in this Internet-Draft > > for > > > > context and a brief summary. > > > > > > > > - CCAMP has primarily responsibility for the descriptions of GMPLS > > terms > > > > > > > > Correct. > > > > > > > > - Q12/15 has responsibility to provide input to the GMPLS-ASON > > > > interpretations (based on the descriptions in this draft) > > > > > > > > We welcome this assumption of responsibility by Q12/15 to provide > > input to > > > > the Internet-Draft. While the Internet-Draft remains under the overall > > > > editorial control of the CCAMP working group, we hope to be able to > > use > > > > the text supplied by Q12/15 with only editorial changes. > > > > > > > > > > > > Additionally, we welcome the pledge to continue work on this > > > > Internet-Draft through correspondence on the Q12/15 mailing list, and > > > > thank Q12/15 for assigning Ben Mack-Crane as the SG15 representative > > on > > > > this matter. Ben has proved very helpful in clarifying and developing > > > > comments received from SG15. > > > > > > > > > > > > We have had several email exchanges with Ben about a few of the > > specific > > > > comments made in your review of the 02 revision of the Internet-Draft. > > > > Although he has taken these comments to the Q12/15 mailing list no > > > > response has been received, perhaps due to pressure of time. We are > > taking > > > > the opportunity to restate several questions here for your convenience > > and > > > > would appreciate your answers as part of the response to this liaison. > > > > > > > > 1. Definition of a Resource (section 3.2) > > > > You have supplied us with the simple text that says... > > > > In ASON applications: > > > > A resource refers only to a link connection (e.g. VC-11, > > > > VC-4 etc.). > > > > > > > > We would like to clarify this because we don't understand how > > > > the original text is wrong. It used to say... > > > > ITU-T [should be ASON] terms for resource: > > > > - Connection point (cp) in the context of link discovery > > > > and resource management (allocation, binding into > > > > cross-connects, etc.); > > > > - Link connection or trail termination in the context of > > > > routing, path computation and signaling. > > > > > > > > Recall that we are talking about the ASON terms for what GMPLS > > > > calls a resource. We think the following... > > > > In ASON a link connection is an association of output cp on one > > > > side of the link and input cp on the other side of the link. In > > > > GMPLS by resource we mean in most of the cases the local end of > > > > the resource (which is cp in ASON terminology). This is true in > > > > the context of link discovery. > > > > > > > > More importantly, this is also true in the context of a > > > > signaling application. Specifically, when a GMPLS signaling > > > > sub-system requests allocation of a resource, the GMPLS > > > > Resource Manager allocates only the local end of the > > > > resource. This happens on both sides of the link, that is, we > > > > allocate input and output resources. This contrasts with ASON, > > > > where there is a much more versatile LRM, and the connection > > > > manager (signaling application) allocates link connections > > > > only on the output side, leaving LRM to coordinate the > > > > allocation with the neighbor. > > > > > > > > However, in GMPLS TE advertisements we account only for > > > > outgoing resources, that is, we don't advertise incoming > > > > resources on the assumption that they match the outgoing > > > > resources for the discovered TE links. > > > > > > > > Bottom line: we believe that we should retain the previous > > > > definitions, and have not made this change in the new > > > > revision. We would like to discuss this point further with > > > > Q12/15. > > > > > > > > 2. Link ends and Data links (Section 3.5 - was section 3.4) > > > > > > > > 2a.Although we understand that ASON does not require the concept > > > > of a link end, GMPLS does. Therefore we should provide some > > > > form of language to help people do the mapping. > > > > > > > > The proposed new text from Q12/15 removes all reference to > > > > link ends from the ASON section, and we feel this is a > > > > mistake and propose the following text: > > > > In ASON, a GMPLS unidirectional data link end is a > > > > collection of connection points from the same client > > > > layer that are supported by a single trail termination > > > > (access point). > > > > > > > > This text depends on the fact that a link end is described as > > > > a set of resources that transfer traffic to a neighbor. > > > > > > > > 2b.The new text from Q12/15 gives us... > > > > A GMPLS data link is an ASON link supported by a single > > > > server trail. > > > > > > > > We are not sure that we understand this correctly. It > > > > appears to say that a unidirectional data link could be > > > > supported by multiple trail terminations (access points). > > > > This seems to suggest that a client would have to make a > > > > choice about where to send data, which seems to imply it > > > > really has to choose between data links! We are confused! > > > > > > > > We think some of this problem may come from the *need* in > > > > GMPLS to identify data link ends. There may be some value in > > > > stressing that a GMPLS link end is supported by exactly one > > > > trail termination. When several trail terminations are > > > > involved, we can still talk about single TE link (as a > > > > bundle), but each component link will be a distinct data > > > > link. We think that in ASON the SNPP (TE link in GMPLS > > > > language) could be supported by multiple terminations. > > > > > > > > 3. Link interfaces (section 3.6 - was 3.5) > > > > The new text from Q12/15 gives us... > > > > A GMPLS interface is "all the stuff between a physical > > > > interface and a matrix in an ASON network element." More > > > > precisely, it is a trail termination function and adaptation > > > > function for which adapted client layer connection points > > > > are bound to a matrix. > > > > > > > > We think "physical interface" may be ambiguous in the multi- > > > > layer context. We think the "stuff" should be between a link > > > > and a matrix, since we are talking about links in distinct > > > > layers. > > > > > > > > > > > > In addition to your feedback on the specific points above, we would > > > > welcome your continued comments and review of the Internet-Draft. > > > > > > > > Regards, > > > > Adrian Farrel and Kireeti Kompella > > > > IETF CCAMP Working Group Co-chairs > > > > > > > > > > > > > > > > > > > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 21 Jul 2005 10:11:00 +0000 Message-ID: <42DF7447.8070103@kddilabs.jp> Date: Thu, 21 Jul 2005 19:09:11 +0900 From: Tomohiro Otani <otani@kddilabs.jp> User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; ja-JP; rv:1.7.2) Gecko/20040803 MIME-Version: 1.0 To: ccamp@ops.ietf.org Subject: draft-otani-ccamp-interas-gmpls-te-03.txt Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi everyone, We updated the GMPLS inter-domain TE req. draft. Since last meeting, we believe this document has been already stable and well-recognized, however, per the last meeting discussion, we have to wait for the charter being revised, in order to be a WG document (This must be initiated by chairs). Until then, we continue to update the contents and indeed welcome your feedback and comments through this list or during the meeting in Paris. Regards, tomo ----------------------------------------------------------------------------- A New Internet-Draft is available from the on-line Internet-Drafts directories. Title : GMPLS Inter-domain Traffic Engineering Requirements Author(s) : T. Otani, et al. Filename : draft-otani-ccamp-interas-gmpls-te-03.txt Pages : 15 Date : 2005-7-20 This draft provides requirements for the support of generalized multi-protocol label switching (GMPLS) inter-domain traffic engineering (TE). Its main objective is to present the differences between MPLS inter-domain TE and GMPLS inter-domain TE. This draft covers not only GMPLS Inter-domain architecture but also functional requirements in terms of GMPLS signaling and routing in order to specify these in a GMPLS Inter-domain environment. A URL for this Internet-Draft is: http://www.ietf.org/internet-drafts/draft-otani-ccamp-interas-gmpls-te-03.txt To remove yourself from the I-D Announcement list, send a message to i-d-announce-request at ietf.org with the word unsubscribe in the body of the message. You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce to change your subscription settings. Internet-Drafts are also available by anonymous FTP. Login with the username "anonymous" and a password of your e-mail address. After logging in, type "cd internet-drafts" and then "get draft-otani-ccamp-interas-gmpls-te-03.txt". A list of Internet-Drafts directories can be found in http://www.ietf.org/shadow.html or ftp://ftp.ietf.org/ietf/1shadow-sites.txt Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 21 Jul 2005 00:31:57 +0000 Message-ID: <42DEEC92.1070309@lab.ntt.co.jp> Date: Thu, 21 Jul 2005 09:30:10 +0900 From: Kohei Shiomoto <shiomoto.kohei@lab.ntt.co.jp> User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; ja-JP; rv:1.7.7) Gecko/20050414 MIME-Version: 1.0 To: ccamp@ops.ietf.org Subject: [Fwd: I-D ACTION:draft-oki-ccamp-gmpls-ip-interworking-06.txt] Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi all A new version of the MPLS-GMPLS interwork and migration draft has been uploaded. "IP/MPLS-GMPLS interworking in support of IP/MPLS to GMPLS migration" http://www.ietf.org/internet-drafts/draft-oki-ccamp-gmpls-ip-interworking-06.txt Summary of change: Section 2 (New) - Addressed interworking MPLS-TE and GMPLS "Three issues on interworking between MPLS-based packet networks and GMPLS-based optical transport network result from the fact that control and data planes are separated in GMPLS-based optical transport networks. These three issues are (1) lack of routing and signaling adjacencies, (2) control plane resource exhaustion, and (3) TE path computation over the border between MPLS and GMPLS domains." Section 3 - Added 3.7 Integrated migration model Section 5 - Removed 4.2.1 Segment stitching of the previous version - Moved 5.1.2 Discovery of GMPLS signaling capability from 4.2.4 of the previous version. - Added 5.3 Path computation Throughout the document - Cleaned up text for readability (Removed redundancy, etc) Comments and feedbacks are highly appreciated. Best regards, -- Kohei Shiomoto -------- Original Message -------- Subject: I-D ACTION:draft-oki-ccamp-gmpls-ip-interworking-06.txt Date: Tue, 19 Jul 2005 10:50:01 -0400 From: Internet-Drafts@ietf.org Reply-To: internet-drafts@ietf.org To: i-d-announce@ietf.org A New Internet-Draft is available from the on-line Internet-Drafts directories. Title : IP/MPLS-GMPLS interworking in support of IP/MPLS to GMPLS migration Author(s) : E. Oki, et al. Filename : draft-oki-ccamp-gmpls-ip-interworking-06.txt Pages : 17 Date : 2005-7-19 Over time Multi-Protocol Label Switching (MPLS) traffic engineered networks may be migrated to use Generalized MPLS (GMPLS). This will allow the networks to benefit from many new features that have been added to the GMPLS protocols. Additionally, such migration will facilitate easy interworking of packet networks that are controlled by IP/MPLS-TE protocols and networks that are controlled by GMPLS protocols. During this migration phase MPLS and GMPLS capable elements will have to co-exist and interwork. GMPLS signaling and routing protocols are subtly different from the MPLS protocols and so interworking and migration strategies are needed. This document describes issues associated with interworking of IP/MPLS-TE networks and GMPLS-based optical networks. Then it describes possible migration scenarios, mechanisms to compensate for the differences between MPLS and GMPLS protocols, and how to apply those mechanisms in order to achieve migration from MPLS to GMPLS. A URL for this Internet-Draft is: http://www.ietf.org/internet-drafts/draft-oki-ccamp-gmpls-ip-interworking-06.txt To remove yourself from the I-D Announcement list, send a message to i-d-announce-request@ietf.org with the word unsubscribe in the body of the message. You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce to change your subscription settings. Internet-Drafts are also available by anonymous FTP. Login with the username "anonymous" and a password of your e-mail address. After logging in, type "cd internet-drafts" and then "get draft-oki-ccamp-gmpls-ip-interworking-06.txt". A list of Internet-Drafts directories can be found in http://www.ietf.org/shadow.html or ftp://ftp.ietf.org/ietf/1shadow-sites.txt Internet-Drafts can also be obtained by e-mail. Send a message to: mailserv@ietf.org. In the body type: "FILE /internet-drafts/draft-oki-ccamp-gmpls-ip-interworking-06.txt". NOTE: The mail server at ietf.org can return the document in MIME-encoded form by using the "mpack" utility. To use this feature, insert the command "ENCODING mime" before the "FILE" command. To decode the response(s), you will need "munpack" or a MIME-compliant mail reader. Different MIME-compliant mail readers exhibit different behavior, especially when dealing with "multipart" MIME messages (i.e. documents which have been split up into multiple messages), so check your local documentation on how to manipulate these messages. Below is the data which will enable a MIME compliant mail reader implementation to automatically retrieve the ASCII version of the Internet-Draft. -- Kohei Shiomoto, Ph.D NTT Network Service Systems Laboratories 3-9-11 Midori, Musashino, Tokyo 180-8585, Japan Phone +81 422 59 4402 Fax +81 422 59 3787 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 20:45:00 +0000 Message-ID: <030501c58d63$5464d460$0601a8c0@pc6> Reply-To: "Tom Petch" <nwnetworks@dial.pipex.com> From: "Tom Petch" <nwnetworks@dial.pipex.com> To: "Adrian Farrel" <adrian@olddog.co.uk>, <ccamp@ops.ietf.org> Subject: Re: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON Lexicography Date: Wed, 20 Jul 2005 21:41:51 +0200 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit <inline> Tom Petch ----- Original Message ----- From: "Adrian Farrel" <adrian@olddog.co.uk> To: "Tom Petch" <nwnetworks@dial.pipex.com>; <ccamp@ops.ietf.org> Sent: Wednesday, July 20, 2005 11:54 AM Subject: Re: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON Lexicography > Tom, > > I don't think it is CCAMP's role to correct the English of another SDO :-) > > Adrian OK, but then I do not think you should reply "Correct" because IMHO it is not correct; the response should be something about what, primarily, CCAMP is responsible for:-) > ----- Original Message ----- > From: "Tom Petch" <nwnetworks@dial.pipex.com> > To: "Adrian Farrel" <adrian@olddog.co.uk>; <ccamp@ops.ietf.org> > Sent: Tuesday, July 19, 2005 7:18 PM > Subject: Re: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON > Lexicography > > > > Adrian > > > > I think it should be > > " > > - CCAMP has primary responsibility for the descriptions of GMPLS terms > > > > Correct " > > > > even if that is not what SG15 said. > > > > Primarily (sic) CCAMP has responsibility for the specification of GMPLS, > not > > just its terms:-) > > > > Tom Petch > > > > ----- Original Message ----- > > From: "Adrian Farrel" <adrian@olddog.co.uk> > > To: <ccamp@ops.ietf.org> > > Cc: "'Kireeti Kompella'" <kireeti@juniper.net>; <zinin@psg.com>; "Bill > Fenner" > > <fenner@research.att.com>; "Scott Bradner" <sob@harvard.edu> > > Sent: Tuesday, July 19, 2005 1:59 PM > > Subject: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON > > Lexicography > > > > > > > Hi, > > > > > > Please comment on this draft response liaison which I intend to send > on > > > Monday 25th July. > > > You can see the original incoming liaison at > > > http://www.olddog.co.uk/incoming.htm > > > > > > Thanks, > > > Adrian > > > ===== > > > To: ITU-T SG15 Questions 12 and 14 > > > From: IETF CCAMP Working Group > > > Subject: Response to your liaison on GMPLS/ASON Lexicography > > > For: Action > > > Deadline: August 31st 2005 > > > > > > The CCAMP Working Group would like to thank SG15 and especially > Questions > > > 12 and 14 for the time and effort they put into providing input and > > > feedback on the GMPLS/ASON lexicography Internet-Draft that is being > > > developed by CCAMP. Much of the material supplied as direct comments > or as > > > mark-ups to the Internet-Draft has been incorporated into the latest > > > revision which is available at > > > > > > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-gmpls-ason-lexicography-03. > > txt > > > > > > In your liaison you state some assumptions which we would like to > comment > > > on as follows. > > > > > > - The purpose of the document is to enable those familiar with ASON > > > to understand GMPLS terminology in an ASON context > > > > > > Correct. > > > > > > - GMPLS terms are described informally in the document in a way that > > > is generally consistent with GMPLS (existing RFCs and work in > progress) > > > > > > While the definition of GMPLS terms would be helpful for understanding > > > GMPLS terms in a GMPLS context, that is not the purpose of this > document. > > > > > > - Based on these descriptions, interpretations of these GMPLS terms > > > in ASON terminology are provided > > > > > > It is intended that the interpretation of the GMPLS terms is not based > on > > > these descriptions, but based on the full meaning of the GMPLS terms. > The > > > descriptions of the GMPLS terms are provided in this Internet-Draft > for > > > context and a brief summary. > > > > > > - CCAMP has primarily responsibility for the descriptions of GMPLS > terms > > > > > > Correct. > > > > > > - Q12/15 has responsibility to provide input to the GMPLS-ASON > > > interpretations (based on the descriptions in this draft) > > > > > > We welcome this assumption of responsibility by Q12/15 to provide > input to > > > the Internet-Draft. While the Internet-Draft remains under the overall > > > editorial control of the CCAMP working group, we hope to be able to > use > > > the text supplied by Q12/15 with only editorial changes. > > > > > > > > > Additionally, we welcome the pledge to continue work on this > > > Internet-Draft through correspondence on the Q12/15 mailing list, and > > > thank Q12/15 for assigning Ben Mack-Crane as the SG15 representative > on > > > this matter. Ben has proved very helpful in clarifying and developing > > > comments received from SG15. > > > > > > > > > We have had several email exchanges with Ben about a few of the > specific > > > comments made in your review of the 02 revision of the Internet-Draft. > > > Although he has taken these comments to the Q12/15 mailing list no > > > response has been received, perhaps due to pressure of time. We are > taking > > > the opportunity to restate several questions here for your convenience > and > > > would appreciate your answers as part of the response to this liaison. > > > > > > 1. Definition of a Resource (section 3.2) > > > You have supplied us with the simple text that says... > > > In ASON applications: > > > A resource refers only to a link connection (e.g. VC-11, > > > VC-4 etc.). > > > > > > We would like to clarify this because we don't understand how > > > the original text is wrong. It used to say... > > > ITU-T [should be ASON] terms for resource: > > > - Connection point (cp) in the context of link discovery > > > and resource management (allocation, binding into > > > cross-connects, etc.); > > > - Link connection or trail termination in the context of > > > routing, path computation and signaling. > > > > > > Recall that we are talking about the ASON terms for what GMPLS > > > calls a resource. We think the following... > > > In ASON a link connection is an association of output cp on one > > > side of the link and input cp on the other side of the link. In > > > GMPLS by resource we mean in most of the cases the local end of > > > the resource (which is cp in ASON terminology). This is true in > > > the context of link discovery. > > > > > > More importantly, this is also true in the context of a > > > signaling application. Specifically, when a GMPLS signaling > > > sub-system requests allocation of a resource, the GMPLS > > > Resource Manager allocates only the local end of the > > > resource. This happens on both sides of the link, that is, we > > > allocate input and output resources. This contrasts with ASON, > > > where there is a much more versatile LRM, and the connection > > > manager (signaling application) allocates link connections > > > only on the output side, leaving LRM to coordinate the > > > allocation with the neighbor. > > > > > > However, in GMPLS TE advertisements we account only for > > > outgoing resources, that is, we don't advertise incoming > > > resources on the assumption that they match the outgoing > > > resources for the discovered TE links. > > > > > > Bottom line: we believe that we should retain the previous > > > definitions, and have not made this change in the new > > > revision. We would like to discuss this point further with > > > Q12/15. > > > > > > 2. Link ends and Data links (Section 3.5 - was section 3.4) > > > > > > 2a.Although we understand that ASON does not require the concept > > > of a link end, GMPLS does. Therefore we should provide some > > > form of language to help people do the mapping. > > > > > > The proposed new text from Q12/15 removes all reference to > > > link ends from the ASON section, and we feel this is a > > > mistake and propose the following text: > > > In ASON, a GMPLS unidirectional data link end is a > > > collection of connection points from the same client > > > layer that are supported by a single trail termination > > > (access point). > > > > > > This text depends on the fact that a link end is described as > > > a set of resources that transfer traffic to a neighbor. > > > > > > 2b.The new text from Q12/15 gives us... > > > A GMPLS data link is an ASON link supported by a single > > > server trail. > > > > > > We are not sure that we understand this correctly. It > > > appears to say that a unidirectional data link could be > > > supported by multiple trail terminations (access points). > > > This seems to suggest that a client would have to make a > > > choice about where to send data, which seems to imply it > > > really has to choose between data links! We are confused! > > > > > > We think some of this problem may come from the *need* in > > > GMPLS to identify data link ends. There may be some value in > > > stressing that a GMPLS link end is supported by exactly one > > > trail termination. When several trail terminations are > > > involved, we can still talk about single TE link (as a > > > bundle), but each component link will be a distinct data > > > link. We think that in ASON the SNPP (TE link in GMPLS > > > language) could be supported by multiple terminations. > > > > > > 3. Link interfaces (section 3.6 - was 3.5) > > > The new text from Q12/15 gives us... > > > A GMPLS interface is "all the stuff between a physical > > > interface and a matrix in an ASON network element." More > > > precisely, it is a trail termination function and adaptation > > > function for which adapted client layer connection points > > > are bound to a matrix. > > > > > > We think "physical interface" may be ambiguous in the multi- > > > layer context. We think the "stuff" should be between a link > > > and a matrix, since we are talking about links in distinct > > > layers. > > > > > > > > > In addition to your feedback on the specific points above, we would > > > welcome your continued comments and review of the Internet-Draft. > > > > > > Regards, > > > Adrian Farrel and Kireeti Kompella > > > IETF CCAMP Working Group Co-chairs > > > > > > > > > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 19:51:57 +0000 Content-Type: Multipart/Mixed; Boundary="NextPart" Mime-Version: 1.0 To: i-d-announce@ietf.org Cc: ccamp@ops.ietf.org From: Internet-Drafts@ietf.org Subject: I-D ACTION:draft-ietf-ccamp-gmpls-rsvp-te-ason-04.txt Message-Id: <E1DvKZv-0007mh-RT@newodin.ietf.org> Date: Wed, 20 Jul 2005 15:50:03 -0400 --NextPart A New Internet-Draft is available from the on-line Internet-Drafts directories. This draft is a work item of the Common Control and Measurement Plane Working Group of the IETF. Title : Generalized MPLS (GMPLS) RSVP-TE Signalling in support of Automatically Switched Optical Network (ASON) Author(s) : J. Drake, et al. Filename : draft-ietf-ccamp-gmpls-rsvp-te-ason-04.txt Pages : 40 Date : 2005-7-20 This document specifies how Generalized MPLS (GMPLS) RSVP-TE signaling may be used and extended to satisfy the requirements of the Automatically Switched Optical Network (ASON) architecture specified by the ITU-T. The requirements are in a companion document "Requirements for Generalized MPLS (GMPLS) Usage and Extensions for Automatically Switched Optical Network (ASON)." In particular, this document details the mechanisms for setting up Soft Permanent Connections (SPC), the necessary extensions in delivering full and logical call/connection separation support, the extended restart capabilities during control plane failures, extended label usage and crankback signalling capability. The mechanisms proposed in this document are applicable to any environment (including multi-area) and for any type of interface: packet, layer-2, time-division multiplexed, lambda or fiber switching. A URL for this Internet-Draft is: http://www.ietf.org/internet-drafts/draft-ietf-ccamp-gmpls-rsvp-te-ason-04.txt To remove yourself from the I-D Announcement list, send a message to i-d-announce-request@ietf.org with the word unsubscribe in the body of the message. You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce to change your subscription settings. Internet-Drafts are also available by anonymous FTP. Login with the username "anonymous" and a password of your e-mail address. After logging in, type "cd internet-drafts" and then "get draft-ietf-ccamp-gmpls-rsvp-te-ason-04.txt". A list of Internet-Drafts directories can be found in http://www.ietf.org/shadow.html or ftp://ftp.ietf.org/ietf/1shadow-sites.txt Internet-Drafts can also be obtained by e-mail. Send a message to: mailserv@ietf.org. In the body type: "FILE /internet-drafts/draft-ietf-ccamp-gmpls-rsvp-te-ason-04.txt". NOTE: The mail server at ietf.org can return the document in MIME-encoded form by using the "mpack" utility. To use this feature, insert the command "ENCODING mime" before the "FILE" command. To decode the response(s), you will need "munpack" or a MIME-compliant mail reader. Different MIME-compliant mail readers exhibit different behavior, especially when dealing with "multipart" MIME messages (i.e. documents which have been split up into multiple messages), so check your local documentation on how to manipulate these messages. Below is the data which will enable a MIME compliant mail reader implementation to automatically retrieve the ASCII version of the Internet-Draft. --NextPart Content-Type: Multipart/Alternative; Boundary="OtherAccess" --OtherAccess Content-Type: Message/External-body; access-type="mail-server"; server="mailserv@ietf.org" Content-Type: text/plain Content-ID: <2005-7-20135038.I-D@ietf.org> ENCODING mime FILE /internet-drafts/draft-ietf-ccamp-gmpls-rsvp-te-ason-04.txt --OtherAccess Content-Type: Message/External-body; name="draft-ietf-ccamp-gmpls-rsvp-te-ason-04.txt"; site="ftp.ietf.org"; access-type="anon-ftp"; directory="internet-drafts" Content-Type: text/plain Content-ID: <2005-7-20135038.I-D@ietf.org> --OtherAccess-- --NextPart-- Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 19:16:05 +0000 DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=Message-ID:Received:Date:From:Subject:To:Cc:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding; b=MdnaU2ljFs7PgYpVIAhMmTTgkBIU25FifH6Q6DQOWxEG3AlU0LA4sH1ygSm8PYhqBN0CSIEaA5bYrD8xK5eZtAjLCzKPe3qZaG0iilsltOGTiprSA95qfjZ8lovSgNLSjNYUhpc7l8VwMOLkRu/NTsdUKou/jk088eHq5vimg10= ; Message-ID: <20050720191546.95622.qmail@web30803.mail.mud.yahoo.com> Date: Wed, 20 Jul 2005 12:15:46 -0700 (PDT) From: Igor Bryskin <i_bryskin@yahoo.com> Subject: Re: LSP stitching questions for discussion To: Dimitri.Papadimitriou@alcatel.be, Igor Bryskin <ibryskin@movaz.com> Cc: Dimitri.Papadimitriou@alcatel.be, Arthi Ayyangar <arthi@juniper.net>, ccamp@ops.ietf.org MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="0-499060946-1121886946=:94207" Content-Transfer-Encoding: 8bit --0-499060946-1121886946=:94207 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit [dp] ok - last try for an end-to-end PSC-1 LSP of 2MB you could make use of a PSC-1 segment of 4MB Ah, Ok. I see your point and agree. Igor You are right: whether PSC1 and PSC2 are separate network layers or not is a topic of a separate (painful :=)) discussion. However, I just wanted to point out here that a PSC2 LSP is a provisioned as such for a reason: it is meant to *nest* PSC1 LSPs rather than to stitch them. From the VNT management point of view PSC2 and PSC1 are separate topological layers. that is PSC2 produces topology for PSC1, just like VC4 produces topology for VC12. The only difference is that in the latter case this relationship is forced by the hardware, and in the former case is not (just logical). Take for example C and C++ programming. Certain things you simply forced not to do in C++, whereas in C you could but shouldn't. See you in Paris. Igor "Igor Bryskin" <ibryskin@movaz.com> Sent by: owner-ccamp@ops.ietf.org 07/20/2005 13:25 AST To: "Arthi Ayyangar" <arthi@juniper.net>, Dimitri PAPADIMITRIOU/BE/ALCATEL@ALCATEL cc: <ccamp@ops.ietf.org> bcc: Subject: Re: LSP stitching questions for discussion Dimitri, It seems to me that I mostly agree(?!) with what you wrote in this mail. I'd still want to confirm your views on the following questions: a) Is it possible (should we allow) to stitch PSC1 e2e LSP by a LSC LSP segment ? b) Is it possible (should we allow) to stitch VC12 e2e LSP by a VC4 LSP segment ? c) Is it possible (should we allow) to stitch PSC1 e2e LSP by a PSC2 LSP segment ? My answers are NO to all three questions. You can sticth PSC1 with PSC1, VC12 with VC12, etc. In all other cases you need, first, to setup a hierarchical LSP (H-LSP) (whether it is FA-LSP (single domain) or not (multi-domain)) in order to create a data link of the e2e LSP switching capability and only after that you can also use a stitching segment on top of this H-LSP). See also in-line. Igor ----- Original Message ----- From: Dimitri.Papadimitriou@alcatel.be To: Arthi Ayyangar Cc: Igor Bryskin ; ccamp@ops.ietf.org Sent: Wednesday, July 20, 2005 4:46 AM Subject: Re: LSP stitching questions for discussion arthi, first a recap: this item on stitching was initiated in order to allow per "domain" control of end-to-end (contiguous) LSP, as part of a separate document, and not to extend its scope (unduly, in this case, see below) but mainly in order to have the set of all related protocol mechanisms within a specific document - from this initial working assumption, it is clear that the segment must have the properties of the incoming end-to-end LSP which are determined among other by its switching type e.g. PSC-3 or TDM, the first requirement of LSP stitching is contiguity of the end-to-end LSP at the data plane level (after stitching operation) - and this by definition of the initial working assumption - IB>> Agree note: the only latitude which is left (for a given switching type e.g. PSC-2) is due to the properties of packet (PSC) networks that allows you stitch a segment having a larger Max Reservable Bandwidth and Unreserved Bandwidth (we would have equivalent latitude for other attributes but this is a local policy decision - which at the end is the sole purpose of this approach) IB>> Disagree. Conceptually PCS1 and PSC2 are different layers and the latter cannot stitch the former, just like in case VC12 and VC4. You should nest PSC1 within PSC2 instead. Here are some of the questions related to LSP stitching to start a discussion on the usage of LSP stitching. >1) Are the control plane signaling procedures for LSP stitching as >described in this ID, (explicit request for LSP stitching on >LSP segment, different label allocation rules; etc) REQUIRED to stitch >packet LSPs in data plane -> yes (beside manual config operation) IB>> Agree >2) Are the control plane signaling procedures for LSP stitching as >described in this ID, (explicit request for LSP stitching on >LSP segment, different label allocation rules; etc) REQUIRED to stitch >non-packet LSPs in data plane ? -> yes (beside manual config operation) IB>> Agree >3) If an e2e LSP crosses a region boundary (based on different TE >link switching capabilities), then can I use LSP stitching >a) control plane procedures for packet LSPs >b) data plane procedures for packet LSPs >c) control plane procedures for non-packet LSPs >d) data plane procedures for non-packet LSPs -> no, when crossing a region boundary, you simply use a document named "draft-mpls-lsp-hierarchy-08.txt" which is in the RFC Editor queue, but this does not prevent from creating an LSP segment on top of an FA-LSP (see below) IB>> Agree with the caviat that it should be H-LSP instead of FA-LSP to cover single and multi-domain scenarios >4) If the switching type of an e2e LSP is different from that of an LSP >segment, can I use LSP stitching >a) control plane procedures for packet LSPs >b) data plane procedures for packet LSPs >c) control plane procedures for non-packet LSPs >d) data plane procedures for non-packet LSPs -> no, in this case progression of the end-to-end LSP establishment, implies that the LSP segment must fulfil a set of constraints carried as part of the incoming end-to-end LSP request, in this case one has to create an LSP segment (with a switching type equal to the incoming end-to-end LSP switching type) over that FA (i.e. referred here above to the LSP segment with a different Switching Type) IB>> Agree with the caviat that it should be H-LSP instead of FA-LSP to cover single and multi-domain scenarios Cheers, Igor >Note that although 3) and 4) appear to be the same, the reason I have >them as different questions is that in one case the TE link switching >capability is examined and in the other case the Generalized Label Request >is examined. -> what it appears to me is that this discussion mixes the issue of the operation describing the stitching operations at the domain boundaries from the LSP segment creation itself i.e. an LSP segment can indeed be an FA link inherited from the creation of an (intra-domain) FA-LSP - __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com --0-499060946-1121886946=:94207 Content-Type: text/html; charset=iso-8859-1 Content-Transfer-Encoding: 8bit <DIV> <BLOCKQUOTE class=replbq style="PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #1010ff 2px solid"> <P>[dp] ok - last try for an end-to-end PSC-1 LSP of 2MB you could make use of a PSC-1 segment of 4MB </P> <P>Ah, Ok. I see your point and agree.</P> <P>Igor</P> <P><BR>You are right: whether PSC1 and PSC2 are separate network layers or not is a topic of a separate (painful :=)) discussion. However, I just wanted to point out here that a PSC2 LSP is a provisioned as such for a reason: it is meant to *nest* PSC1 LSPs rather than to stitch them. From the VNT management point of view PSC2 and PSC1 are separate topological layers. that is PSC2 produces topology for PSC1, just like VC4 produces topology for VC12.<BR>The only difference is that in the latter case this relationship is forced by the hardware, and in the former case is not (just logical). Take for example C and C++ programming. Certain things you simply forced not to do in C++, whereas in C you could but shouldn't.<BR><BR>See you in Paris.<BR>Igor<BR><FONT size=4></FONT><BR><BR><BR><BR><B>"Igor Bryskin" <ibryskin@movaz.com></B><BR>Sent by: owner-ccamp@ops.ietf.org<BR>07/20/2005 13:25 AST<BR><BR>To:<FONT size=4> </FONT>"Arthi Ayyangar" <arthi@juniper.net>, Dimitri PAPADIMITRIOU/BE/ALCATEL@ALCATEL<BR>cc:<FONT size=4> </FONT><ccamp@ops.ietf.org><BR>bcc:<FONT size=4> </FONT><BR>Subject:<FONT size=4> </FONT>Re: LSP stitching questions for discussion<BR><BR><BR><FONT size=4>Dimitri,</FONT><BR><BR><FONT size=4>It seems to me that I mostly agree(?!) with what you wrote in this mail. I'd still want to confirm your views on the following questions:</FONT><BR><BR><BR><FONT size=4>a) Is it possible (should we allow) to stitch PSC1 e2e LSP by a LSC LSP segment ? </FONT><BR><BR><FONT size=4>b) Is it possible (should we allow) to stitch VC12 e2e LSP by a VC4 LSP segment ? </FONT><BR><FONT size=4>c) Is it possible (should we allow) to stitch PSC1 e2e LSP by a PSC2 LSP segment ?</FONT><BR><BR><FONT size=4>My answers are NO to all three questions. You can sticth PSC1 with PSC1, VC12 with VC12, etc. In all other cases you need, first, to setup a hierarchical LSP (H-LSP) (whether it is FA-LSP (single domain) or not (multi-domain)) in order to create a data link of the e2e LSP switching capability and only after that you can also use a stitching segment on top of this H-LSP).</FONT><BR><BR><FONT size=4>See also in-line.</FONT><BR><BR><FONT size=4>Igor</FONT><BR><BR><BR><FONT size=5>----- Original Message ----- </FONT><BR><FONT size=5><B>From:</B> </FONT><FONT color=blue size=5><U><A href="mailto:Dimitri.Papadimitriou@alcatel.be">Dimitri.Papadimitriou@alcatel.be</A></U></FONT><FONT color=black size=5> </FONT><BR><FONT size=5><B>To:</B> </FONT><FONT color=blue size=5><U><A href="mailto:arthi@juniper.net">Arthi Ayyangar</A></U></FONT><FONT color=black size=5> </FONT><BR><FONT size=5><B>Cc:</B> </FONT><FONT color=blue size=5><U><A href="mailto:ibryskin@movaz.com">Igor Bryskin</A></U></FONT><FONT color=black size=5> ; </FONT><FONT color=blue size=5><U><A href="mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A></U></FONT><FONT color=black size=5> </FONT><BR><FONT size=5><B>Sent:</B> Wednesday, July 20, 2005 4:46 AM</FONT><BR><FONT size=5><B>Subject:</B> Re: LSP stitching questions for discussion</FONT><BR><BR><FONT size=5>arthi, </FONT><BR><BR><BR><BR><FONT size=5>first a recap: this item on stitching was initiated in order to allow per "domain" control of end-to-end (contiguous) LSP, as part of a separate document, and not to extend its scope (unduly, in this case, see below) but mainly in order to have the set of all related protocol mechanisms within a specific document -</FONT><BR><BR><BR><BR><FONT size=5>from this initial working assumption, it is clear that the segment must have the properties of the incoming end-to-end LSP which are determined among other by its switching type e.g. PSC-3 or TDM, the first requirement of LSP stitching is contiguity of the end-to-end LSP at the data plane level (after stitching operation) - and this by definition of the initial working assumption -</FONT><BR><BR><FONT size=5>IB>> Agree</FONT><BR><BR><FONT size=5>note: the only latitude which is left (for a given switching type e.g. PSC-2) is due to the properties of packet (PSC) networks that allows you stitch a segment having a larger Max Reservable Bandwidth and Unreserved Bandwidth (we would have equivalent latitude for other attributes but this is a local policy decision - which at the end is the sole purpose of this approach)</FONT><BR><BR><FONT size=5>IB>> Disagree. Conceptually PCS1 and PSC2 are different layers and the latter cannot stitch the former, just like in case VC12 and VC4. You should nest PSC1 within PSC2 instead.</FONT><BR><BR><BR><BR><FONT size=5>Here are some of the questions related to LSP stitching to start a</FONT><BR><FONT size=5>discussion on the usage of LSP stitching.</FONT><BR><BR><FONT size=5>>1) Are the control plane signaling procedures for LSP stitching as</FONT><BR><FONT size=5>>described in this ID, (explicit request for LSP stitching on</FONT><BR><FONT size=5>>LSP segment, different label allocation rules; etc) REQUIRED to stitch</FONT><BR><FONT size=5>>packet LSPs in data plane </FONT><BR><BR><BR><FONT size=5>-> yes (beside manual config operation)</FONT><BR><BR><FONT size=5>IB>> Agree</FONT><BR><BR><BR><FONT size=5>>2) Are the control plane signaling procedures for LSP stitching as</FONT><BR><FONT size=5>>described in this ID, (explicit request for LSP stitching on</FONT><BR><FONT size=5>>LSP segment, different label allocation rules; etc) REQUIRED to stitch</FONT><BR><FONT size=5>>non-packet LSPs in data plane ?</FONT><BR><BR><BR><FONT size=5>-> yes (beside manual config operation)</FONT><BR><BR><FONT size=5>IB>> Agree</FONT><BR><BR><BR><FONT size=5>>3) If an e2e LSP crosses a region boundary (based on different TE</FONT><BR><FONT size=5>>link switching capabilities), then can I use LSP stitching </FONT><BR><FONT size=5>>a) control plane procedures for packet LSPs</FONT><BR><FONT size=5>>b) data plane procedures for packet LSPs</FONT><BR><FONT size=5>>c) control plane procedures for non-packet LSPs</FONT><BR><FONT size=5>>d) data plane procedures for non-packet LSPs</FONT><BR><BR><BR><FONT size=5>-> no, when crossing a region boundary, you simply use a document named "draft-mpls-lsp-hierarchy-08.txt" which is in the RFC Editor queue, but this does not prevent from creating an LSP segment on top of an FA-LSP (see below)</FONT><BR><BR><FONT size=5>IB>> Agree with the caviat that it should be H-LSP instead of FA-LSP to cover single and multi-domain scenarios</FONT><BR><BR><BR><FONT size=5>>4) If the switching type of an e2e LSP is different from that of an LSP</FONT><BR><FONT size=5>>segment, can I use LSP stitching </FONT><BR><FONT size=5>>a) control plane procedures for packet LSPs</FONT><BR><FONT size=5>>b) data plane procedures for packet LSPs</FONT><BR><FONT size=5>>c) control plane procedures for non-packet LSPs</FONT><BR><FONT size=5>>d) data plane procedures for non-packet LSPs</FONT><BR><BR><BR><FONT size=5>-> no, in this case progression of the end-to-end LSP establishment, implies that the LSP segment must fulfil a set of constraints carried as part of the incoming end-to-end LSP request, in this case one has to create an LSP segment (with a switching type equal to the incoming end-to-end LSP switching type) over that FA (i.e. referred here above to the LSP segment with a different Switching Type)</FONT><BR><BR><BR><FONT size=5>IB>> Agree with the caviat that it should be H-LSP instead of FA-LSP to cover single and multi-domain scenarios</FONT><BR><BR><FONT size=4>Cheers,</FONT><BR><BR><FONT size=4>Igor</FONT><BR><BR><FONT size=5>>Note that although 3) and 4) appear to be the same, the reason I have</FONT><BR><FONT size=5>>them as different questions is that in one case the TE link switching</FONT><BR><FONT size=5>>capability is examined and in the other case the Generalized Label Request</FONT><BR><FONT size=5>>is examined.</FONT><BR><BR><BR><BR><FONT size=5>-> what it appears to me is that this discussion mixes the issue of the operation describing the stitching operations at the domain boundaries from the LSP segment creation itself i.e. an LSP segment can indeed be an FA link inherited from the creation of an (intra-domain) FA-LSP - </FONT></P></BLOCKQUOTE></DIV><p>__________________________________________________<br>Do You Yahoo!?<br>Tired of spam? Yahoo! Mail has the best spam protection around <br>http://mail.yahoo.com --0-499060946-1121886946=:94207-- Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 19:08:22 +0000 To: "Igor Bryskin" <ibryskin@movaz.com> Cc: <Dimitri.Papadimitriou@alcatel.be>, "Arthi Ayyangar" <arthi@juniper.net>, <ccamp@ops.ietf.org> MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: Re: LSP stitching questions for discussion Date: Wed, 20 Jul 2005 21:08:01 +0200 Message-ID: <OFBD3B496D.F86F66A5-ONC1257044.00691A1A-C1257044.00691AB5@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+aWdvciAtIGluLWxpbmUgPEJSPjxCUj48QlI+PEZPTlQgU0laRT00Pml0IGlzICZxdW90O25v JnF1b3Q7IHRvIGFsbCB0aGUgYmVsb3cgYSksIGIpIGFuZCBjKSAtIGFzIGxvZ2ljYWwgY29uc2Vx dWVuY2Ugb2YgdGhlIGN1cnJlbnQgR01QTFMgYXJjaGl0ZWN0dXJlPC9GT05UPjxCUj48QlI+PEJS PjxCUj48Rk9OVCBTSVpFPTQ+bm90ZTogdG8gY2xhcmlmeSBvbmUgb2YgbXkgbm90ZSBoZXJlIGJl bG93ICh3aGVyZSBpdCBzZWVtcyB5b3Ugd2VyZSBub3QgaW4gYWdyZWVtZW50KSBhYm91dCBQU0Mg TFNQIGJhbmR3aWR0aCBhbmQgc2VnbWVudCBzdGl0Y2hpbmcgaS5lLiBsYXJnZXIgYmFuZHdpZHRo IHJlc2VydmF0aW9uIGZvciB0aGUgc2VnbWVudCB0aGFuIGZvciB0aGUgZW5kLXRvLWVuZCBMU1Ag ZG9lcyBub3QgbWVhbiBkaWZmZXJlbnQgUFNDIGxldmVsIChub3RlOiBQU0MtMSwgLTIsIC0zIGFu ZCAtNCBhcmUgbm90ICZxdW90O2xheWVycyZxdW90OyB0aGV5IHNob3VsZCBzaW1wbHkgYmUgcmVm ZXJyZWQgdG8gYXMgbGV2ZWxzIGJ1dCB0aGlzIGlzIGFub3RoZXIgZGlzY3Vzc2lvbikgc28gaSB3 YW50ZWQgdG8gY2xhcmlmeSB0aGF0IHRoZSBkZWZhdWx0IGFkbWlzc2lvbiBjb250cm9sIHJ1bGVz IGZvciB0aGUgcGFja2V0IGNhc2UgLSB3aXRoaW4gdGhlIHNhbWUgUFNDIGxldmVsIC0gY291bGQg YmUgcmVsYXhlZCBjb21wYXJlZCB0byB0aGUgb3RoZXIgc3dpdGNoaW5nIHRlY2hub2xvZ2llcyB3 aXRob3V0IHN0cmV0Y2hpbmcgdGhlIHN0aXRjaGluZyBjb25jZXB0PC9GT05UPjxCUj48QlI+PEZP TlQgU0laRT00PklCJmd0OyZndDsgTXkgcXVlc3Rpb24gYykgd2FzOjwvRk9OVD48QlI+PEZPTlQg U0laRT00PmMpIElzIGl0IHBvc3NpYmxlIChzaG91bGQgd2UgYWxsb3cpIHRvIHN0aXRjaCZuYnNw OyZuYnNwO1BTQzEgZTJlIExTUCBieSBhJm5ic3A7UFNDMiBMU1Agc2VnbWVudCA8L0ZPTlQ+PEJS PjxCUj48Rk9OVCBTSVpFPTQ+SWYgeW91IHNheSBOTyB0byB0aGlzIHF1ZXN0aW9uLCB0aGFuIHRo ZSBhbnN3ZXIgY29udHJhZGljdHMgdG8gd2hhdCB5b3Ugc2FpZCBhYm92ZSBpbiB5b3VyIG5vdGUu PC9GT05UPjxCUj48L1A+PFA+W2RwXSBvayAtIGxhc3QgdHJ5IGZvciBhbiBlbmQtdG8tZW5kIFBT Qy0xIExTUCBvZiAyTUIgeW91IGNvdWxkIG1ha2UgdXNlIG9mIGEgUFNDLTEgc2VnbWVudCBvZiA0 TUIgPC9QPjxQPjxCUj5Zb3UgYXJlIHJpZ2h0OiB3aGV0aGVyIFBTQzEgYW5kIFBTQzIgYXJlIHNl cGFyYXRlIG5ldHdvcmsgbGF5ZXJzIG9yIG5vdCBpcyBhIHRvcGljIG9mIGEgc2VwYXJhdGUgKHBh aW5mdWwgOj0pKSBkaXNjdXNzaW9uLiBIb3dldmVyLCBJIGp1c3Qgd2FudGVkIHRvIHBvaW50IG91 dCBoZXJlIHRoYXQgYSBQU0MyIExTUCBpcyBhIHByb3Zpc2lvbmVkIGFzIHN1Y2ggZm9yIGEgcmVh c29uOiZuYnNwO2l0IGlzIG1lYW50IHRvICpuZXN0KiBQU0MxIExTUHMgcmF0aGVyIHRoYW4gdG8g c3RpdGNoIHRoZW0uIEZyb20gdGhlIFZOVCBtYW5hZ2VtZW50IHBvaW50IG9mIHZpZXcgUFNDMiBh bmQgUFNDMSBhcmUgc2VwYXJhdGUgdG9wb2xvZ2ljYWwgbGF5ZXJzLiB0aGF0IGlzIFBTQzIgcHJv ZHVjZXMgdG9wb2xvZ3kgZm9yIFBTQzEsIGp1c3QgbGlrZSBWQzQgcHJvZHVjZXMgdG9wb2xvZ3kg Zm9yIFZDMTIuPEJSPlRoZSBvbmx5IGRpZmZlcmVuY2UgaXMgdGhhdCBpbiB0aGUgbGF0dGVyIGNh c2UgdGhpcyByZWxhdGlvbnNoaXAgaXMgZm9yY2VkIGJ5IHRoZSBoYXJkd2FyZSwgYW5kIGluIHRo ZSBmb3JtZXIgY2FzZSBpcyBub3QgKGp1c3QgbG9naWNhbCkuIFRha2UgZm9yIGV4YW1wbGUgQyBh bmQgQysrIHByb2dyYW1taW5nLiBDZXJ0YWluIHRoaW5ncyB5b3Ugc2ltcGx5IGZvcmNlZCBub3Qg dG8gZG8gaW4gQysrLCB3aGVyZWFzIGluIEMgeW91IGNvdWxkIGJ1dCBzaG91bGRuJ3QuPEJSPjxC Uj5TZWUgeW91IGluIFBhcmlzLjxCUj5JZ29yPEJSPjxGT05UIFNJWkU9ND4gPC9GT05UPjxCUj48 QlI+PEJSPjxCUj48Qj4mcXVvdDtJZ29yIEJyeXNraW4mcXVvdDsgJmx0O2licnlza2luQG1vdmF6 LmNvbSZndDs8L0I+PEJSPlNlbnQgYnk6IG93bmVyLWNjYW1wQG9wcy5pZXRmLm9yZzxCUj4wNy8y MC8yMDA1IDEzOjI1IEFTVDxCUj48QlI+VG86PEZPTlQgU0laRT00PiA8L0ZPTlQ+JnF1b3Q7QXJ0 aGkgQXl5YW5nYXImcXVvdDsgJmx0O2FydGhpQGp1bmlwZXIubmV0Jmd0OywgRGltaXRyaSBQQVBB RElNSVRSSU9VL0JFL0FMQ0FURUxAQUxDQVRFTDxCUj5jYzo8Rk9OVCBTSVpFPTQ+IDwvRk9OVD4m bHQ7Y2NhbXBAb3BzLmlldGYub3JnJmd0OzxCUj5iY2M6PEZPTlQgU0laRT00PiA8L0ZPTlQ+PEJS PlN1YmplY3Q6PEZPTlQgU0laRT00PiA8L0ZPTlQ+UmU6IExTUCBzdGl0Y2hpbmcgcXVlc3Rpb25z IGZvciBkaXNjdXNzaW9uPEJSPjxCUj48QlI+PEZPTlQgU0laRT00PkRpbWl0cmksPC9GT05UPjxC Uj48QlI+PEZPTlQgU0laRT00Pkl0IHNlZW1zIHRvIG1lIHRoYXQgSSBtb3N0bHkgYWdyZWUoPyEp IHdpdGggd2hhdCB5b3Ugd3JvdGUgaW4gdGhpcyBtYWlsLiBJJ2Qgc3RpbGwgd2FudCB0byBjb25m aXJtIHlvdXIgdmlld3MmbmJzcDsgb24gdGhlIGZvbGxvd2luZyBxdWVzdGlvbnM6PC9GT05UPjxC Uj48QlI+PEJSPjxGT05UIFNJWkU9ND5hKSBJcyBpdCBwb3NzaWJsZSAoc2hvdWxkIHdlIGFsbG93 KSB0byBzdGl0Y2gmbmJzcDsgUFNDMSBlMmUgTFNQIGJ5IGEgTFNDIExTUCBzZWdtZW50ID8gPC9G T05UPjxCUj48QlI+PEZPTlQgU0laRT00PmIpIElzIGl0IHBvc3NpYmxlIChzaG91bGQgd2UgYWxs b3cpIHRvIHN0aXRjaCZuYnNwOyZuYnNwO1ZDMTIgZTJlIExTUCBieSBhJm5ic3A7VkM0IExTUCBz ZWdtZW50ID8gPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+YykgSXMgaXQgcG9zc2libGUgKHNob3Vs ZCB3ZSBhbGxvdykgdG8gc3RpdGNoJm5ic3A7Jm5ic3A7UFNDMSBlMmUgTFNQIGJ5IGEmbmJzcDtQ U0MyIExTUCBzZWdtZW50ID88L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTQ+TXkgYW5zd2VycyBh cmUgTk8gdG8gYWxsIHRocmVlIHF1ZXN0aW9ucy4gWW91IGNhbiBzdGljdGggUFNDMSB3aXRoIFBT QzEsIFZDMTIgd2l0aCBWQzEyLCBldGMuIEluIGFsbCBvdGhlciBjYXNlcyB5b3UgbmVlZCwgZmly c3QsIHRvIHNldHVwIGEgaGllcmFyY2hpY2FsIExTUCAoSC1MU1ApICh3aGV0aGVyIGl0IGlzIEZB LUxTUCAoc2luZ2xlIGRvbWFpbikgb3Igbm90Jm5ic3A7KG11bHRpLWRvbWFpbikpIGluIG9yZGVy IHRvIGNyZWF0ZSBhIGRhdGEgbGluayBvZiB0aGUgZTJlIExTUCBzd2l0Y2hpbmcgY2FwYWJpbGl0 eSBhbmQgb25seSBhZnRlciB0aGF0IHlvdSBjYW4gYWxzbyZuYnNwO3VzZSBhIHN0aXRjaGluZyBz ZWdtZW50IG9uIHRvcCBvZiAmbmJzcDt0aGlzIEgtTFNQKS48L0ZPTlQ+PEJSPjxCUj48Rk9OVCBT SVpFPTQ+U2VlIGFsc28gaW4tbGluZS48L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTQ+SWdvcjwv Rk9OVD48QlI+PEJSPjxCUj48Rk9OVCBTSVpFPTU+LS0tLS0gT3JpZ2luYWwgTWVzc2FnZSAtLS0t LSA8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9NT48Qj5Gcm9tOjwvQj4gPC9GT05UPjxGT05UIFNJWkU9 NSBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOkRpbWl0cmkuUGFwYWRpbWl0cmlvdUBhbGNh dGVsLmJlPkRpbWl0cmkuUGFwYWRpbWl0cmlvdUBhbGNhdGVsLmJlPC9BPjwvVT48L0ZPTlQ+PEZP TlQgU0laRT01IENPTE9SPUJMQUNLPiA8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9NT48Qj5Ubzo8L0I+ IDwvRk9OVD48Rk9OVCBTSVpFPTUgQ09MT1I9QkxVRT48VT48QSBIUkVGPW1haWx0bzphcnRoaUBq dW5pcGVyLm5ldD5BcnRoaSBBeXlhbmdhcjwvQT48L1U+PC9GT05UPjxGT05UIFNJWkU9NSBDT0xP Uj1CTEFDSz4gPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+PEI+Q2M6PC9CPiA8L0ZPTlQ+PEZPTlQg U0laRT01IENPTE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86aWJyeXNraW5AbW92YXouY29tPkln b3IgQnJ5c2tpbjwvQT48L1U+PC9GT05UPjxGT05UIFNJWkU9NSBDT0xPUj1CTEFDSz4gOyA8L0ZP TlQ+PEZPTlQgU0laRT01IENPTE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86Y2NhbXBAb3BzLmll dGYub3JnPmNjYW1wQG9wcy5pZXRmLm9yZzwvQT48L1U+PC9GT05UPjxGT05UIFNJWkU9NSBDT0xP Uj1CTEFDSz4gPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+PEI+U2VudDo8L0I+IFdlZG5lc2RheSwg SnVseSAyMCwgMjAwNSA0OjQ2IEFNPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+PEI+U3ViamVjdDo8 L0I+IFJlOiBMU1Agc3RpdGNoaW5nIHF1ZXN0aW9ucyBmb3IgZGlzY3Vzc2lvbjwvRk9OVD48QlI+ PEJSPjxGT05UIFNJWkU9NT5hcnRoaSwgPC9GT05UPjxCUj48QlI+PEJSPjxCUj48Rk9OVCBTSVpF PTU+Zmlyc3QgYSByZWNhcDogdGhpcyBpdGVtIG9uIHN0aXRjaGluZyB3YXMgaW5pdGlhdGVkIGlu IG9yZGVyIHRvIGFsbG93IHBlciAmcXVvdDtkb21haW4mcXVvdDsgY29udHJvbCBvZiBlbmQtdG8t ZW5kIChjb250aWd1b3VzKSBMU1AsIGFzIHBhcnQgb2YgYSBzZXBhcmF0ZSBkb2N1bWVudCwgYW5k IG5vdCB0byBleHRlbmQgaXRzIHNjb3BlICh1bmR1bHksIGluIHRoaXMgY2FzZSwgc2VlIGJlbG93 KSBidXQgbWFpbmx5IGluIG9yZGVyIHRvIGhhdmUgdGhlIHNldCBvZiBhbGwgcmVsYXRlZCBwcm90 b2NvbCBtZWNoYW5pc21zIHdpdGhpbiBhIHNwZWNpZmljIGRvY3VtZW50IC08L0ZPTlQ+PEJSPjxC Uj48QlI+PEJSPjxGT05UIFNJWkU9NT5mcm9tIHRoaXMgaW5pdGlhbCB3b3JraW5nIGFzc3VtcHRp b24sIGl0IGlzIGNsZWFyIHRoYXQgdGhlIHNlZ21lbnQgbXVzdCBoYXZlIHRoZSBwcm9wZXJ0aWVz IG9mIHRoZSBpbmNvbWluZyBlbmQtdG8tZW5kIExTUCB3aGljaCBhcmUgZGV0ZXJtaW5lZCBhbW9u ZyBvdGhlciBieSBpdHMgc3dpdGNoaW5nIHR5cGUgZS5nLiBQU0MtMyBvciBURE0sIHRoZSBmaXJz dCByZXF1aXJlbWVudCBvZiBMU1Agc3RpdGNoaW5nIGlzIGNvbnRpZ3VpdHkgb2YgdGhlIGVuZC10 by1lbmQgTFNQIGF0IHRoZSBkYXRhIHBsYW5lIGxldmVsIChhZnRlciBzdGl0Y2hpbmcgb3BlcmF0 aW9uKSAtIGFuZCB0aGlzIGJ5IGRlZmluaXRpb24gb2YgdGhlIGluaXRpYWwgd29ya2luZyBhc3N1 bXB0aW9uIC08L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTU+SUImZ3Q7Jmd0OyBBZ3JlZTwvRk9O VD48QlI+PEJSPjxGT05UIFNJWkU9NT5ub3RlOiB0aGUgb25seSBsYXRpdHVkZSB3aGljaCBpcyBs ZWZ0IChmb3IgYSBnaXZlbiBzd2l0Y2hpbmcgdHlwZSBlLmcuIFBTQy0yKSBpcyBkdWUgdG8gdGhl IHByb3BlcnRpZXMgb2YgcGFja2V0IChQU0MpIG5ldHdvcmtzIHRoYXQgYWxsb3dzIHlvdSBzdGl0 Y2ggYSBzZWdtZW50IGhhdmluZyBhIGxhcmdlciBNYXggUmVzZXJ2YWJsZSBCYW5kd2lkdGggYW5k IFVucmVzZXJ2ZWQgQmFuZHdpZHRoICh3ZSB3b3VsZCBoYXZlIGVxdWl2YWxlbnQgbGF0aXR1ZGUg Zm9yIG90aGVyIGF0dHJpYnV0ZXMgYnV0IHRoaXMgaXMgYSBsb2NhbCBwb2xpY3kgZGVjaXNpb24g LSB3aGljaCBhdCB0aGUgZW5kIGlzIHRoZSBzb2xlIHB1cnBvc2Ugb2YgdGhpcyBhcHByb2FjaCk8 L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTU+SUImZ3Q7Jmd0OyBEaXNhZ3JlZS4gQ29uY2VwdHVh bGx5IFBDUzEgYW5kIFBTQzIgYXJlIGRpZmZlcmVudCBsYXllcnMgYW5kIHRoZSBsYXR0ZXIgY2Fu bm90IHN0aXRjaCB0aGUgZm9ybWVyLCBqdXN0IGxpa2UgaW4gY2FzZSBWQzEyIGFuZCBWQzQuIFlv dSBzaG91bGQgbmVzdCBQU0MxIHdpdGhpbiBQU0MyIGluc3RlYWQuPC9GT05UPjxCUj48QlI+PEJS PjxCUj48Rk9OVCBTSVpFPTU+SGVyZSBhcmUgc29tZSBvZiB0aGUgcXVlc3Rpb25zIHJlbGF0ZWQg dG8gTFNQIHN0aXRjaGluZyB0byBzdGFydCBhPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+ZGlzY3Vz c2lvbiBvbiB0aGUgdXNhZ2Ugb2YgTFNQIHN0aXRjaGluZy48L0ZPTlQ+PEJSPjxCUj48Rk9OVCBT SVpFPTU+Jmd0OzEpIEFyZSB0aGUgY29udHJvbCBwbGFuZSBzaWduYWxpbmcgcHJvY2VkdXJlcyBm b3IgTFNQIHN0aXRjaGluZyBhczwvRk9OVD48QlI+PEZPTlQgU0laRT01PiZndDtkZXNjcmliZWQg aW4gdGhpcyBJRCwgKGV4cGxpY2l0IHJlcXVlc3QgZm9yIExTUCBzdGl0Y2hpbmcgb248L0ZPTlQ+ PEJSPjxGT05UIFNJWkU9NT4mZ3Q7TFNQIHNlZ21lbnQsIGRpZmZlcmVudCBsYWJlbCBhbGxvY2F0 aW9uIHJ1bGVzOyBldGMpIFJFUVVJUkVEIHRvIHN0aXRjaDwvRk9OVD48QlI+PEZPTlQgU0laRT01 PiZndDtwYWNrZXQgTFNQcyBpbiBkYXRhIHBsYW5lIDwvRk9OVD48QlI+PEJSPjxCUj48Rk9OVCBT SVpFPTU+LSZndDsgeWVzIChiZXNpZGUgbWFudWFsIGNvbmZpZyBvcGVyYXRpb24pPC9GT05UPjxC Uj48QlI+PEZPTlQgU0laRT01PklCJmd0OyZndDsgQWdyZWU8L0ZPTlQ+PEJSPjxCUj48QlI+PEZP TlQgU0laRT01PiZndDsyKSBBcmUgdGhlIGNvbnRyb2wgcGxhbmUgc2lnbmFsaW5nIHByb2NlZHVy ZXMgZm9yIExTUCBzdGl0Y2hpbmcgYXM8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9NT4mZ3Q7ZGVzY3Jp YmVkIGluIHRoaXMgSUQsIChleHBsaWNpdCByZXF1ZXN0IGZvciBMU1Agc3RpdGNoaW5nIG9uPC9G T05UPjxCUj48Rk9OVCBTSVpFPTU+Jmd0O0xTUCBzZWdtZW50LCBkaWZmZXJlbnQgbGFiZWwgYWxs b2NhdGlvbiBydWxlczsgZXRjKSBSRVFVSVJFRCB0byBzdGl0Y2g8L0ZPTlQ+PEJSPjxGT05UIFNJ WkU9NT4mZ3Q7bm9uLXBhY2tldCBMU1BzIGluIGRhdGEgcGxhbmUgPzwvRk9OVD48QlI+PEJSPjxC Uj48Rk9OVCBTSVpFPTU+LSZndDsgeWVzIChiZXNpZGUgbWFudWFsIGNvbmZpZyBvcGVyYXRpb24p PC9GT05UPjxCUj48QlI+PEZPTlQgU0laRT01PklCJmd0OyZndDsgQWdyZWU8L0ZPTlQ+PEJSPjxC Uj48QlI+PEZPTlQgU0laRT01PiZndDszKSBJZiBhbiBlMmUgTFNQIGNyb3NzZXMgYSByZWdpb24g Ym91bmRhcnkgKGJhc2VkIG9uIGRpZmZlcmVudCBURTwvRk9OVD48QlI+PEZPTlQgU0laRT01PiZn dDtsaW5rIHN3aXRjaGluZyBjYXBhYmlsaXRpZXMpLCB0aGVuIGNhbiBJIHVzZSBMU1Agc3RpdGNo aW5nIDwvRk9OVD48QlI+PEZPTlQgU0laRT01PiZndDthKSBjb250cm9sIHBsYW5lIHByb2NlZHVy ZXMgZm9yIHBhY2tldCBMU1BzPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+Jmd0O2IpIGRhdGEgcGxh bmUgcHJvY2VkdXJlcyBmb3IgcGFja2V0IExTUHM8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9NT4mZ3Q7 YykgY29udHJvbCBwbGFuZSBwcm9jZWR1cmVzIGZvciBub24tcGFja2V0IExTUHM8L0ZPTlQ+PEJS PjxGT05UIFNJWkU9NT4mZ3Q7ZCkgZGF0YSBwbGFuZSBwcm9jZWR1cmVzIGZvciBub24tcGFja2V0 IExTUHM8L0ZPTlQ+PEJSPjxCUj48QlI+PEZPTlQgU0laRT01Pi0mZ3Q7IG5vLCB3aGVuIGNyb3Nz aW5nIGEgcmVnaW9uIGJvdW5kYXJ5LCB5b3Ugc2ltcGx5IHVzZSBhIGRvY3VtZW50IG5hbWVkICZx dW90O2RyYWZ0LW1wbHMtbHNwLWhpZXJhcmNoeS0wOC50eHQmcXVvdDsgd2hpY2ggaXMgaW4gdGhl IFJGQyBFZGl0b3IgcXVldWUsIGJ1dCB0aGlzIGRvZXMgbm90IHByZXZlbnQgZnJvbSBjcmVhdGlu ZyBhbiBMU1Agc2VnbWVudCBvbiB0b3Agb2YgYW4gRkEtTFNQIChzZWUgYmVsb3cpPC9GT05UPjxC Uj48QlI+PEZPTlQgU0laRT01PklCJmd0OyZndDsgQWdyZWUgd2l0aCB0aGUgY2F2aWF0IHRoYXQg aXQgc2hvdWxkIGJlIEgtTFNQIGluc3RlYWQgb2YgRkEtTFNQIHRvIGNvdmVyIHNpbmdsZSBhbmQg bXVsdGktZG9tYWluIHNjZW5hcmlvczwvRk9OVD48QlI+PEJSPjxCUj48Rk9OVCBTSVpFPTU+Jmd0 OzQpIElmIHRoZSBzd2l0Y2hpbmcgdHlwZSBvZiBhbiBlMmUgTFNQIGlzIGRpZmZlcmVudCBmcm9t IHRoYXQgb2YgYW4gTFNQPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+Jmd0O3NlZ21lbnQsIGNhbiBJ IHVzZSBMU1Agc3RpdGNoaW5nIDwvRk9OVD48QlI+PEZPTlQgU0laRT01PiZndDthKSBjb250cm9s IHBsYW5lIHByb2NlZHVyZXMgZm9yIHBhY2tldCBMU1BzPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+ Jmd0O2IpIGRhdGEgcGxhbmUgcHJvY2VkdXJlcyBmb3IgcGFja2V0IExTUHM8L0ZPTlQ+PEJSPjxG T05UIFNJWkU9NT4mZ3Q7YykgY29udHJvbCBwbGFuZSBwcm9jZWR1cmVzIGZvciBub24tcGFja2V0 IExTUHM8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9NT4mZ3Q7ZCkgZGF0YSBwbGFuZSBwcm9jZWR1cmVz IGZvciBub24tcGFja2V0IExTUHM8L0ZPTlQ+PEJSPjxCUj48QlI+PEZPTlQgU0laRT01Pi0mZ3Q7 IG5vLCBpbiB0aGlzIGNhc2UgcHJvZ3Jlc3Npb24gb2YgdGhlIGVuZC10by1lbmQgTFNQIGVzdGFi bGlzaG1lbnQsIGltcGxpZXMgdGhhdCB0aGUgTFNQIHNlZ21lbnQgbXVzdCBmdWxmaWwgYSBzZXQg b2YgY29uc3RyYWludHMgY2FycmllZCBhcyBwYXJ0IG9mIHRoZSBpbmNvbWluZyBlbmQtdG8tZW5k IExTUCByZXF1ZXN0LCBpbiB0aGlzIGNhc2Ugb25lIGhhcyB0byBjcmVhdGUgYW4gTFNQIHNlZ21l bnQgKHdpdGggYSBzd2l0Y2hpbmcgdHlwZSBlcXVhbCB0byB0aGUgaW5jb21pbmcgZW5kLXRvLWVu ZCBMU1Agc3dpdGNoaW5nIHR5cGUpIG92ZXIgdGhhdCBGQSAoaS5lLiByZWZlcnJlZCBoZXJlIGFi b3ZlIHRvIHRoZSBMU1Agc2VnbWVudCB3aXRoIGEgZGlmZmVyZW50IFN3aXRjaGluZyBUeXBlKTwv Rk9OVD48QlI+PEJSPjxCUj48Rk9OVCBTSVpFPTU+SUImZ3Q7Jmd0OyBBZ3JlZSB3aXRoIHRoZSBj YXZpYXQgdGhhdCBpdCBzaG91bGQgYmUgSC1MU1AgaW5zdGVhZCBvZiBGQS1MU1AgdG8gY292ZXIg c2luZ2xlIGFuZCBtdWx0aS1kb21haW4gc2NlbmFyaW9zPC9GT05UPjxCUj48QlI+PEZPTlQgU0la RT00PkNoZWVycyw8L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTQ+SWdvcjwvRk9OVD48QlI+PEJS PjxGT05UIFNJWkU9NT4mZ3Q7Tm90ZSB0aGF0IGFsdGhvdWdoIDMpIGFuZCA0KSBhcHBlYXIgdG8g YmUgdGhlIHNhbWUsIHRoZSByZWFzb24gSSBoYXZlPC9GT05UPjxCUj48Rk9OVCBTSVpFPTU+Jmd0 O3RoZW0gYXMgZGlmZmVyZW50IHF1ZXN0aW9ucyBpcyB0aGF0IGluIG9uZSBjYXNlIHRoZSBURSBs aW5rIHN3aXRjaGluZzwvRk9OVD48QlI+PEZPTlQgU0laRT01PiZndDtjYXBhYmlsaXR5IGlzIGV4 YW1pbmVkIGFuZCBpbiB0aGUgb3RoZXIgY2FzZSB0aGUgR2VuZXJhbGl6ZWQgTGFiZWwgUmVxdWVz dDwvRk9OVD48QlI+PEZPTlQgU0laRT01PiZndDtpcyBleGFtaW5lZC48L0ZPTlQ+PEJSPjxCUj48 QlI+PEJSPjxGT05UIFNJWkU9NT4tJmd0OyB3aGF0IGl0IGFwcGVhcnMgdG8gbWUgaXMgdGhhdCB0 aGlzIGRpc2N1c3Npb24gbWl4ZXMgdGhlIGlzc3VlIG9mIHRoZSBvcGVyYXRpb24gZGVzY3JpYmlu ZyB0aGUgc3RpdGNoaW5nIG9wZXJhdGlvbnMgYXQgdGhlIGRvbWFpbiBib3VuZGFyaWVzIGZyb20g dGhlIExTUCBzZWdtZW50IGNyZWF0aW9uIGl0c2VsZiBpLmUuIGFuIExTUCBzZWdtZW50IGNhbiBp bmRlZWQgYmUgYW4gRkEgbGluayBpbmhlcml0ZWQgZnJvbSB0aGUgY3JlYXRpb24gb2YgYW4gKGlu dHJhLWRvbWFpbikgRkEtTFNQIC0gPC9GT05UPjwvUD4= Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 19:01:16 +0000 Message-ID: <013d01c58d5d$54912340$7a1810ac@movaz.com> From: "Igor Bryskin" <ibryskin@movaz.com> To: <Dimitri.Papadimitriou@alcatel.be> Cc: "Arthi Ayyangar" <arthi@juniper.net>, <Dimitri.Papadimitriou@alcatel.be>, <ccamp@ops.ietf.org> Subject: Re: LSP stitching questions for discussion Date: Wed, 20 Jul 2005 15:00:43 -0400 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_013A_01C58D3B.CD519570" This is a multi-part message in MIME format. ------=_NextPart_000_013A_01C58D3B.CD519570 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Dimitri, igor -=20 it is "no" to all the below a), b) and c) - as logical consequence of = the current GMPLS architecture note: to clarify one of my note here below (where it seems you were = not in agreement) about PSC LSP bandwidth and segment stitching i.e. = larger bandwidth reservation for the segment than for the end-to-end LSP = does not mean different PSC level (note: PSC-1, -2, -3 and -4 are not = "layers" they should simply be referred to as levels but this is another = discussion) so i wanted to clarify that the default admission control = rules for the packet case - within the same PSC level - could be relaxed = compared to the other switching technologies without stretching the = stitching concept IB>> My question c) was: c) Is it possible (should we allow) to stitch PSC1 e2e LSP by a PSC2 = LSP segment ? If you say NO to this question, than the answer contradicts to what = you said above in your note. You are right: whether PSC1 and PSC2 are separate network layers or = not is a topic of a separate (painful :=3D)) discussion. However, I just = wanted to point out here that a PSC2 LSP is a provisioned as such for a = reason: it is meant to *nest* PSC1 LSPs rather than to stitch them. From = the VNT management point of view PSC2 and PSC1 are separate topological = layers. that is PSC2 produces topology for PSC1, just like VC4 produces = topology for VC12. The only difference is that in the latter case this relationship is = forced by the hardware, and in the former case is not (just logical). = Take for example C and C++ programming. Certain things you simply forced = not to do in C++, whereas in C you could but shouldn't. See you in Paris. Igor "Igor Bryskin" <ibryskin@movaz.com> Sent by: owner-ccamp@ops.ietf.org 07/20/2005 13:25 AST To: "Arthi Ayyangar" <arthi@juniper.net>, Dimitri = PAPADIMITRIOU/BE/ALCATEL@ALCATEL cc: <ccamp@ops.ietf.org> bcc:=20 Subject: Re: LSP stitching questions for discussion Dimitri, It seems to me that I mostly agree(?!) with what you wrote in this = mail. I'd still want to confirm your views on the following questions: a) Is it possible (should we allow) to stitch PSC1 e2e LSP by a LSC = LSP segment ?=20 b) Is it possible (should we allow) to stitch VC12 e2e LSP by a VC4 = LSP segment ?=20 c) Is it possible (should we allow) to stitch PSC1 e2e LSP by a PSC2 = LSP segment ? My answers are NO to all three questions. You can sticth PSC1 with = PSC1, VC12 with VC12, etc. In all other cases you need, first, to setup = a hierarchical LSP (H-LSP) (whether it is FA-LSP (single domain) or not = (multi-domain)) in order to create a data link of the e2e LSP switching = capability and only after that you can also use a stitching segment on = top of this H-LSP). See also in-line. Igor ----- Original Message -----=20 From: Dimitri.Papadimitriou@alcatel.be=20 To: Arthi Ayyangar=20 Cc: Igor Bryskin ; ccamp@ops.ietf.org=20 Sent: Wednesday, July 20, 2005 4:46 AM Subject: Re: LSP stitching questions for discussion arthi,=20 first a recap: this item on stitching was initiated in order to allow = per "domain" control of end-to-end (contiguous) LSP, as part of a = separate document, and not to extend its scope (unduly, in this case, = see below) but mainly in order to have the set of all related protocol = mechanisms within a specific document - from this initial working assumption, it is clear that the segment = must have the properties of the incoming end-to-end LSP which are = determined among other by its switching type e.g. PSC-3 or TDM, the = first requirement of LSP stitching is contiguity of the end-to-end LSP = at the data plane level (after stitching operation) - and this by = definition of the initial working assumption - IB>> Agree note: the only latitude which is left (for a given switching type e.g. = PSC-2) is due to the properties of packet (PSC) networks that allows you = stitch a segment having a larger Max Reservable Bandwidth and Unreserved = Bandwidth (we would have equivalent latitude for other attributes but = this is a local policy decision - which at the end is the sole purpose = of this approach) IB>> Disagree. Conceptually PCS1 and PSC2 are different layers and the = latter cannot stitch the former, just like in case VC12 and VC4. You = should nest PSC1 within PSC2 instead. Here are some of the questions related to LSP stitching to start a discussion on the usage of LSP stitching. >1) Are the control plane signaling procedures for LSP stitching as >described in this ID, (explicit request for LSP stitching on >LSP segment, different label allocation rules; etc) REQUIRED to = stitch >packet LSPs in data plane ? -> yes (beside manual config operation) IB>> Agree >2) Are the control plane signaling procedures for LSP stitching as >described in this ID, (explicit request for LSP stitching on >LSP segment, different label allocation rules; etc) REQUIRED to = stitch >non-packet LSPs in data plane ? -> yes (beside manual config operation) IB>> Agree >3) If an e2e LSP crosses a region boundary (based on different TE >link switching capabilities), then can I use LSP stitching - >a) control plane procedures for packet LSPs >b) data plane procedures for packet LSPs >c) control plane procedures for non-packet LSPs >d) data plane procedures for non-packet LSPs -> no, when crossing a region boundary, you simply use a document = named "draft-mpls-lsp-hierarchy-08.txt" which is in the RFC Editor = queue, but this does not prevent from creating an LSP segment on top of = an FA-LSP (see below) IB>> Agree with the caviat that it should be H-LSP instead of FA-LSP = to cover single and multi-domain scenarios >4) If the switching type of an e2e LSP is different from that of an = LSP >segment, can I use LSP stitching=20 >a) control plane procedures for packet LSPs >b) data plane procedures for packet LSPs >c) control plane procedures for non-packet LSPs >d) data plane procedures for non-packet LSPs -> no, in this case progression of the end-to-end LSP establishment, = implies that the LSP segment must fulfil a set of constraints carried as = part of the incoming end-to-end LSP request, in this case one has to = create an LSP segment (with a switching type equal to the incoming = end-to-end LSP switching type) over that FA (i.e. referred here above to = the LSP segment with a different Switching Type) IB>> Agree with the caviat that it should be H-LSP instead of FA-LSP = to cover single and multi-domain scenarios Cheers, Igor >Note that although 3) and 4) appear to be the same, the reason I have >them as different questions is that in one case the TE link switching >capability is examined and in the other case the Generalized Label = Request >is examined. -> what it appears to me is that this discussion mixes the issue of = the operation describing the stitching operations at the domain = boundaries from the LSP segment creation itself i.e. an LSP segment can = indeed be an FA link inherited from the creation of an (intra-domain) = FA-LSP -=20 ------=_NextPart_000_013A_01C58D3B.CD519570 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META http-equiv=3DContent-Type content=3D"text/html; = charset=3Diso-8859-1"> <META content=3D"MSHTML 6.00.2800.1400" name=3DGENERATOR> <STYLE></STYLE> </HEAD> <BODY bgColor=3D#ffffff> <DIV><FONT face=3DArial size=3D2>Dimitri,</FONT></DIV> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <DIV><FONT face=3DMonospace,Courier>igor - </FONT></DIV> <BLOCKQUOTE dir=3Dltr=20 style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; = BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px"> <P> </P> <P><FONT face=3DMonospace,Courier>it is "no" to all the below a), b) = and c) - as=20 logical consequence of the current GMPLS architecture</FONT></P> <P> </P> <P><FONT face=3DMonospace,Courier>note: to clarify one of my note here = below=20 (where it seems you were not in agreement) about PSC LSP bandwidth and = segment=20 stitching i.e. larger bandwidth reservation for the segment than for = the=20 end-to-end LSP does not mean different PSC level (note: PSC-1, -2, -3 = and -4=20 are not "layers" they should simply be referred to as levels but this = is=20 another discussion) so i wanted to clarify that the default admission = control=20 rules for the packet case - within the same PSC level - could be = relaxed=20 compared to the other switching technologies without stretching the = stitching=20 concept</FONT></P><FONT face=3D"Courier New"></FONT></BLOCKQUOTE> <BLOCKQUOTE dir=3Dltr=20 style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; = BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px"> <DIV><FONT face=3D"Courier New">IB>> My question c) = was:</FONT></DIV> <DIV><FONT face=3D"Courier New"><FONT face=3D"Times New Roman">c) Is = it possible=20 (should we allow) to stitch PSC1 e2e LSP by a PSC2 LSP = segment=20 ?</FONT><BR></FONT></DIV> <DIV><FONT face=3D"Courier New">If you say NO to this question, than = the answer=20 contradicts to what you said above in your note.</FONT></DIV> <DIV><FONT face=3D"Courier New"></FONT> </DIV> <DIV><FONT face=3D"Courier New"><FONT face=3DArial size=3D2>You are = right: whether=20 PSC1 and PSC2 are separate network layers or not is a topic of a = separate=20 (painful :=3D)) discussion. However, I just wanted to point out here = that a PSC2=20 LSP is a provisioned as such for a reason: it is meant to *nest* = PSC1=20 LSPs rather than to stitch them. From the VNT management point of view = PSC2=20 and PSC1 are separate topological layers. that is PSC2 produces = topology for=20 PSC1, just like VC4 produces topology for VC12.</FONT></FONT></DIV> <DIV><FONT face=3D"Courier New"><FONT face=3DArial size=3D2>The only = difference is=20 that in the latter case this relationship is forced by the hardware, = and in=20 the former case is not (just logical). Take for example C and C++ = programming.=20 Certain things you simply forced not to do in C++, whereas in C you = could but=20 shouldn't.</FONT></FONT></DIV> <DIV><FONT face=3D"Courier New"><FONT face=3DArial=20 size=3D2></FONT></FONT> </DIV> <DIV><FONT face=3D"Courier New"><FONT face=3DArial size=3D2>See you in = Paris.</FONT></FONT></DIV> <DIV><FONT face=3D"Courier New"><FONT face=3DArial = size=3D2>Igor</FONT></DIV></FONT> <P><FONT face=3DArial size=3D2></FONT><BR><BR></P> <P><FONT size=3D2><B>"Igor Bryskin"=20 <ibryskin@movaz.com></B></FONT><BR><FONT size=3D2>Sent by:=20 owner-ccamp@ops.ietf.org</FONT><BR><FONT size=3D2>07/20/2005 13:25=20 AST</FONT><BR><BR><FONT size=3D2>To:</FONT> <FONT size=3D2>"Arthi = Ayyangar"=20 <arthi@juniper.net>, Dimitri=20 PAPADIMITRIOU/BE/ALCATEL@ALCATEL</FONT><BR><FONT size=3D2>cc:</FONT> = <FONT=20 size=3D2><ccamp@ops.ietf.org></FONT><BR><FONT = size=3D2>bcc:</FONT> <BR><FONT=20 size=3D2>Subject:</FONT> <FONT size=3D2>Re: LSP stitching questions = for=20 discussion</FONT><BR><BR><BR></P> <P>Dimitri,<BR><BR>It seems to me that I mostly agree(?!) with what = you wrote=20 in this mail. I'd still want to confirm your views on the = following=20 questions:<BR><BR><BR>a) Is it possible (should we allow) to = stitch PSC1=20 e2e LSP by a LSC LSP segment ? <BR><BR>b) Is it possible (should we = allow) to=20 stitch VC12 e2e LSP by a VC4 LSP segment ? <BR>c) Is = it=20 possible (should we allow) to stitch PSC1 e2e LSP by = a PSC2=20 LSP segment ?<BR><BR>My answers are NO to all three questions. You can = sticth=20 PSC1 with PSC1, VC12 with VC12, etc. In all other cases you need, = first, to=20 setup a hierarchical LSP (H-LSP) (whether it is FA-LSP (single domain) = or=20 not (multi-domain)) in order to create a data link of the e2e LSP = switching capability and only after that you can also use a = stitching=20 segment on top of this H-LSP).<BR><BR>See also=20 in-line.<BR><BR>Igor<BR><BR><BR><FONT size=3D4>----- Original Message = -----=20 </FONT><BR><FONT size=3D4><B>From:</B> </FONT><FONT color=3Dblue = size=3D4><U><A=20 = href=3D"mailto:Dimitri.Papadimitriou@alcatel.be">Dimitri.Papadimitriou@al= catel.be</A></U></FONT><FONT=20 color=3Dblack size=3D4> </FONT><BR><FONT size=3D4><B>To:</B> = </FONT><FONT color=3Dblue=20 size=3D4><U><A href=3D"mailto:arthi@juniper.net">Arthi=20 Ayyangar</A></U></FONT><FONT color=3Dblack size=3D4> </FONT><BR><FONT=20 size=3D4><B>Cc:</B> </FONT><FONT color=3Dblue size=3D4><U><A=20 href=3D"mailto:ibryskin@movaz.com">Igor Bryskin</A></U></FONT><FONT = color=3Dblack=20 size=3D4> ; </FONT><FONT color=3Dblue size=3D4><U><A=20 = href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A></U></FONT><FONT= =20 color=3Dblack size=3D4> </FONT><BR><FONT size=3D4><B>Sent:</B> = Wednesday, July 20,=20 2005 4:46 AM</FONT><BR><FONT size=3D4><B>Subject:</B> Re: LSP = stitching=20 questions for discussion</FONT><BR><BR><FONT size=3D4>arthi,=20 </FONT><BR><BR><BR><BR><FONT size=3D4>first a recap: this item on = stitching was=20 initiated in order to allow per "domain" control of end-to-end = (contiguous)=20 LSP, as part of a separate document, and not to extend its scope = (unduly, in=20 this case, see below) but mainly in order to have the set of all = related=20 protocol mechanisms within a specific document = -</FONT><BR><BR><BR><BR><FONT=20 size=3D4>from this initial working assumption, it is clear that the = segment must=20 have the properties of the incoming end-to-end LSP which are = determined among=20 other by its switching type e.g. PSC-3 or TDM, the first requirement = of LSP=20 stitching is contiguity of the end-to-end LSP at the data plane level = (after=20 stitching operation) - and this by definition of the initial working=20 assumption -</FONT><BR><BR><FONT size=3D4>IB>> = Agree</FONT><BR><BR><FONT=20 size=3D4>note: the only latitude which is left (for a given switching = type e.g.=20 PSC-2) is due to the properties of packet (PSC) networks that allows = you=20 stitch a segment having a larger Max Reservable Bandwidth and = Unreserved=20 Bandwidth (we would have equivalent latitude for other attributes but = this is=20 a local policy decision - which at the end is the sole purpose of this = approach)</FONT><BR><BR><FONT size=3D4>IB>> Disagree. = Conceptually PCS1=20 and PSC2 are different layers and the latter cannot stitch the former, = just=20 like in case VC12 and VC4. You should nest PSC1 within PSC2=20 instead.</FONT><BR><BR><BR><BR><FONT size=3D4>Here are some of the = questions=20 related to LSP stitching to start a</FONT><BR><FONT = size=3D4>discussion on the=20 usage of LSP stitching.</FONT><BR><BR><FONT size=3D4>>1) Are the = control=20 plane signaling procedures for LSP stitching as</FONT><BR><FONT=20 size=3D4>>described in this ID, (explicit request for LSP stitching = on</FONT><BR><FONT size=3D4>>LSP segment, different label = allocation rules;=20 etc) REQUIRED to stitch</FONT><BR><FONT size=3D4>>packet LSPs in = data plane=20 ?</FONT><BR><BR><BR><FONT size=3D4>-> yes (beside manual config=20 operation)</FONT><BR><BR><FONT size=3D4>IB>> = Agree</FONT><BR><BR><BR><FONT=20 size=3D4>>2) Are the control plane signaling procedures for LSP = stitching=20 as</FONT><BR><FONT size=3D4>>described in this ID, (explicit = request for LSP=20 stitching on</FONT><BR><FONT size=3D4>>LSP segment, different label = allocation rules; etc) REQUIRED to stitch</FONT><BR><FONT=20 size=3D4>>non-packet LSPs in data plane ?</FONT><BR><BR><BR><FONT=20 size=3D4>-> yes (beside manual config = operation)</FONT><BR><BR><FONT=20 size=3D4>IB>> Agree</FONT><BR><BR><BR><FONT size=3D4>>3) If = an e2e LSP=20 crosses a region boundary (based on different TE</FONT><BR><FONT=20 size=3D4>>link switching capabilities), then can I use LSP = stitching=20 -</FONT><BR><FONT size=3D4>>a) control plane procedures for packet=20 LSPs</FONT><BR><FONT size=3D4>>b) data plane procedures for packet=20 LSPs</FONT><BR><FONT size=3D4>>c) control plane procedures for = non-packet=20 LSPs</FONT><BR><FONT size=3D4>>d) data plane procedures for = non-packet=20 LSPs</FONT><BR><BR><BR><FONT size=3D4>-> no, when crossing a region = boundary,=20 you simply use a document named "draft-mpls-lsp-hierarchy-08.txt" = which is in=20 the RFC Editor queue, but this does not prevent from creating an LSP = segment=20 on top of an FA-LSP (see below)</FONT><BR><BR><FONT = size=3D4>IB>> Agree=20 with the caviat that it should be H-LSP instead of FA-LSP to cover = single and=20 multi-domain scenarios</FONT><BR><BR><BR><FONT size=3D4>>4) If the = switching=20 type of an e2e LSP is different from that of an LSP</FONT><BR><FONT=20 size=3D4>>segment, can I use LSP stitching </FONT><BR><FONT = size=3D4>>a)=20 control plane procedures for packet LSPs</FONT><BR><FONT = size=3D4>>b) data=20 plane procedures for packet LSPs</FONT><BR><FONT size=3D4>>c) = control plane=20 procedures for non-packet LSPs</FONT><BR><FONT size=3D4>>d) data = plane=20 procedures for non-packet LSPs</FONT><BR><BR><BR><FONT size=3D4>-> = no, in=20 this case progression of the end-to-end LSP establishment, implies = that the=20 LSP segment must fulfil a set of constraints carried as part of the = incoming=20 end-to-end LSP request, in this case one has to create an LSP segment = (with a=20 switching type equal to the incoming end-to-end LSP switching type) = over that=20 FA (i.e. referred here above to the LSP segment with a different = Switching=20 Type)</FONT><BR><FONT size=3D4></FONT><BR><BR><FONT = size=3D4>IB>> Agree with=20 the caviat that it should be H-LSP instead of FA-LSP to cover single = and=20 multi-domain scenarios</FONT><BR><BR>Cheers,<BR><BR>Igor<BR><BR><FONT=20 size=3D4>>Note that although 3) and 4) appear to be the same, the = reason I=20 have</FONT><BR><FONT size=3D4>>them as different questions is that = in one=20 case the TE link switching</FONT><BR><FONT size=3D4>>capability is = examined=20 and in the other case the Generalized Label Request</FONT><BR><FONT=20 size=3D4>>is examined.</FONT><BR><BR><BR><BR><FONT size=3D4>-> = what it=20 appears to me is that this discussion mixes the issue of the operation = describing the stitching operations at the domain boundaries from the = LSP=20 segment creation itself i.e. an LSP segment can indeed be an FA link = inherited=20 from the creation of an (intra-domain) FA-LSP -=20 </FONT></P></BLOCKQUOTE></BODY></HTML> ------=_NextPart_000_013A_01C58D3B.CD519570-- Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 18:34:31 +0000 To: "Igor Bryskin" <ibryskin@movaz.com> Cc: "Arthi Ayyangar" <arthi@juniper.net>, <Dimitri.Papadimitriou@alcatel.be>, <ccamp@ops.ietf.org> MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: Re: LSP stitching questions for discussion Date: Wed, 20 Jul 2005 20:33:07 +0200 Message-ID: <OF7ADBF5F6.38B3280C-ONC1257044.0065E822-C1257044.0065E8AF@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmlnb3IgLSA8L0ZPTlQ+PC9QPjxQPiZu YnNwOzwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+aXQgaXMgJnF1b3Q7bm8m cXVvdDsgdG8gYWxsIHRoZSBiZWxvdyBhKSwgYikgYW5kIGMpIC0gYXMgbG9naWNhbCBjb25zZXF1 ZW5jZSBvZiB0aGUgY3VycmVudCBHTVBMUyBhcmNoaXRlY3R1cmU8L0ZPTlQ+PC9QPjxQPiZuYnNw OzwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+bm90ZTogdG8gY2xhcmlmeSBv bmUgb2YgbXkgbm90ZSBoZXJlIGJlbG93ICh3aGVyZSBpdCBzZWVtcyB5b3Ugd2VyZSBub3QgaW4g YWdyZWVtZW50KSBhYm91dCBQU0MgTFNQIGJhbmR3aWR0aCBhbmQgc2VnbWVudCBzdGl0Y2hpbmcg aS5lLiBsYXJnZXIgYmFuZHdpZHRoIHJlc2VydmF0aW9uIGZvciB0aGUgc2VnbWVudCB0aGFuIGZv ciB0aGUgZW5kLXRvLWVuZCBMU1AgZG9lcyBub3QgbWVhbiBkaWZmZXJlbnQgUFNDIGxldmVsIChu b3RlOiBQU0MtMSwgLTIsIC0zIGFuZCAtNCBhcmUgbm90ICZxdW90O2xheWVycyZxdW90OyB0aGV5 IHNob3VsZCBzaW1wbHkgYmUgcmVmZXJyZWQgdG8gYXMgbGV2ZWxzIGJ1dCB0aGlzIGlzIGFub3Ro ZXIgZGlzY3Vzc2lvbikgc28gaSB3YW50ZWQgdG8gY2xhcmlmeSB0aGF0IHRoZSBkZWZhdWx0IGFk bWlzc2lvbiBjb250cm9sIHJ1bGVzIGZvciB0aGUgcGFja2V0IGNhc2UgLSB3aXRoaW4gdGhlIHNh bWUgUFNDIGxldmVsIC0gY291bGQgYmUgcmVsYXhlZCBjb21wYXJlZCB0byB0aGUgb3RoZXIgc3dp dGNoaW5nIHRlY2hub2xvZ2llcyB3aXRob3V0IHN0cmV0Y2hpbmcgdGhlIHN0aXRjaGluZyBjb25j ZXB0PC9GT05UPjxCUj48QlI+PC9QPjxQPjxGT05UIFNJWkU9Mj48Qj4mcXVvdDtJZ29yIEJyeXNr aW4mcXVvdDsgJmx0O2licnlza2luQG1vdmF6LmNvbSZndDs8L0I+PC9GT05UPjxCUj48Rk9OVCBT SVpFPTI+U2VudCBieTogb3duZXItY2NhbXBAb3BzLmlldGYub3JnPC9GT05UPjxCUj48Rk9OVCBT SVpFPTI+MDcvMjAvMjAwNSAxMzoyNSBBU1Q8L0ZPTlQ+PEJSPjxCUj4gPEZPTlQgU0laRT0yPlRv OjwvRk9OVD4gPEZPTlQgU0laRT0yPiZxdW90O0FydGhpIEF5eWFuZ2FyJnF1b3Q7ICZsdDthcnRo aUBqdW5pcGVyLm5ldCZndDssIERpbWl0cmkgUEFQQURJTUlUUklPVS9CRS9BTENBVEVMQEFMQ0FU RUw8L0ZPTlQ+PEJSPiA8Rk9OVCBTSVpFPTI+Y2M6PC9GT05UPiA8Rk9OVCBTSVpFPTI+Jmx0O2Nj YW1wQG9wcy5pZXRmLm9yZyZndDs8L0ZPTlQ+PEJSPiA8Rk9OVCBTSVpFPTI+YmNjOjwvRk9OVD4g PEJSPiA8Rk9OVCBTSVpFPTI+U3ViamVjdDo8L0ZPTlQ+IDxGT05UIFNJWkU9Mj5SZTogTFNQIHN0 aXRjaGluZyBxdWVzdGlvbnMgZm9yIGRpc2N1c3Npb248L0ZPTlQ+PEJSPiA8QlI+PEJSPjwvUD48 UD5EaW1pdHJpLDxCUj48QlI+SXQgc2VlbXMgdG8gbWUgdGhhdCBJIG1vc3RseSBhZ3JlZSg/ISkg d2l0aCB3aGF0IHlvdSB3cm90ZSBpbiB0aGlzIG1haWwuIEknZCBzdGlsbCB3YW50IHRvIGNvbmZp cm0geW91ciB2aWV3cyZuYnNwOyBvbiB0aGUgZm9sbG93aW5nIHF1ZXN0aW9uczo8QlI+PEJSPjxC Uj5hKSBJcyBpdCBwb3NzaWJsZSAoc2hvdWxkIHdlIGFsbG93KSB0byBzdGl0Y2gmbmJzcDsgUFND MSBlMmUgTFNQIGJ5IGEgTFNDIExTUCBzZWdtZW50ID8gPEJSPjxCUj5iKSBJcyBpdCBwb3NzaWJs ZSAoc2hvdWxkIHdlIGFsbG93KSB0byBzdGl0Y2gmbmJzcDsmbmJzcDtWQzEyIGUyZSBMU1AgYnkg YSZuYnNwO1ZDNCBMU1Agc2VnbWVudCA/IDxCUj5jKSBJcyBpdCBwb3NzaWJsZSAoc2hvdWxkIHdl IGFsbG93KSB0byBzdGl0Y2gmbmJzcDsmbmJzcDtQU0MxIGUyZSBMU1AgYnkgYSZuYnNwO1BTQzIg TFNQIHNlZ21lbnQgPzxCUj48QlI+TXkgYW5zd2VycyBhcmUgTk8gdG8gYWxsIHRocmVlIHF1ZXN0 aW9ucy4gWW91IGNhbiBzdGljdGggUFNDMSB3aXRoIFBTQzEsIFZDMTIgd2l0aCBWQzEyLCBldGMu IEluIGFsbCBvdGhlciBjYXNlcyB5b3UgbmVlZCwgZmlyc3QsIHRvIHNldHVwIGEgaGllcmFyY2hp Y2FsIExTUCAoSC1MU1ApICh3aGV0aGVyIGl0IGlzIEZBLUxTUCAoc2luZ2xlIGRvbWFpbikgb3Ig bm90Jm5ic3A7KG11bHRpLWRvbWFpbikpIGluIG9yZGVyIHRvIGNyZWF0ZSBhIGRhdGEgbGluayBv ZiB0aGUgZTJlIExTUCBzd2l0Y2hpbmcgY2FwYWJpbGl0eSBhbmQgb25seSBhZnRlciB0aGF0IHlv dSBjYW4gYWxzbyZuYnNwO3VzZSBhIHN0aXRjaGluZyBzZWdtZW50IG9uIHRvcCBvZiAmbmJzcDt0 aGlzIEgtTFNQKS48QlI+PEJSPlNlZSBhbHNvIGluLWxpbmUuPEJSPjxCUj5JZ29yPEJSPjxCUj48 QlI+PEZPTlQgU0laRT00Pi0tLS0tIE9yaWdpbmFsIE1lc3NhZ2UgLS0tLS0gPC9GT05UPjxCUj48 Rk9OVCBTSVpFPTQ+PEI+RnJvbTo8L0I+IDwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxVRT48 VT48QSBIUkVGPW1haWx0bzpEaW1pdHJpLlBhcGFkaW1pdHJpb3VAYWxjYXRlbC5iZT5EaW1pdHJp LlBhcGFkaW1pdHJpb3VAYWxjYXRlbC5iZTwvQT48L1U+PC9GT05UPjxGT05UIFNJWkU9NCBDT0xP Uj1CTEFDSz4gPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+PEI+VG86PC9CPiA8L0ZPTlQ+PEZPTlQg U0laRT00IENPTE9SPUJMVUU+PFU+PEEgSFJFRj1tYWlsdG86YXJ0aGlAanVuaXBlci5uZXQ+QXJ0 aGkgQXl5YW5nYXI8L0E+PC9VPjwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxBQ0s+IDwvRk9O VD48QlI+PEZPTlQgU0laRT00PjxCPkNjOjwvQj4gPC9GT05UPjxGT05UIFNJWkU9NCBDT0xPUj1C TFVFPjxVPjxBIEhSRUY9bWFpbHRvOmlicnlza2luQG1vdmF6LmNvbT5JZ29yIEJyeXNraW48L0E+ PC9VPjwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxBQ0s+IDsgPC9GT05UPjxGT05UIFNJWkU9 NCBDT0xPUj1CTFVFPjxVPjxBIEhSRUY9bWFpbHRvOmNjYW1wQG9wcy5pZXRmLm9yZz5jY2FtcEBv cHMuaWV0Zi5vcmc8L0E+PC9VPjwvRk9OVD48Rk9OVCBTSVpFPTQgQ09MT1I9QkxBQ0s+IDwvRk9O VD48QlI+PEZPTlQgU0laRT00PjxCPlNlbnQ6PC9CPiBXZWRuZXNkYXksIEp1bHkgMjAsIDIwMDUg NDo0NiBBTTwvRk9OVD48QlI+PEZPTlQgU0laRT00PjxCPlN1YmplY3Q6PC9CPiBSZTogTFNQIHN0 aXRjaGluZyBxdWVzdGlvbnMgZm9yIGRpc2N1c3Npb248L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpF PTQ+YXJ0aGksIDwvRk9OVD48QlI+PEJSPjxCUj48QlI+PEZPTlQgU0laRT00PmZpcnN0IGEgcmVj YXA6IHRoaXMgaXRlbSBvbiBzdGl0Y2hpbmcgd2FzIGluaXRpYXRlZCBpbiBvcmRlciB0byBhbGxv dyBwZXIgJnF1b3Q7ZG9tYWluJnF1b3Q7IGNvbnRyb2wgb2YgZW5kLXRvLWVuZCAoY29udGlndW91 cykgTFNQLCBhcyBwYXJ0IG9mIGEgc2VwYXJhdGUgZG9jdW1lbnQsIGFuZCBub3QgdG8gZXh0ZW5k IGl0cyBzY29wZSAodW5kdWx5LCBpbiB0aGlzIGNhc2UsIHNlZSBiZWxvdykgYnV0IG1haW5seSBp biBvcmRlciB0byBoYXZlIHRoZSBzZXQgb2YgYWxsIHJlbGF0ZWQgcHJvdG9jb2wgbWVjaGFuaXNt cyB3aXRoaW4gYSBzcGVjaWZpYyBkb2N1bWVudCAtPC9GT05UPjxCUj48QlI+PEJSPjxCUj48Rk9O VCBTSVpFPTQ+ZnJvbSB0aGlzIGluaXRpYWwgd29ya2luZyBhc3N1bXB0aW9uLCBpdCBpcyBjbGVh ciB0aGF0IHRoZSBzZWdtZW50IG11c3QgaGF2ZSB0aGUgcHJvcGVydGllcyBvZiB0aGUgaW5jb21p bmcgZW5kLXRvLWVuZCBMU1Agd2hpY2ggYXJlIGRldGVybWluZWQgYW1vbmcgb3RoZXIgYnkgaXRz IHN3aXRjaGluZyB0eXBlIGUuZy4gUFNDLTMgb3IgVERNLCB0aGUgZmlyc3QgcmVxdWlyZW1lbnQg b2YgTFNQIHN0aXRjaGluZyBpcyBjb250aWd1aXR5IG9mIHRoZSBlbmQtdG8tZW5kIExTUCBhdCB0 aGUgZGF0YSBwbGFuZSBsZXZlbCAoYWZ0ZXIgc3RpdGNoaW5nIG9wZXJhdGlvbikgLSBhbmQgdGhp cyBieSBkZWZpbml0aW9uIG9mIHRoZSBpbml0aWFsIHdvcmtpbmcgYXNzdW1wdGlvbiAtPC9GT05U PjxCUj48QlI+PEZPTlQgU0laRT00PklCJmd0OyZndDsgQWdyZWU8L0ZPTlQ+PEJSPjxCUj48Rk9O VCBTSVpFPTQ+bm90ZTogdGhlIG9ubHkgbGF0aXR1ZGUgd2hpY2ggaXMgbGVmdCAoZm9yIGEgZ2l2 ZW4gc3dpdGNoaW5nIHR5cGUgZS5nLiBQU0MtMikgaXMgZHVlIHRvIHRoZSBwcm9wZXJ0aWVzIG9m IHBhY2tldCAoUFNDKSBuZXR3b3JrcyB0aGF0IGFsbG93cyB5b3Ugc3RpdGNoIGEgc2VnbWVudCBo YXZpbmcgYSBsYXJnZXIgTWF4IFJlc2VydmFibGUgQmFuZHdpZHRoIGFuZCBVbnJlc2VydmVkIEJh bmR3aWR0aCAod2Ugd291bGQgaGF2ZSBlcXVpdmFsZW50IGxhdGl0dWRlIGZvciBvdGhlciBhdHRy aWJ1dGVzIGJ1dCB0aGlzIGlzIGEgbG9jYWwgcG9saWN5IGRlY2lzaW9uIC0gd2hpY2ggYXQgdGhl IGVuZCBpcyB0aGUgc29sZSBwdXJwb3NlIG9mIHRoaXMgYXBwcm9hY2gpPC9GT05UPjxCUj48QlI+ PEZPTlQgU0laRT00PklCJmd0OyZndDsgRGlzYWdyZWUuIENvbmNlcHR1YWxseSBQQ1MxIGFuZCBQ U0MyIGFyZSBkaWZmZXJlbnQgbGF5ZXJzIGFuZCB0aGUgbGF0dGVyIGNhbm5vdCBzdGl0Y2ggdGhl IGZvcm1lciwganVzdCBsaWtlIGluIGNhc2UgVkMxMiBhbmQgVkM0LiBZb3Ugc2hvdWxkIG5lc3Qg UFNDMSB3aXRoaW4gUFNDMiBpbnN0ZWFkLjwvRk9OVD48QlI+PEJSPjxCUj48QlI+PEZPTlQgU0la RT00PkhlcmUgYXJlIHNvbWUgb2YgdGhlIHF1ZXN0aW9ucyByZWxhdGVkIHRvIExTUCBzdGl0Y2hp bmcgdG8gc3RhcnQgYTwvRk9OVD48QlI+PEZPTlQgU0laRT00PmRpc2N1c3Npb24gb24gdGhlIHVz YWdlIG9mIExTUCBzdGl0Y2hpbmcuPC9GT05UPjxCUj48QlI+PEZPTlQgU0laRT00PiZndDsxKSBB cmUgdGhlIGNvbnRyb2wgcGxhbmUgc2lnbmFsaW5nIHByb2NlZHVyZXMgZm9yIExTUCBzdGl0Y2hp bmcgYXM8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ZGVzY3JpYmVkIGluIHRoaXMgSUQsIChl eHBsaWNpdCByZXF1ZXN0IGZvciBMU1Agc3RpdGNoaW5nIG9uPC9GT05UPjxCUj48Rk9OVCBTSVpF PTQ+Jmd0O0xTUCBzZWdtZW50LCBkaWZmZXJlbnQgbGFiZWwgYWxsb2NhdGlvbiBydWxlczsgZXRj KSBSRVFVSVJFRCB0byBzdGl0Y2g8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7cGFja2V0IExT UHMgaW4gZGF0YSBwbGFuZSA/PC9GT05UPjxCUj48QlI+PEJSPjxGT05UIFNJWkU9ND4tJmd0OyB5 ZXMgKGJlc2lkZSBtYW51YWwgY29uZmlnIG9wZXJhdGlvbik8L0ZPTlQ+PEJSPjxCUj48Rk9OVCBT SVpFPTQ+SUImZ3Q7Jmd0OyBBZ3JlZTwvRk9OVD48QlI+PEJSPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0 OzIpIEFyZSB0aGUgY29udHJvbCBwbGFuZSBzaWduYWxpbmcgcHJvY2VkdXJlcyBmb3IgTFNQIHN0 aXRjaGluZyBhczwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDtkZXNjcmliZWQgaW4gdGhpcyBJ RCwgKGV4cGxpY2l0IHJlcXVlc3QgZm9yIExTUCBzdGl0Y2hpbmcgb248L0ZPTlQ+PEJSPjxGT05U IFNJWkU9ND4mZ3Q7TFNQIHNlZ21lbnQsIGRpZmZlcmVudCBsYWJlbCBhbGxvY2F0aW9uIHJ1bGVz OyBldGMpIFJFUVVJUkVEIHRvIHN0aXRjaDwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDtub24t cGFja2V0IExTUHMgaW4gZGF0YSBwbGFuZSA/PC9GT05UPjxCUj48QlI+PEJSPjxGT05UIFNJWkU9 ND4tJmd0OyB5ZXMgKGJlc2lkZSBtYW51YWwgY29uZmlnIG9wZXJhdGlvbik8L0ZPTlQ+PEJSPjxC Uj48Rk9OVCBTSVpFPTQ+SUImZ3Q7Jmd0OyBBZ3JlZTwvRk9OVD48QlI+PEJSPjxCUj48Rk9OVCBT SVpFPTQ+Jmd0OzMpIElmIGFuIGUyZSBMU1AgY3Jvc3NlcyBhIHJlZ2lvbiBib3VuZGFyeSAoYmFz ZWQgb24gZGlmZmVyZW50IFRFPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0O2xpbmsgc3dpdGNo aW5nIGNhcGFiaWxpdGllcyksIHRoZW4gY2FuIEkgdXNlIExTUCBzdGl0Y2hpbmcgLTwvRk9OVD48 QlI+PEZPTlQgU0laRT00PiZndDthKSBjb250cm9sIHBsYW5lIHByb2NlZHVyZXMgZm9yIHBhY2tl dCBMU1BzPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0O2IpIGRhdGEgcGxhbmUgcHJvY2VkdXJl cyBmb3IgcGFja2V0IExTUHM8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7YykgY29udHJvbCBw bGFuZSBwcm9jZWR1cmVzIGZvciBub24tcGFja2V0IExTUHM8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9 ND4mZ3Q7ZCkgZGF0YSBwbGFuZSBwcm9jZWR1cmVzIGZvciBub24tcGFja2V0IExTUHM8L0ZPTlQ+ PEJSPjxCUj48QlI+PEZPTlQgU0laRT00Pi0mZ3Q7IG5vLCB3aGVuIGNyb3NzaW5nIGEgcmVnaW9u IGJvdW5kYXJ5LCB5b3Ugc2ltcGx5IHVzZSBhIGRvY3VtZW50IG5hbWVkICZxdW90O2RyYWZ0LW1w bHMtbHNwLWhpZXJhcmNoeS0wOC50eHQmcXVvdDsgd2hpY2ggaXMgaW4gdGhlIFJGQyBFZGl0b3Ig cXVldWUsIGJ1dCB0aGlzIGRvZXMgbm90IHByZXZlbnQgZnJvbSBjcmVhdGluZyBhbiBMU1Agc2Vn bWVudCBvbiB0b3Agb2YgYW4gRkEtTFNQIChzZWUgYmVsb3cpPC9GT05UPjxCUj48QlI+PEZPTlQg U0laRT00PklCJmd0OyZndDsgQWdyZWUgd2l0aCB0aGUgY2F2aWF0IHRoYXQgaXQgc2hvdWxkIGJl IEgtTFNQIGluc3RlYWQgb2YgRkEtTFNQIHRvIGNvdmVyIHNpbmdsZSBhbmQgbXVsdGktZG9tYWlu IHNjZW5hcmlvczwvRk9OVD48QlI+PEJSPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0OzQpIElmIHRoZSBz d2l0Y2hpbmcgdHlwZSBvZiBhbiBlMmUgTFNQIGlzIGRpZmZlcmVudCBmcm9tIHRoYXQgb2YgYW4g TFNQPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0O3NlZ21lbnQsIGNhbiBJIHVzZSBMU1Agc3Rp dGNoaW5nIDwvRk9OVD48QlI+PEZPTlQgU0laRT00PiZndDthKSBjb250cm9sIHBsYW5lIHByb2Nl ZHVyZXMgZm9yIHBhY2tldCBMU1BzPC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0O2IpIGRhdGEg cGxhbmUgcHJvY2VkdXJlcyBmb3IgcGFja2V0IExTUHM8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4m Z3Q7YykgY29udHJvbCBwbGFuZSBwcm9jZWR1cmVzIGZvciBub24tcGFja2V0IExTUHM8L0ZPTlQ+ PEJSPjxGT05UIFNJWkU9ND4mZ3Q7ZCkgZGF0YSBwbGFuZSBwcm9jZWR1cmVzIGZvciBub24tcGFj a2V0IExTUHM8L0ZPTlQ+PEJSPjxCUj48QlI+PEZPTlQgU0laRT00Pi0mZ3Q7IG5vLCBpbiB0aGlz IGNhc2UgcHJvZ3Jlc3Npb24gb2YgdGhlIGVuZC10by1lbmQgTFNQIGVzdGFibGlzaG1lbnQsIGlt cGxpZXMgdGhhdCB0aGUgTFNQIHNlZ21lbnQgbXVzdCBmdWxmaWwgYSBzZXQgb2YgY29uc3RyYWlu dHMgY2FycmllZCBhcyBwYXJ0IG9mIHRoZSBpbmNvbWluZyBlbmQtdG8tZW5kIExTUCByZXF1ZXN0 LCBpbiB0aGlzIGNhc2Ugb25lIGhhcyB0byBjcmVhdGUgYW4gTFNQIHNlZ21lbnQgKHdpdGggYSBz d2l0Y2hpbmcgdHlwZSBlcXVhbCB0byB0aGUgaW5jb21pbmcgZW5kLXRvLWVuZCBMU1Agc3dpdGNo aW5nIHR5cGUpIG92ZXIgdGhhdCBGQSAoaS5lLiByZWZlcnJlZCBoZXJlIGFib3ZlIHRvIHRoZSBM U1Agc2VnbWVudCB3aXRoIGEgZGlmZmVyZW50IFN3aXRjaGluZyBUeXBlKTwvRk9OVD48QlI+PEZP TlQgU0laRT00PiA8L0ZPTlQ+PEJSPjxCUj48Rk9OVCBTSVpFPTQ+SUImZ3Q7Jmd0OyBBZ3JlZSB3 aXRoIHRoZSBjYXZpYXQgdGhhdCBpdCBzaG91bGQgYmUgSC1MU1AgaW5zdGVhZCBvZiBGQS1MU1Ag dG8gY292ZXIgc2luZ2xlIGFuZCBtdWx0aS1kb21haW4gc2NlbmFyaW9zPC9GT05UPjxCUj48QlI+ Q2hlZXJzLDxCUj48QlI+SWdvcjxCUj48QlI+PEZPTlQgU0laRT00PiZndDtOb3RlIHRoYXQgYWx0 aG91Z2ggMykgYW5kIDQpIGFwcGVhciB0byBiZSB0aGUgc2FtZSwgdGhlIHJlYXNvbiBJIGhhdmU8 L0ZPTlQ+PEJSPjxGT05UIFNJWkU9ND4mZ3Q7dGhlbSBhcyBkaWZmZXJlbnQgcXVlc3Rpb25zIGlz IHRoYXQgaW4gb25lIGNhc2UgdGhlIFRFIGxpbmsgc3dpdGNoaW5nPC9GT05UPjxCUj48Rk9OVCBT SVpFPTQ+Jmd0O2NhcGFiaWxpdHkgaXMgZXhhbWluZWQgYW5kIGluIHRoZSBvdGhlciBjYXNlIHRo ZSBHZW5lcmFsaXplZCBMYWJlbCBSZXF1ZXN0PC9GT05UPjxCUj48Rk9OVCBTSVpFPTQ+Jmd0O2lz IGV4YW1pbmVkLjwvRk9OVD48QlI+PEJSPjxCUj48QlI+PEZPTlQgU0laRT00Pi0mZ3Q7IHdoYXQg aXQgYXBwZWFycyB0byBtZSBpcyB0aGF0IHRoaXMgZGlzY3Vzc2lvbiBtaXhlcyB0aGUgaXNzdWUg b2YgdGhlIG9wZXJhdGlvbiBkZXNjcmliaW5nIHRoZSBzdGl0Y2hpbmcgb3BlcmF0aW9ucyBhdCB0 aGUgZG9tYWluIGJvdW5kYXJpZXMgZnJvbSB0aGUgTFNQIHNlZ21lbnQgY3JlYXRpb24gaXRzZWxm IGkuZS4gYW4gTFNQIHNlZ21lbnQgY2FuIGluZGVlZCBiZSBhbiBGQSBsaW5rIGluaGVyaXRlZCBm cm9tIHRoZSBjcmVhdGlvbiBvZiBhbiAoaW50cmEtZG9tYWluKSBGQS1MU1AgLSA8L0ZPTlQ+PC9Q Pg== Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 18:19:12 +0000 DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=Message-ID:Received:Date:From:Subject:To:Cc:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding; b=fGPZubv8Y3uwT5SE0F//pveBtaHtXUnQzoXBovOI7+BZUCJe0GIMMYsWeHrGA0jUFYKDyH+yfmd/8uBt0n6zCLx6QB76CnGi91Thigp8ksrz0mQC1f4fo8tOfdAvIx8mNq+HqL/rbvi29uQ0o6yLYSn3kR4pnKYDKBP/Hqd7wYQ= ; Message-ID: <20050720181838.1056.qmail@web30811.mail.mud.yahoo.com> Date: Wed, 20 Jul 2005 11:18:37 -0700 (PDT) From: Igor Bryskin <i_bryskin@yahoo.com> Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt To: Arthi Ayyangar <arthi@juniper.net>, Igor Bryskin <ibryskin@movaz.com> Cc: ccamp@ops.ietf.org MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="0-512571639-1121883517=:99174" Content-Transfer-Encoding: 8bit --0-512571639-1121883517=:99174 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit Thanks, sounds like a good idea. Igor Arthi Ayyangar <arthi@juniper.net> wrote: Hi Igor, > One more comment. I think the draft should give a clear and unambiguous > answer to how e2e LSP and a stitching LSP (S-LSP) relate from the switching > capability point of view. Specifically, should they belong to the same > network layer, should they share the same switching type or maybe there are > no limitations at all, and you can stitch your e2e LSP pretty much with > anything you got, provided that this "anything" has enough bandwidth. > > Let me give you a specific example. Is it possible to stitch a PSC1 e2e LSP > with a PSC2 S-LSP? You probably remember our discussion with Kireeti in > Minneapolis (and you were also part of it), and he said (and I 100% agree > with him) that although it is doable (you can do anything you want with the > label stack), it is not right and this should be logically prohibited. What > you should do in this case is to nest the e2e LSP within the PSC2 LSP (that > is, use it as H-LSP) even in case it consumes the entire PSC2 LSP bandwidth. ---> I do remember this discussion. > It is more clear in non-PSC cases. Although it is possible to "stitch" an > e2e VC12 LSP with a VC4 S-LSP, I could not imagine that a network operator > would ever agree with such solution. He is likely to say " What ? If the > service is supposed to stay for 12 month, than for entire year I pretty much > burry my 62 VC12 channels - I cannot use them even conceptually in case > when, say, on the next day I'll have to provide another 62 VC12 services > that could've used these wasted resources" > > On the other hand I know plenty GMPLS people who would say: "Sure, no > problem. Of course we can stitch a PSC e2e LSP with lambda S-LSP: If > somebody wants to do a trick stupid as it is, why should we prevent him from > doing so?" ----> I know some of these people too. :-) > The bottom line is that a discussion like this in the draft and consensus > on how we see the stitching IMO would be extremely valuable for the GMPLS > multi-layer story, and I don't see a reason why this could not be achieved > in the frame of this work. ------> I agree that we should have a discussion on this, but I would like folks to have this discussion on the list first, which I thank you for initiating. Yes, the draft has to state clearly whether LSP stitching between different switching capabilities is allowed or not. But it's more than that. Currently, the draft talks about this under sections 3.1 and 4.1 and 4.1.2. But, let us have this discussion on the list and we will be glad to add what's missing (with required details) to the ID, depending on the consensus reached. I will send out the questions to start the discussion in the next email. thanks, -arthi > > ----- Original Message ----- > From: "Arthi Ayyangar" > To: "Igor Bryskin" > Cc: ; ; > > Sent: Monday, July 18, 2005 2:51 PM > Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > > > > Hi Igor, > > > > Please see inline. > > > > > 1. Why are saying that LSP Stitching is a private case of LSP Hierarchy? > > > > > there is more differences than similarities: > > > The differences are: > > > 1) In case of H-LSP there is a data plane adjacency, while in case of > S-LSP > > > there is none (as you correctly pointed out); > > > 2) In case of H-LSP there is an adaptation in data plane (label push/pop > for > > > PSC), while in case of S-LSP there is none - just simple > cross-connecting > > > ( label swap) as in case of two "native " e2e LSP adjacent segments ; > > > 3) H-LSP could be used by many e2e LSPs, while S-LSP could be used by > > > exactly one e2e LSP > > > 4) Signaling is different - there is no label negotiation in stitching > > > 5) H-LSP is used as a "true" data link, specifically there is a resource > > > allocation on the H-LSP edges, while in case of S-LSP there is none > > > 6) From MLN point of view, H-LSP is created in a server (lower) layer, > while > > > the S-LSP is created in the client (same as e2e LSP) layer. > > > > > > There are two similarities that I can think of: > > > 1) There is a signaling and possibly routing (see below) adjacencies > between > > > the ends; > > > 2) Both H-LSP and S-LSP could be advertised as separate TE links or as > TE > > > bundles > > > > > > I would recommend to dedicate a paragraph and enlist there similarities > and > > > differencies > > -------> Based on the discussions on the list over the last revision, the > > message that the authors received was that we do not want to go about such > > a lengthy discussion of similarities and differences (NOTE that the draft > > already does clearly highlight them where applicable). Instead we want > > this ID to simply explain how LSP stitching functions. In other words the > > idea was to make this ID a complete document by itself. But since it does > > borrow concepts from the LSP hierarchy ID, just state the concepts that > > are applicable or inapplicable. > > > > > 2. Why are you saying that a TE Link based on S-LSP can be used for > exactly > > > one e2e LSP? Parallel S-LSPs could be advertised as a single TE link > > > (bundle, see above) and hence can accomadate several e2e LSPs. > > ----> Igor, the draft does talk about this Bundling case. > > > > > 3. You are saying that S-LSP does not have a routing peering. Actually, > in > > > this respect it is no different from H-LSP: if it is advertised as a TE > link > > > into the same TE domain that was used for S-LSP creation (unlikely IMO > > > scenario) than it does not require the routing adjacency (in other > words, it > > > is an FA according to LSP-HIER definition), otherwise, it IS NOT and FA > and > > > does require the direct routing peering in the domain it is advertised > to > > > make it useful as a TE link in this domain > > -----> I am sorry, but I am missing your point here completely. Are you > > saying that the statement "the end points of an LSP segment do not have a > > routing adjacency", is incorrect or are you saying that "this is > > obvious" ? It is unclear to me what exactly your argument is. Please > > clarify. > > > > thanks, > > -arthi > > > > > > > > Cheers, > > > Igor > > > > > > > > > > > > ----- Original Message ----- > > > From: > > > To: > > > Cc: > > > Sent: Friday, July 15, 2005 3:50 PM > > > Subject: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > > > > > > > > > > A New Internet-Draft is available from the on-line Internet-Drafts > > > directories. > > > > This draft is a work item of the Common Control and Measurement Plane > > > Working Group of the IETF. > > > > > > > > Title : Label Switched Path Stitching with Generalized MPLS Traffic > > > Engineering > > > > Author(s) : A. Ayyangar, J. Vasseur > > > > Filename : draft-ietf-ccamp-lsp-stitching-01.txt > > > > Pages : 19 > > > > Date : 2005-7-15 > > > > > > > > In certain scenarios, there may be a need to combine together two > > > > different Generalized Multi-Protocol Label Switching (GMPLS) Label > > > > Switched Paths (LSPs) such that in the data plane, a single end-to- > > > > end (e2e) LSP is achieved and all traffic from one LSP is switched > > > > onto the other LSP. We will refer to this as "LSP stitching". > This > > > > document covers cases where: a) the node performing the stitching > > > > does not require configuration of every LSP pair to be stitched > > > > together b) the node performing the stitching is not the egress of > > > > any of the LSPs c) LSP stitching not only results in an end-to-end > > > > LSP in the data plane, but there is also a corresponding end-to-end > > > > LSP (RSVP session) in the control plane. It might be possible to > > > > configure a GMPLS node to switch the traffic from an LSP for which > it > > > > is the egress, to another LSP for which it is the ingress, without > > > > requiring any signaling or routing extensions whatsoever, > completely > > > > transparent to other nodes. This will also result in LSP stitching > > > > in the data plane. However, this document does not cover this > > > > scenario of LSP stitching. > > > > > > > > A URL for this Internet-Draft is: > > > > > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt > > > > > > > > To remove yourself from the I-D Announcement list, send a message to > > > > i-d-announce-request@ietf.org with the word unsubscribe in the body of > the > > > message. > > > > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > > > > to change your subscription settings. > > > > > > > > > > > > Internet-Drafts are also available by anonymous FTP. Login with the > > > username > > > > "anonymous" and a password of your e-mail address. After logging in, > > > > type "cd internet-drafts" and then > > > > "get draft-ietf-ccamp-lsp-stitching-01.txt". > > > > > > > > A list of Internet-Drafts directories can be found in > > > > http://www.ietf.org/shadow.html > > > > or ftp://ftp.ietf.org/ietf/1shadow-sites.txt > > > > > > > > > > > > Internet-Drafts can also be obtained by e-mail. > > > > > > > > Send a message to: > > > > mailserv@ietf.org. > > > > In the body type: > > > > "FILE /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt". > > > > > > > > NOTE: The mail server at ietf.org can return the document in > > > > MIME-encoded form by using the "mpack" utility. To use this > > > > feature, insert the command "ENCODING mime" before the "FILE" > > > > command. To decode the response(s), you will need "munpack" or > > > > a MIME-compliant mail reader. Different MIME-compliant mail readers > > > > exhibit different behavior, especially when dealing with > > > > "multipart" MIME messages (i.e. documents which have been split > > > > up into multiple messages), so check your local documentation on > > > > how to manipulate these messages. > > > > > > > > > > > > Below is the data which will enable a MIME compliant mail reader > > > > implementation to automatically retrieve the ASCII version of the > > > > Internet-Draft. > > > > > > > > > > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com --0-512571639-1121883517=:99174 Content-Type: text/html; charset=iso-8859-1 Content-Transfer-Encoding: 8bit <DIV>Thanks, sounds like a good idea.</DIV> <DIV> </DIV> <DIV>Igor<BR><BR><B><I>Arthi Ayyangar <arthi@juniper.net></I></B> wrote:</DIV> <BLOCKQUOTE class=replbq style="PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #1010ff 2px solid">Hi Igor,<BR><BR>> One more comment. I think the draft should give a clear and unambiguous<BR>> answer to how e2e LSP and a stitching LSP (S-LSP) relate from the switching<BR>> capability point of view. Specifically, should they belong to the same<BR>> network layer, should they share the same switching type or maybe there are<BR>> no limitations at all, and you can stitch your e2e LSP pretty much with<BR>> anything you got, provided that this "anything" has enough bandwidth.<BR>><BR>> Let me give you a specific example. Is it possible to stitch a PSC1 e2e LSP<BR>> with a PSC2 S-LSP? You probably remember our discussion with Kireeti in<BR>> Minneapolis (and you were also part of it), and he said (and I 100% agree<BR>> with him) that although it is doable (you can do anything you want with the<BR>> label stack), it is not right and this should be logically prohibited. What<BR>> you should do in this case is to nest the e2e LSP within the PSC2 LSP (that<BR>> is, use it as H-LSP) even in case it consumes the entire PSC2 LSP bandwidth.<BR><AA>---> I do remember this discussion.<BR><BR><BR>> It is more clear in non-PSC cases. Although it is possible to "stitch" an<BR>> e2e VC12 LSP with a VC4 S-LSP, I could not imagine that a network operator<BR>> would ever agree with such solution. He is likely to say " What ? If the<BR>> service is supposed to stay for 12 month, than for entire year I pretty much<BR>> burry my 62 VC12 channels - I cannot use them even conceptually in case<BR>> when, say, on the next day I'll have to provide another 62 VC12 services<BR>> that could've used these wasted resources"<BR>><BR>> On the other hand I know plenty GMPLS people who would say: "Sure, no<BR>> problem. Of course we can stitch a PSC e2e LSP with lambda S-LSP: If<BR>> somebody wants to do a trick stupid as it is, why should we prevent him from<BR>> doing so?"<BR><AA>----> I know some of these people too. :-)<BR><BR>> The bottom line is that a discussion like this in the draft and consensus<BR>> on how we see the stitching IMO would be extremely valuable for the GMPLS<BR>> multi-layer story, and I don't see a reason why this could not be achieved<BR>> in the frame of this work.<BR><AA>------> I agree that we should have a discussion on this, but I would<BR>like folks to have this discussion on the list first, which I thank<BR>you for initiating.<BR><BR>Yes, the draft has to state clearly whether LSP stitching between<BR>different switching capabilities is allowed or not. But it's more than<BR>that. Currently, the draft talks about this under sections 3.1 and 4.1<BR>and 4.1.2.<BR><BR>But, let us have this discussion on the list and we will be glad to add<BR>what's missing (with required details) to the ID, depending on the<BR>consensus reached. I will send out the questions to start the discussion<BR>in the next email.<BR><BR>thanks,<BR>-arthi<BR><BR>><BR>> ----- Original Message -----<BR>> From: "Arthi Ayyangar" <ARTHI@JUNIPER.NET><BR>> To: "Igor Bryskin" <IBRYSKIN@MOVAZ.COM><BR>> Cc: <I-D-ANNOUNCE@IETF.ORG>; <INTERNET-DRAFTS@IETF.ORG>;<BR>> <CCAMP@OPS.IETF.ORG><BR>> Sent: Monday, July 18, 2005 2:51 PM<BR>> Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt<BR>><BR>><BR>> > Hi Igor,<BR>> ><BR>> > Please see inline.<BR>> ><BR>> > > 1. Why are saying that LSP Stitching is a private case of LSP Hierarchy?<BR>> ><BR>> > > there is more differences than similarities:<BR>> > > The differences are:<BR>> > > 1) In case of H-LSP there is a data plane adjacency, while in case of<BR>> S-LSP<BR>> > > there is none (as you correctly pointed out);<BR>> > > 2) In case of H-LSP there is an adaptation in data plane (label push/pop<BR>> for<BR>> > > PSC), while in case of S-LSP there is none - just simple<BR>> cross-connecting<BR>> > > ( label swap) as in case of two "native " e2e LSP adjacent segments ;<BR>> > > 3) H-LSP could be used by many e2e LSPs, while S-LSP could be used by<BR>> > > exactly one e2e LSP<BR>> > > 4) Signaling is different - there is no label negotiation in stitching<BR>> > > 5) H-LSP is used as a "true" data link, specifically there is a resource<BR>> > > allocation on the H-LSP edges, while in case of S-LSP there is none<BR>> > > 6) From MLN point of view, H-LSP is created in a server (lower) layer,<BR>> while<BR>> > > the S-LSP is created in the client (same as e2e LSP) layer.<BR>> > ><BR>> > > There are two similarities that I can think of:<BR>> > > 1) There is a signaling and possibly routing (see below) adjacencies<BR>> between<BR>> > > the ends;<BR>> > > 2) Both H-LSP and S-LSP could be advertised as separate TE links or as<BR>> TE<BR>> > > bundles<BR>> > ><BR>> > > I would recommend to dedicate a paragraph and enlist there similarities<BR>> and<BR>> > > differencies<BR>> > -------> Based on the discussions on the list over the last revision, the<BR>> > message that the authors received was that we do not want to go about such<BR>> > a lengthy discussion of similarities and differences (NOTE that the draft<BR>> > already does clearly highlight them where applicable). Instead we want<BR>> > this ID to simply explain how LSP stitching functions. In other words the<BR>> > idea was to make this ID a complete document by itself. But since it does<BR>> > borrow concepts from the LSP hierarchy ID, just state the concepts that<BR>> > are applicable or inapplicable.<BR>> ><BR>> > > 2. Why are you saying that a TE Link based on S-LSP can be used for<BR>> exactly<BR>> > > one e2e LSP? Parallel S-LSPs could be advertised as a single TE link<BR>> > > (bundle, see above) and hence can accomadate several e2e LSPs.<BR>> > ----> Igor, the draft does talk about this Bundling case.<BR>> ><BR>> > > 3. You are saying that S-LSP does not have a routing peering. Actually,<BR>> in<BR>> > > this respect it is no different from H-LSP: if it is advertised as a TE<BR>> link<BR>> > > into the same TE domain that was used for S-LSP creation (unlikely IMO<BR>> > > scenario) than it does not require the routing adjacency (in other<BR>> words, it<BR>> > > is an FA according to LSP-HIER definition), otherwise, it IS NOT and FA<BR>> and<BR>> > > does require the direct routing peering in the domain it is advertised<BR>> to<BR>> > > make it useful as a TE link in this domain<BR>> > -----> I am sorry, but I am missing your point here completely. Are you<BR>> > saying that the statement "the end points of an LSP segment do not have a<BR>> > routing adjacency", is incorrect or are you saying that "this is<BR>> > obvious" ? It is unclear to me what exactly your argument is. Please<BR>> > clarify.<BR>> ><BR>> > thanks,<BR>> > -arthi<BR>> ><BR>> > ><BR>> > > Cheers,<BR>> > > Igor<BR>> > ><BR>> > ><BR>> > ><BR>> > > ----- Original Message -----<BR>> > > From: <INTERNET-DRAFTS@IETF.ORG><BR>> > > To: <I-D-ANNOUNCE@IETF.ORG><BR>> > > Cc: <CCAMP@OPS.IETF.ORG><BR>> > > Sent: Friday, July 15, 2005 3:50 PM<BR>> > > Subject: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt<BR>> > ><BR>> > ><BR>> > > > A New Internet-Draft is available from the on-line Internet-Drafts<BR>> > > directories.<BR>> > > > This draft is a work item of the Common Control and Measurement Plane<BR>> > > Working Group of the IETF.<BR>> > > ><BR>> > > > Title : Label Switched Path Stitching with Generalized MPLS Traffic<BR>> > > Engineering<BR>> > > > Author(s) : A. Ayyangar, J. Vasseur<BR>> > > > Filename : draft-ietf-ccamp-lsp-stitching-01.txt<BR>> > > > Pages : 19<BR>> > > > Date : 2005-7-15<BR>> > > ><BR>> > > > In certain scenarios, there may be a need to combine together two<BR>> > > > different Generalized Multi-Protocol Label Switching (GMPLS) Label<BR>> > > > Switched Paths (LSPs) such that in the data plane, a single end-to-<BR>> > > > end (e2e) LSP is achieved and all traffic from one LSP is switched<BR>> > > > onto the other LSP. We will refer to this as "LSP stitching".<BR>> This<BR>> > > > document covers cases where: a) the node performing the stitching<BR>> > > > does not require configuration of every LSP pair to be stitched<BR>> > > > together b) the node performing the stitching is not the egress of<BR>> > > > any of the LSPs c) LSP stitching not only results in an end-to-end<BR>> > > > LSP in the data plane, but there is also a corresponding end-to-end<BR>> > > > LSP (RSVP session) in the control plane. It might be possible to<BR>> > > > configure a GMPLS node to switch the traffic from an LSP for which<BR>> it<BR>> > > > is the egress, to another LSP for which it is the ingress, without<BR>> > > > requiring any signaling or routing extensions whatsoever,<BR>> completely<BR>> > > > transparent to other nodes. This will also result in LSP stitching<BR>> > > > in the data plane. However, this document does not cover this<BR>> > > > scenario of LSP stitching.<BR>> > > ><BR>> > > > A URL for this Internet-Draft is:<BR>> > > ><BR>> http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt<BR>> > > ><BR>> > > > To remove yourself from the I-D Announcement list, send a message to<BR>> > > > i-d-announce-request@ietf.org with the word unsubscribe in the body of<BR>> the<BR>> > > message.<BR>> > > > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce<BR>> > > > to change your subscription settings.<BR>> > > ><BR>> > > ><BR>> > > > Internet-Drafts are also available by anonymous FTP. Login with the<BR>> > > username<BR>> > > > "anonymous" and a password of your e-mail address. After logging in,<BR>> > > > type "cd internet-drafts" and then<BR>> > > > "get draft-ietf-ccamp-lsp-stitching-01.txt".<BR>> > > ><BR>> > > > A list of Internet-Drafts directories can be found in<BR>> > > > http://www.ietf.org/shadow.html<BR>> > > > or ftp://ftp.ietf.org/ietf/1shadow-sites.txt<BR>> > > ><BR>> > > ><BR>> > > > Internet-Drafts can also be obtained by e-mail.<BR>> > > ><BR>> > > > Send a message to:<BR>> > > > mailserv@ietf.org.<BR>> > > > In the body type:<BR>> > > > "FILE /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt".<BR>> > > ><BR>> > > > NOTE: The mail server at ietf.org can return the document in<BR>> > > > MIME-encoded form by using the "mpack" utility. To use this<BR>> > > > feature, insert the command "ENCODING mime" before the "FILE"<BR>> > > > command. To decode the response(s), you will need "munpack" or<BR>> > > > a MIME-compliant mail reader. Different MIME-compliant mail readers<BR>> > > > exhibit different behavior, especially when dealing with<BR>> > > > "multipart" MIME messages (i.e. documents which have been split<BR>> > > > up into multiple messages), so check your local documentation on<BR>> > > > how to manipulate these messages.<BR>> > > ><BR>> > > ><BR>> > > > Below is the data which will enable a MIME compliant mail reader<BR>> > > > implementation to automatically retrieve the ASCII version of the<BR>> > > > Internet-Draft.<BR>> > > ><BR>> > ><BR>> ><BR>><BR><BR></BLOCKQUOTE><p>__________________________________________________<br>Do You Yahoo!?<br>Tired of spam? Yahoo! Mail has the best spam protection around <br>http://mail.yahoo.com --0-512571639-1121883517=:99174-- Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 17:47:34 +0000 Message-ID: <00de01c58d53$06b9f7a0$7a1810ac@movaz.com> From: "Igor Bryskin" <ibryskin@movaz.com> To: "Stephen Shew" <sdshew@nortel.com>, <ccamp@ops.ietf.org> Subject: Re: Date: Wed, 20 Jul 2005 13:46:58 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Stephen, Great to hear from you. I agree with what you said. As you correctly pointed out, there are cases already when provisioning of an LSP does not require operations in the data plane (e.g. make-before-break). In these cases, however, the CP is aware that necessary data plane related provisioning is in-place already. When we move ownership from MP to CP, the procedures from the CP point of view are almost the same - one exception is that the CP is not aware of the fact that the network resources are in place already, and with simple signaling extensions it is possible to achieve the goal with very little pain and as reliable as make-before-break procedure is. One other similar example. When the CP on some controller is re-synched (after the crash or software upgrade) the situation is almost identical - the crossconnects are in place already, the only thing to do is to restore the CP states. The only difference here. though, that the CP can figure out by itself that resources are in-place Igor ----- Original Message ----- From: "Stephen Shew" <sdshew@nortel.com> To: <ccamp@ops.ietf.org> Sent: Wednesday, July 20, 2005 5:54 AM Adrian, see comment below. > -----Original Message----- > From: owner-ccamp@ops.ietf.org > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Adrian Farrel > Sent: Tuesday, July 19, 2005 10:02 > To: George Newsome > Cc: ccamp@ops.ietf.org > Subject: Re: Moving LSP ownership between control plane and > management plane > > > Thanks George, > > Useful input. <snip> > > - Call awareness of migration vs. connection being unaware > of migration. > > When connection state is being created to match an existing > > connection, the connection controllers (CCs) do not require > awareness > > of why this is happening as the context could be migration or > > recovery. A mechanism to create control state without affecting > > transport plane state is needed in the CCs. Call > controllers though > > do need migration awareness as they need to obtain/derive call > > identification from the MP. > > I disagree. > Connection controllers must be able to distinguish between > resources that are already in use and cannot be assigned to a > new connection, and those that are already in use and can be > assigned to a new connection because the connection is > replacing an existing connection. If this were happening > entirely within the control plane we would use resource > sharing, but when the operation spans the control and > management planes resource ownership/sharing is more complex > and the conneciotn controller needs to be able to issue the > approriate instrucitons for the assignment of the resources. > On the other hand, call controllers do not need awareness of > the details of the realisation of the connections. I'm not sure if the difference between migration awareness and the action of creating connection state for an existing transport connection is clear in the original point. I agree with your first sentence and think that signalling could be extended to include the case where connection details (in a complete ERO) are supplied with the attribute that the cross connections are already in place. Similarly, teardown of connection state can occur where cross connections are not undone. I still agree though with the original point and note that it does not say that call controllers are aware of connection details. Call controllers are aware that a migration action is the reason for triggering the connection controller to establish state. The point about awareness refers to an understanding of why LSP state is being created. If existing LSP state is being replaced by another one (e.g., a recovery scenario or maybe a change of the extent of the transport connection), then I think it's sufficient for the connection controller to be told that it should not change the cross connection (matrix level subnetwork connection). If the overall action fails for some reason, the call entity can then decide whether to try again, and/or tell the management plane what happened. I believe that this is consistent with what the existing protocol(s) know when they are creating LSPs. For example, a protocol instance is not aware whether the connection being created is a result of a "break before make" or "make before break" scenario, or that it is the 3rd connection being created for a VCAT group. This understanding should only be known by the initiator of the action. Stephen Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 17:26:27 +0000 Message-ID: <00be01c58d50$00f56730$7a1810ac@movaz.com> From: "Igor Bryskin" <ibryskin@movaz.com> To: "Arthi Ayyangar" <arthi@juniper.net>, <Dimitri.Papadimitriou@alcatel.be> Cc: <ccamp@ops.ietf.org> Subject: Re: LSP stitching questions for discussion Date: Wed, 20 Jul 2005 13:25:19 -0400 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_00BB_01C58D2E.79B452C0" This is a multi-part message in MIME format. ------=_NextPart_000_00BB_01C58D2E.79B452C0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Dimitri, It seems to me that I mostly agree(?!) with what you wrote in this mail. = I'd still want to confirm your views on the following questions: a) Is it possible (should we allow) to stitch PSC1 e2e LSP by a LSC LSP = segment ? b) Is it possible (should we allow) to stitch VC12 e2e LSP by a VC4 LSP = segment ? c) Is it possible (should we allow) to stitch PSC1 e2e LSP by a PSC2 = LSP segment ? My answers are NO to all three questions. You can sticth PSC1 with PSC1, = VC12 with VC12, etc. In all other cases you need, first, to setup a = hierarchical LSP (H-LSP) (whether it is FA-LSP (single domain) or not = (multi-domain)) in order to create a data link of the e2e LSP switching = capability and only after that you can also use a stitching segment on = top of this H-LSP). See also in-line. Igor ----- Original Message -----=20 From: Dimitri.Papadimitriou@alcatel.be=20 To: Arthi Ayyangar=20 Cc: Igor Bryskin ; ccamp@ops.ietf.org=20 Sent: Wednesday, July 20, 2005 4:46 AM Subject: Re: LSP stitching questions for discussion arthi,=20 first a recap: this item on stitching was initiated in order to allow = per "domain" control of end-to-end (contiguous) LSP, as part of a = separate document, and not to extend its scope (unduly, in this case, = see below) but mainly in order to have the set of all related protocol = mechanisms within a specific document - from this initial working assumption, it is clear that the segment = must have the properties of the incoming end-to-end LSP which are = determined among other by its switching type e.g. PSC-3 or TDM, the = first requirement of LSP stitching is contiguity of the end-to-end LSP = at the data plane level (after stitching operation) - and this by = definition of the initial working assumption - IB>> Agree note: the only latitude which is left (for a given switching type e.g. = PSC-2) is due to the properties of packet (PSC) networks that allows you = stitch a segment having a larger Max Reservable Bandwidth and Unreserved = Bandwidth (we would have equivalent latitude for other attributes but = this is a local policy decision - which at the end is the sole purpose = of this approach) IB>> Disagree. Conceptually PCS1 and PSC2 are different layers and the = latter cannot stitch the former, just like in case VC12 and VC4. You = should nest PSC1 within PSC2 instead. Here are some of the questions related to LSP stitching to start a discussion on the usage of LSP stitching. >1) Are the control plane signaling procedures for LSP stitching as >described in this ID, (explicit request for LSP stitching on >LSP segment, different label allocation rules; etc) REQUIRED to = stitch >packet LSPs in data plane ? -> yes (beside manual config operation) IB>> Agree >2) Are the control plane signaling procedures for LSP stitching as >described in this ID, (explicit request for LSP stitching on >LSP segment, different label allocation rules; etc) REQUIRED to = stitch >non-packet LSPs in data plane ? -> yes (beside manual config operation) IB>> Agree >3) If an e2e LSP crosses a region boundary (based on different TE >link switching capabilities), then can I use LSP stitching - >a) control plane procedures for packet LSPs >b) data plane procedures for packet LSPs >c) control plane procedures for non-packet LSPs >d) data plane procedures for non-packet LSPs -> no, when crossing a region boundary, you simply use a document = named "draft-mpls-lsp-hierarchy-08.txt" which is in the RFC Editor = queue, but this does not prevent from creating an LSP segment on top of = an FA-LSP (see below) IB>> Agree with the caviat that it should be H-LSP instead of FA-LSP = to cover single and multi-domain scenarios >4) If the switching type of an e2e LSP is different from that of an = LSP >segment, can I use LSP stitching - >a) control plane procedures for packet LSPs >b) data plane procedures for packet LSPs >c) control plane procedures for non-packet LSPs >d) data plane procedures for non-packet LSPs -> no, in this case progression of the end-to-end LSP establishment, = implies that the LSP segment must fulfil a set of constraints carried as = part of the incoming end-to-end LSP request, in this case one has to = create an LSP segment (with a switching type equal to the incoming = end-to-end LSP switching type) over that FA (i.e. referred here above to = the LSP segment with a different Switching Type) IB>> Agree with the caviat that it should be H-LSP instead of FA-LSP = to cover single and multi-domain scenarios Cheers, Igor >Note that although 3) and 4) appear to be the same, the reason I have >them as different questions is that in one case the TE link switching >capability is examined and in the other case the Generalized Label = Request >is examined. -> what it appears to me is that this discussion mixes the issue of = the operation describing the stitching operations at the domain = boundaries from the LSP segment creation itself i.e. an LSP segment can = indeed be an FA link inherited from the creation of an (intra-domain) = FA-LSP -=20 ------=_NextPart_000_00BB_01C58D2E.79B452C0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META http-equiv=3DContent-Type content=3D"text/html; = charset=3Diso-8859-1"> <META content=3D"MSHTML 6.00.2800.1400" name=3DGENERATOR> <STYLE></STYLE> </HEAD> <BODY bgColor=3D#ffffff> <DIV><FONT face=3DArial size=3D2>Dimitri,</FONT></DIV> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <DIV><FONT face=3DArial size=3D2>It seems to me that I mostly agree(?!) = with what=20 you wrote in this mail. I'd still want to confirm your views on = the=20 following questions:</FONT></DIV> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <DIV><FONT face=3DArial size=3D2> <DIV><FONT face=3DArial size=3D2>a) Is it possible (should we allow) to = stitch =20 PSC1 e2e LSP by a LSC LSP segment ? <DIV><FONT face=3DArial size=3D2> <DIV><FONT face=3DArial size=3D2>b) Is it possible (should we allow) to=20 stitch VC12 e2e LSP by a VC4 LSP segment ? <DIV><FONT face=3DArial size=3D2>c) Is it possible (should we allow) to=20 stitch PSC1 e2e LSP by a PSC2 LSP segment ?</FONT></DIV> <DIV> </DIV> <DIV>My answers are NO to all three questions. You can sticth PSC1 with = PSC1,=20 VC12 with VC12, etc. In all other cases you need, first, to setup a = hierarchical=20 LSP (H-LSP) (whether it is FA-LSP (single domain) or = not (multi-domain)) in=20 order to create a data link of the e2e LSP switching capability and only = after=20 that you can also use a stitching segment on top of this=20 H-LSP).</DIV> <DIV> </DIV> <DIV>See also in-line.</DIV> <DIV> </DIV> <DIV>Igor</DIV></FONT></DIV></FONT></FONT></FONT><FONT face=3DArial=20 size=3D2></FONT></DIV></DIV></DIV> <BLOCKQUOTE dir=3Dltr=20 style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; = BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px"> <DIV style=3D"FONT: 10pt arial">----- Original Message ----- </DIV> <DIV=20 style=3D"BACKGROUND: #e4e4e4; FONT: 10pt arial; font-color: = black"><B>From:</B>=20 <A title=3DDimitri.Papadimitriou@alcatel.be=20 = href=3D"mailto:Dimitri.Papadimitriou@alcatel.be">Dimitri.Papadimitriou@al= catel.be</A>=20 </DIV> <DIV style=3D"FONT: 10pt arial"><B>To:</B> <A = title=3Darthi@juniper.net=20 href=3D"mailto:arthi@juniper.net">Arthi Ayyangar</A> </DIV> <DIV style=3D"FONT: 10pt arial"><B>Cc:</B> <A = title=3Dibryskin@movaz.com=20 href=3D"mailto:ibryskin@movaz.com">Igor Bryskin</A> ; <A=20 title=3Dccamp@ops.ietf.org=20 href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A> </DIV> <DIV style=3D"FONT: 10pt arial"><B>Sent:</B> Wednesday, July 20, 2005 = 4:46=20 AM</DIV> <DIV style=3D"FONT: 10pt arial"><B>Subject:</B> Re: LSP stitching = questions for=20 discussion</DIV> <DIV><BR></DIV> <P><FONT face=3DMonospace,Courier>arthi, </FONT></P> <P> </P> <P><FONT face=3DMonospace,Courier>first a recap: this item on = stitching was=20 initiated in order to allow per "domain" control of end-to-end = (contiguous)=20 LSP, as part of a separate document, and not to extend its scope = (unduly, in=20 this case, see below) but mainly in order to have the set of all = related=20 protocol mechanisms within a specific document -</FONT></P> <P> </P> <P><FONT face=3DMonospace,Courier>from this initial working = assumption, it is=20 clear that the segment must have the properties of the incoming = end-to-end LSP=20 which are determined among other by its switching type e.g. PSC-3 or = TDM, the=20 first requirement of LSP stitching is contiguity of the end-to-end LSP = at the=20 data plane level (after stitching operation) - and this by definition = of the=20 initial working assumption -</FONT></P> <P><FONT face=3D"Courier New">IB>> Agree</FONT></P> <P><FONT face=3DMonospace,Courier>note: the only latitude which is = left (for a=20 given switching type e.g. PSC-2) is due to the properties of packet = (PSC)=20 networks that allows you stitch a segment having a larger Max = Reservable=20 Bandwidth and Unreserved Bandwidth (we would have equivalent latitude = for=20 other attributes but this is a local policy decision - which at the = end is the=20 sole purpose of this approach)</FONT></P> <P><FONT face=3D"Courier New">IB>> Disagree. Conceptually PCS1 = and PSC2=20 are different layers and the latter cannot stitch the former, just = like in=20 case VC12 and VC4. You should nest PSC1 within PSC2 = instead.</FONT></P> <P><FONT face=3DMonospace,Courier><BR><BR>Here are some of the = questions related=20 to LSP stitching to start a<BR>discussion on the usage of LSP=20 stitching.<BR><BR>>1) Are the control plane signaling procedures = for LSP=20 stitching as<BR>>described in this ID, (explicit request for LSP = stitching=20 on<BR>>LSP segment, different label allocation rules; etc) REQUIRED = to=20 stitch<BR>>packet LSPs in data plane ?<BR></FONT></P> <P><FONT face=3DMonospace,Courier>-> yes (beside manual config=20 operation)</FONT></P> <P><FONT face=3D"Courier New">IB>> Agree</FONT></P> <P><FONT face=3DMonospace,Courier><BR>>2) Are the control plane = signaling=20 procedures for LSP stitching as<BR>>described in this ID, (explicit = request=20 for LSP stitching on<BR>>LSP segment, different label allocation = rules;=20 etc) REQUIRED to stitch<BR>>non-packet LSPs in data plane = ?<BR></FONT></P> <P><FONT face=3DMonospace,Courier>-> yes (beside manual config=20 operation)</FONT></P> <P><FONT face=3D"Courier New">IB>> Agree</FONT></P> <P><FONT face=3DMonospace,Courier><BR>>3) If an e2e LSP crosses a = region=20 boundary (based on different TE<BR>>link switching capabilities), = then can=20 I use LSP stitching -<BR>>a) control plane procedures for packet=20 LSPs<BR>>b) data plane procedures for packet LSPs<BR>>c) control = plane=20 procedures for non-packet LSPs<BR>>d) data plane procedures for = non-packet=20 LSPs<BR></FONT></P> <P><FONT face=3DMonospace,Courier>-> no, when crossing a region = boundary, you=20 simply use a document named "draft-mpls-lsp-hierarchy-08.txt" which is = in the=20 RFC Editor queue, but this does not prevent from creating an LSP = segment on=20 top of an FA-LSP (see below)</FONT></P> <P><FONT face=3D"Courier New">IB>> Agree with the caviat that it = should be=20 H-LSP instead of FA-LSP to cover single and multi-domain = scenarios</FONT></P> <P><FONT face=3DMonospace,Courier><BR>>4) If the switching type of = an e2e LSP=20 is different from that of an LSP<BR>>segment, can I use LSP = stitching=20 -<BR>>a) control plane procedures for packet LSPs<BR>>b) data = plane=20 procedures for packet LSPs<BR>>c) control plane procedures for = non-packet=20 LSPs<BR>>d) data plane procedures for non-packet = LSPs<BR></FONT></P> <P><FONT face=3DMonospace,Courier>-> no, in this case progression = of the=20 end-to-end LSP establishment, implies that the LSP segment must fulfil = a set=20 of constraints carried as part of the incoming end-to-end LSP request, = in this=20 case one has to create an LSP segment (with a switching type equal to = the=20 incoming end-to-end LSP switching type) over that FA (i.e. referred = here above=20 to the LSP segment with a different Switching Type)</FONT></P><FONT=20 face=3DMonospace,Courier> <P><FONT face=3D"Courier New">IB>> Agree with the caviat that it = should be=20 H-LSP instead of FA-LSP to cover single and multi-domain = scenarios</FONT></P> <P><FONT face=3DArial size=3D2>Cheers,</FONT></P> <P><FONT face=3DArial size=3D2>Igor</FONT></P> <P>>Note that although 3) and 4) appear to be the same, the reason = I=20 have<BR>>them as different questions is that in one case the TE = link=20 switching<BR>>capability is examined and in the other case the = Generalized=20 Label Request<BR>>is examined.<BR><BR></P></FONT> <P><FONT face=3DMonospace,Courier>-> what it appears to me is that = this=20 discussion mixes the issue of the operation describing the stitching=20 operations at the domain boundaries from the LSP segment creation = itself i.e.=20 an LSP segment can indeed be an FA link inherited from the creation of = an=20 (intra-domain) FA-LSP - </FONT></P></BLOCKQUOTE></BODY></HTML> ------=_NextPart_000_00BB_01C58D2E.79B452C0-- Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 15:40:08 +0000 Sensitivity: Subject: RE: Moving LSP ownership between control plane and management plane To: Dimitri.Papadimitriou@alcatel.be Cc: ccamp@ops.ietf.org Message-ID: <OF38D13803.20690BAC-ONC1257044.00518238-C1257044.00560DC2@uk.marconicomms.com> From: "Dino Bramanti" <Dino.Bramanti@marconi.com> Date: Wed, 20 Jul 2005 17:39:41 +0200 MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii Dimitri, as Diego says in his post, the main use for the procedure defined in our ID is for MP->CP handover. This at least could be its immediate possible use and surely it is the reason why we started thinking about a similar mechanism within CP. But we have proposed the complete procedure including the reverse CP->MP handover for several reasons. The first one is related to the need of reverting back what MP-CP procedure does in case of some kind of failure during its action. In addition, the possibility of doing a CP->MP handover of LSP, leaving untouched corresponding data plane state, could be useful for specific needs. An example could be a Control Plane software upgrade to be performed on network elements that are in service and carrying traffic over data plane. In such a case it could be possible to perform a CP->MP handover of a number of LSP, transfering their ownership to MP for the time required. Furthermore, taking Stephen's suggestion about it and looking at the proposed mechanism (i.e. handover flag usage within GRSVPTE signaling flow) as a general way to act on CP state without affecting data plane state, the need for a two-way procedure is required to provide it with maximum flexibility. Thank you dino >diego >as part of the motivation section of this document it would be indeed interesting to be much more explicit about the CP->MP >handover purpose(s) before progressing on this part of the mechanism >- d. Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 15:10:30 +0000 Content-Class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: End-to-end L2-LSP Date: Wed, 20 Jul 2005 16:09:56 +0100 Message-ID: <B5E87B043D4C514389141E2661D255EC0A835AD2@i2km41-ukdy.domain1.systemhost.net> Thread-Topic: End-to-end L2-LSP Thread-Index: AcWII4lsKzFN0JPAQs66qP37pnL8MAFB1CaQ From: <richard.spencer@bt.com> To: <jaihyung@etri.re.kr>, <ccamp@ops.ietf.org> Jaihyung, I agree that extending signalling down to the end-user is one possible = solution to address charging/billing issues, another possible solution = would be to use session based billing. I also agree that using L2-LSPs = could be a viable solution for providing traffic separation and = bandwidth guarantees in a service providers network, another possible = solution would be to extended MPLS into the access network. However, if you want to extend L2-LSPs (not just signalling) down to the = end user, then there are a number of complicated technical and = commercial issues that must be overcome. Please see below for some of = the key issues: 1. If you want to eliminate the possibility of congestion in the = enterprise/home network using L2-LSPs for the reasons you give below, = then on every node in the network you will need: - A CPU: for control plane protocol processing - Routing: At least static, an IGP for large networks and OSPF-TE if TE = is required - Signalling: RSVP-TE - Policing: Per connection policing, e.g. per VLAN I do not know how much it would cost to upgrade the $20 five port 10/100 = switch sitting on my desk to support all this, but I do not believe this = would be a trivial amount, particularly at the volumes demanded by = enterprises. In addition, a hierarchical architecture would be required = for the solution to scale and therefore aggregation switches within the = enterprise network would also need to support hierarchical CAC and = policing. 2. Unless you want to reinvent LANE, then you will need to continue to = support CL Ethernet broadcasts and multicasts on GMPLS enabled switches. = To protect the L2-LSP traffic from CL broadcast traffic you could rate = limit it to say 1% of the bandwidth available. However, how would you = capacity plan for and rate limit (dynamic, receiver driven) multicast = traffic for which usage patterns are difficult to predict? In my opinion = you will need to use over provisioning along with CoS, which would raise = the question, why is over provisioning and CoS OK for multicast traffic = (which could be critical, e.g. trading floor information), but not OK = for unicast traffic? 3. Routers are relatively stable systems that are switched on = permanently and therefore RSVP-TE sessions are stable and only go down = when there is a failure. End user systems however such as PCs, set top = boxes, games consoles, etc. are not always switched on and are not as = stable. What will the effect on the network be when large number of = RSVP-TE sessions are going up/down, e.g. as a result of game console = resets or PC crashes/freezes? These effects will ripple through the = enterprise/home network into the operators network, and therefore = mechanisms must be put in place to protect against them. 4. When an IP device wants to forward a packet it uses ARP to discover = the MAC address of the IP next hop (unless its already in the MAC table) = and adds the Ethernet header before forwarding the packet on. Now, if = you want to extend connections down to the end user device, then the = device will need to map IP traffic to different L2-LSPs. A simple 1-to-1 = mapping between an IP next hop address and a MAC address will not be = sufficient as there may be multiple connections (e.g. using different = VLAN IDs) associated with an IP next hop. So, policies will need to be = configured on end user devices to map IP traffic (e.g. based on src/dest = IP address, port numbers, etc.) to a specific L2-LSP (e.g. based on = src/dest MAC address and/or VLAN ID). =20 5. End user devices such as PCs and VoIP phones are likely to only need = to support a relatively low number of L2-LSPs. However, devices such as = content servers and call servers/gateways will need to support = hundreds/thousands of L2-LSPs. Connection setup and state maintenance = for large numbers of L2-LSPs will consume a significant amount of = processing resources, as will policies to map IP traffic to the = different L2-LSPs. Therefore, I would expect that the system software = (and possibly hardware) will need to be designed from the ground up to = be able to support large numbers of L2-LSPs. 6. Enterprise networks are built using not just switches but routers as = well. Firstly, for L2-LSPs to be set up between switches and routers, = they must belong to the same address space to communicate with each = other. However, this address space must be separate from the IPv4 = forwarding address space. To use an example of why this is true, routers = connected via Ethernet segments are considered to be directly connected, = however, if the IPv4 and L2-LSP address spaces were the same, then IP = routers separated by GMPLS switches would be multiple hops away from = each other. In addition to address space separation, routing protocol = instances must also be separated. This need for this separation is = obvious when considering using a link state routing protocol which = relies on a topological view of the network, i.e. the IPv4 and L2-LSP = topologies are different.=20 Secondly, routers break up Ethernet segments and must therefore must = always be the headend or tailend for a L2-LSPs, i.e. they cannot be = midpoints. This means that from a control plane perspective, on = receiving a L2-LSP path message a router must terminate the LSP and send = back a resv message whether it is the intended destination/sink or not. = It must then originate a path message for a new L2-LSP and send it onto = the next hop GMPLS switch. This is not how RSVP-TE works today and = therefore will require modifications. In addition, traffic will be = blackholed if transmitted by the source before all the separate L2-LSPs = have been setup end-to-end to the sink/destination. This means that some = kind of end-to-end signalling message will need to be sent by the = sink/destination once the final L2-LSP has been established to inform = the source that the end-to-end connectivity has been provided. This = functionality is not supported by RSVP-TE today and will therefore = require protocol modifications. Thirdly, routers forward packets based on information in the IP header. = So, when a router receives a packet via a L2-LSP it will strip of the L2 = header and perform an IP lookup. Now, for the packet to be forwarded = using the correct L2-LSP the router must have a policy that tells it = which L2-LSP to use as there may be multiple L2-LSPs to the same = destination, e.g. for different voice calls or video downloads. However, = before a policy can even be configured, the router needs to know what = traffic (e.g. based on src/dest address, port number, DSCP, etc.) to map = to which L2-LSP, and this information needs to come from the traffic = source using signalling, i.e. RSVP-TE. I do not believe this is = supported today and will therefore require modifications. Regards, Richard > -----Original Message----- > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > Sent: 14 July 2005 04:24 > To: Spencer,R,Richard,XDE73 R; ccamp@ops.ietf.org > Subject: RE: End-to-end L2-LSP >=20 > Dear Richard, >=20 > See in-line please, >=20 > >Thank you for taking the time to respond to my questions. I=20 > anticipated=20 > >that your response would include "over-provisioning is the=20 > best answer"=20 > >to decrease the chance of congestion on non-GMPLS switches.=20 > >This is what is done in the LAN environment today in=20 > conjunction with CoS,=20 > >and this is why I believe there will be two opposing views on how=20 > >far L2-LSP connections should be extended down to the user. One view=20 > >is that over provisioning and CoS work fine in the=20 > enterprise network today,=20 > >and therefore the L2-LSP should only be extended down to the point=20 > >at which the enterprise network connects to the SP network,=20 > >e.g. Ethernet DSLAM. I think the other view will be that=20 > over provisioning=20 > >and CoS in the enterprise network don't work and therefore L2-LSPs=20 > >should be provided end-to-end using only GMPLS capable switches.=20 > >I wouldn't have thought there would be much call for mixed=20 > >GMPLS/non-GMPLS enterprise networks (except for during migration)=20 > >due to the opposing views, e.g. if over provisioning is good=20 > enough for=20 > >one switch, why isn't it good enough for the other? >=20 > Over-provisioned network is non-guaranteed network. It can't prevent=20 > occasional congestion at any rate. Further, network condition=20 > in private=20 > network is very diverse. You can't predict minimum service level=20 > inside every private network.=20 >=20 > Suppose that an ISP want to provide high-quality, video-phone service=20 > for enterprise users. Nothing can be guaranteed inside the=20 > over-provisioned=20 > customer network. If the service level is not satisfactory,=20 > customers will complain=20 > to service provider even if the main cause is in his network=20 > or destination network.=20 > We need some structure that coordinate service quality in=20 > both private sector,=20 > as well as in public sector consistently.=20 > An end-to-end signal is an effective means to communicate between=20 > user terminal, private switches and provider switches, and=20 > concert service=20 > quality at each different operations network. This is a goal=20 > that RSVP (RFC2205)=20 > pursued, however ineffective in local network because=20 > Ethernet industry=20 > didn't support RSVP over Ethernet (RFC2814, 2815, 2816).=20 > Now we have another opportunity to realize the goal using L2-LSP.=20 > Therefore, I believe end-to-end LSP is necessary even in=20 > enterprise network=20 > for the purpose of providing guaranteed commercial service. >=20 > >I think where most people will agree is that using connections=20 > >in the SPs access network does make sense in order to=20 > provide efficient=20 > >link utilisation and per connection policing to ensure one customers=20 > >traffic can not affect another's. Regarding the ability to=20 > set up new=20 > >connections based on a request from the user (rather than=20 > constraining=20 > >RSVP signalling to inside the SP network), I think this will come=20 > >down to enterprise customer demand for improved service performance.=20 > >If a customer wasn't getting the performance they needed from their=20 > >SP and was offered a new service that allowed them to setup=20 > new L2-LSP=20 > >connections on demand that would offer better performance=20 > (e.g. for voice=20 > >calls or video conferences), then upgrading the user devices=20 > to support=20 > >the new service could well be an attractive proposition. If=20 > on the other=20 >=20 > Yes, this will be one of prospective service newly enabled=20 > using L2-LSP. >=20 > >hand a customer buys a best effort service and considers the=20 > performance=20 > >to be 'good enough', then there would be no reason to go to=20 > the expense=20 > >of upgrading the user devices to support the set up of L2-LSPs. >=20 > Ever increasing besteffort bandwidth will eventually not satisfy=20 > both providers and customers for many reasons. Service providers=20 > will have difficulty in collecting expenses enough to invest=20 > on increased=20 > bandwidth. This will in turn gradually downgrade overall=20 > communication quality,=20 > and applications requiring real-time service will never=20 > satisfy customers.=20 > We already see such symptom in some low-price service providers.=20 > I believe new service structure that can provide difference=20 > of service=20 > quality will offer benefit to customers, providers as well as=20 > vendors in the end. >=20 > I also look forward to see you at Paris and continue helpful=20 > discussion. >=20 > Sincerely >=20 > Jaihyung >=20 >=20 > =20 >=20 > Dr. Jaihyung Cho > ETRI, Korea > phone : 042) 860-5514 > oversea: +82-42-860-5514 > fax: +82-42-861-5550=20 >=20 >=20 >=20 >=20 > -----?? ???-----=20 > ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com>=20 > ?? ??: 2005-07-14 ?? 1:10:30=20 > ?? ??: "jaihyung@etri.re.kr" <jaihyung@etri.re.kr>,=20 > "ccamp@ops.ietf.org" <ccamp@ops.ietf.org>=20 > ??:=20 > ??: RE: End-to-end L2-LSP=20 >=20 >=20 >=20 > Jaihyung,=20 >=20 > Thank you for taking the time to respond to my questions. I=20 > anticipated that your response would include=20 > "over-provisioning is the best answer" to decrease the chance=20 > of congestion on non-GMPLS switches. This is what is done in=20 > the LAN environment today in conjunction with CoS, and this=20 > is why I believe there will be two opposing views on how far=20 > L2-LSP connections should be extended down to the user. One=20 > view is that over provisioning and CoS work fine in the=20 > enterprise network today, and therefore the L2-LSP should=20 > only be extended down to the point at which the enterprise=20 > network connects to the SP network, e.g. Ethernet DSLAM. I=20 > think the other view will be that over provisioning and CoS=20 > in the enterprise network don't work and therefore L2-LSPs=20 > should be provided end-to-end using only GMPLS capable=20 > switches. I wouldn't have thought there would be much call=20 > for mixed GMPLS/non-GMPLS enterprise networks (except for=20 > during migration) due to the opposing views, e.g. if over=20 > provisioning is good enough for one switch, why isn't it good=20 > enough for the other?=20 >=20 > I think where most people will agree is that using=20 > connections in the SPs access network does make sense in=20 > order to provide efficient link utilisation and per=20 > connection policing to ensure one customers traffic can not=20 > affect another's. Regarding the ability to set up new=20 > connections based on a request from the user (rather than=20 > constraining RSVP signalling to inside the SP network), I=20 > think this will come down to enterprise customer demand for=20 > improved service performance. If a customer wasn't getting=20 > the performance they needed from their SP and was offered a=20 > new service that allowed them to setup new L2-LSP connections=20 > on demand that would offer better performance (e.g. for voice=20 > calls or video conferences), then upgrading the user devices=20 > to support the new service could well be an attractive=20 > proposition. If on the other hand a customer buys a best=20 > effort service and considers the performance to be 'good=20 > enough', then there would be no reason to go to the expense=20 > of upgrading the user devices to support the set up of L2-LSPs.=20 >=20 > I look forward to the discussions in Paris.=20 >=20 > Regards,=20 > Richard=20 >=20 > > -----Original Message-----=20 > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr]=20 > > Sent: 13 July 2005 15:14=20 > > To: Spencer,R,Richard,XDE73 R; ccamp@ops.ietf.org=20 > > Subject: RE: End-to-end L2-LSP=20 > >=20 > >=20 > > =20 > > =20 > > Dear Richard Spencer=20 > > =20 > > see in-line=20 > > =20 > > >Jaihyung,=20 > > >=20 > > >Perhaps I'm becoming too much of a pessimist;-)=20 > > =20 > > oh, not at all,=20 > > you are reasonable enough to dig up unclear knowledge=20 > > please go ahead until you finally get satisfactory answer.=20 > > =20 > >=20 > > >Ignoring the commercial aspects, please can you help clarify=20 > > a few technical=20 > > >questions I have in order to help me gain a better=20 > > understanding of the proposal?=20 > > >=20 > > >1. You say that one of the key reasons for using GMPLS in=20 > > the enterprise=20 > > >network is to provide end-to-end QoS control. However, you=20 > > also say that not=20 > > >all the switches will need to support GMPLS, only those at=20 > > bottle necks.=20 > > >Please can you expand on this to explain how you can have=20 > end-to-end=20 > > >QoS control when the L2-LSP is not end-to-end?=20 > > =20 > >=20 > > The contextual meaning of 'End-to-end LSP' is that GMPLS label is=20 > > assigned from source terminal and removed at destination terminal,=20 > > because RSVP-TE message is exchanged from host to host.=20 > > It may not perfectly guarantee QoS when the GMPLS deployment=20 > > is not complete, however it will at least be good enough=20 > > for end-user applications to tell those GMPLS implemented=20 > L2 switches=20 > > what frame must be given priority service when congestion occur.=20 > > =20 > >=20 > > >As being discussed=20 > > >on the PWE3 mailing list, the performance a service receives=20 > > is only as=20 > > >good as the worst performing server layer. Therefore, if=20 > there is a=20 > > >standard 802.1 workgroup switch between the user and the=20 > > L2-LSP that is=20 > > >experiencing congestion (e.g. due to a large amount of=20 > > peer-to-peer traffic),=20 > > >then the QoS that the user receives will only be as good as=20 > > that provided=20 > > >by the congested workgroup switch.=20 > > =20 > >=20 > > of course, ETH-GMPLS does not prevent congestion inevitably=20 > occurring=20 > > in those non-GMPLS deployed part.=20 > > =20 > >=20 > > >The only way I can see a network using a mixture of GMPLS=20 > > and non-GMPLS=20 > > >switches guaranteeing end-to-end QoS would be to design the=20 > > network in=20 > > >such a way that the non-GMPLS switches could never become=20 > congested.=20 > > =20 > >=20 > > Yes, that will be the way wise network designer improve=20 > their network.=20 > > =20 > >=20 > > >To do this would require that RSVP requests be used for ALL=20 > > traffic flows=20 > > >to ensure that there was enough bandwidth available, along=20 > > with strict=20 > > >per L2-LSP policing on the user devices (e.g. PC).=20 > > >However, the obvious flaw in this approach is that someone could=20 > > >just come along and connect=20 > > >an end device to a non-GMPLS switch and start transmitting traffic=20 > > >(without performing CAC/policing) causing the switch to=20 > > become congested.=20 > > >Therefore, please can you explain how end-to-end QoS can be=20 > > guaranteed=20 > > >in a mixed non-GMPLS and GMPLS switching environment?=20 > > =20 > >=20 > > Of course traffic came from non controlled ingress can not be=20 > > prevented=20 > > in non-GMPLS node. There's no magic. Only the=20 > > over-provisioning is the=20 > > best answer. However, once the aggregated traffic pass 'ANY' one of=20 > > GMPLS enabled node, those uncontrolled traffic will suffer=20 > > queuing discipline.=20 > > =20 > >=20 > > >2. Section 5.1 currently states "When the customer initiates=20 > > data transmission,=20 > > >the access switch maps the user flow into the L2 LSP.=20 > > Mapping procedure is=20 > > >subject to implementation choices, and is out of the scope of this=20 > > >document." In order to be able to map any traffic, first of=20 > > all the switch needs=20 > > >to know what type of traffic (e.g. based on source/dest MAC,=20 > > VLAN ID, 802.1p=20 > > >bits, etc) should be mapped to the L2-LSP. In Figure 1, if=20 > > an L2-LSP is set up=20 > > >from the DSLAM to the edge router based on a request from=20 > > the user (e.g. via=20 > > >a PC or VoIP phone), then won't the user also need to signal=20 > > what type of=20 > > >traffic should to be mapped to the L2-LSP?=20 > > =20 > >=20 > > Thank you for pointing out important thing.=20 > > That's the most unfortunate part I reluctantly agreed because the=20 > > work condition in the framework DT was not to span L2-LSP=20 > > beyond provider network.=20 > > As you noticed, mapping user traffic at provider access is awkward.=20 > > A smooth method will be for user terminal aware of the L2 label=20 > > to use by receiving RSVP-TE message.=20 > > However, there was strong argument that RSVP-TE must not=20 > > go beyond provider network because the DT is not chartered=20 > > to work in customer network.=20 > > That's why I requested amendment of CCAMP charter=20 > > to include user terminal explicitly in CCAMP work scope, so that=20 > > avoid such silly argument in future.=20 > > =20 > >=20 > > >3. In section 5.1, the text states that the aggregation=20 > > switch between the=20 > > >DSLAM and edge router may be a 802.1 switch, i.e. it doesn't=20 > > have to support=20 > > >GMPLS. To set up a connection, first of all there needs to=20 > > be a data plane=20 > > >capable of carrying control plane traffic between GMPLS=20 > > peers. Secondly,=20 > > >the GMPLS peers need reachability information in order to=20 > > forward control=20 > > >messages onto the next peer (i.e. a static or IGP route). If=20 > > the GMPLS peers=20 > > >(e.g. the DSLAM and edge router in Figure 1) are not=20 > > directly connected,=20 > > >then the GMPLS control packets will need to be forwarded by=20 > > the 802.1 switch.=20 > > >Would the idea be to have say a "control traffic VLAN"=20 > > configured between the=20 > > >DSLAM and edge router via the transit 802.1 switch so that=20 > > control packets=20 > > >could be forwarded transparently by the 802.1 switch?=20 > > =20 > >=20 > > Although it doesn't preclude use of VLAN, I think such=20 > > special configuration=20 > > for RSVP-TE control message is not necessary in legacy switches.=20 > > Legacy switches only need to pass RSVP-TE messages as they=20 > > pass normal data frames. Only the GMPLS implemented switches=20 > > need to distinguish frames containing RSVP-TE messages and=20 > > snoop them.=20 > > Detail of implementation is, of course, for future discussion.=20 > > =20 > >=20 > > >4. If we assume that control messages can be transparently=20 > forwarded=20 > > >by the intermediate 802.1 aggregation switch then how are=20 > > user frames=20 > > >forwarded in the data plane by the 802.1 switch?=20 > > =20 > >=20 > > Legacy switches do not distinguish control/data frames.=20 > > =20 > >=20 > > >If the DSLAM sets up a L2-LSP=20 > > >to the edge router using say VLAN ID 99, how does the switch=20 > > know that any=20 > > >packets coming in with a VLAN ID of 99 should be forwarded=20 > > out of the port=20 > > >connecting to the edge router?=20 > > =20 > >=20 > > oh, you misunderstood the L2-LSP.=20 > > The question you are asking depends on how L2 label is encoded.=20 > > there was hot debate in the DT about the issue how L2 label=20 > > is encoded.=20 > > My proposal is to use part of MAC address for label and the=20 > > other people=20 > > argued to use VLAN ID for label. The debate was not concluded=20 > > because the goal of the L2SC framework document is not to select=20 > > implementation method but to draw necessary work scope.=20 > > I believe this issue will be dealt in depth after Paris meeting.=20 > > So, please reserve your question until then and ask to those=20 > > VLAN proponents.=20 > > =20 > >=20 > > >5. During connection setup, I take it that the GMPLS peers=20 > > assume that=20 > > >resources are available within the intermediate 802.1 switch=20 > > to support=20 > > >the connection (as its transparent). However, as there is no=20 > > >CAC/policing being performed by the 802.1 switch=20 > > >(the L2-LSPs are transparent), and due to the any-to-any and=20 > > broadcast=20 > > >based nature of Ethernet (i.e. its CL properties), then=20 > how can this=20 > > >assumption be made? This comes back to my original question, how=20 > > >can end-to-end QoS be guaranteed in an environment where there is=20 > > >a mixture of 802.1 and GMPLS enabled switches?=20 > > =20 > >=20 > > yes, you came back to the question I answered in above ;-)=20 > > so, please see above.=20 > > GMPLS enabled nodes must be positioned at points congestion=20 > > is likely to be expected.=20 > > =20 > > Thanks=20 > > =20 > > Jaihyung=20 > > =20 > >=20 > > = =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=20 > >=20 > > =20 > >=20 > > -----?? ???-----=20 > > ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com>=20 > > ?? ??: 2005-07-13 ?? 8:53:13=20 > > ?? ??: "jaihyung@etri.re.kr" <jaihyung@etri.re.kr>,=20 > > "ccamp@ops.ietf.org" <ccamp@ops.ietf.org>=20 > > ??:=20 > > ??: RE: End-to-end L2-LSP=20 > >=20 > >=20 > >=20 > >=20 > > Jaihyung,=20 > >=20 > > Perhaps I'm becoming too much of a pessimist;-) Ignoring the=20 > > commercial aspects, please can you help clarify a few=20 > > technical questions I have in order to help me gain a better=20 > > understanding of the proposal?=20 > >=20 > > 1. You say that one of the key reasons for using GMPLS in the=20 > > enterprise network is to provide end-to-end QoS control.=20 > > However, you also say that not all the switches will need to=20 > > support GMPLS, only those at bottle necks. Please can you=20 > > expand on this to explain how you can have end-to-end QoS=20 > > control when the L2-LSP is not end-to-end? As being discussed=20 > > on the PWE3 mailing list, the performance a service receives=20 > > is only as good as the worst performing server layer.=20 > > Therefore, if there is a standard 802.1 workgroup switch=20 > > between the user and the L2-LSP that is experiencing=20 > > congestion (e.g. due to a large amount of peer-to-peer=20 > > traffic), then the QoS that the user receives will only be as=20 > > good as that provided by the congested workgroup switch.=20 > >=20 > > The only way I can see a network using a mixture of GMPLS and=20 > > non-GMPLS switches guaranteeing end-to-end QoS would be to=20 > > design the network in such a way that the non-GMPLS switches=20 > > could never become congested. To do this would require that=20 > > RSVP requests be used for ALL traffic flows to ensure that=20 > > there was enough bandwidth available, along with strict per=20 > > L2-LSP policing on the user devices (e.g. PC). However, the=20 > > obvious flaw in this approach is that someone could just come=20 > > along and connect an end device to a non-GMPLS switch and=20 > > start transmitting traffic (without performing CAC/policing)=20 > > causing the switch to become congested. Therefore, please can=20 > > you explain how end-to-end QoS can be guaranteed in a mixed=20 > > non-GMPLS and GMPLS switching environment?=20 > >=20 > > 2. Section 5.1 currently states "When the customer initiates=20 > > data transmission, the access switch maps the user flow into=20 > > the L2 LSP. Mapping procedure is subject to implementation=20 > > choices, and is out of the scope of this document." In order=20 > > to be able to map any traffic, first of all the switch needs=20 > > to know what type of traffic (e.g. based on source/dest MAC,=20 > > VLAN ID, 802.1p bits, etc) should be mapped to the L2-LSP. In=20 > > Figure 1, if an L2-LSP is set up from the DSLAM to the edge=20 > > router based on a request from the user (e.g. via a PC or=20 > > VoIP phone), then won't the user also need to signal what=20 > > type of traffic should to be mapped to the L2-LSP?=20 > >=20 > > 3. In section 5.1, the text states that the aggregation=20 > > switch between the DSLAM and edge router may be a 802.1=20 > > switch, i.e. it doesn't have to support GMPLS. To set up a=20 > > connection, first of all there needs to be a data plane=20 > > capable of carrying control plane traffic between GMPLS=20 > > peers. Secondly, the GMPLS peers need reachability=20 > > information in order to forward control messages onto the=20 > > next peer (i.e. a static or IGP route). If the GMPLS peers=20 > > (e.g. the DSLAM and edge router in Figure 1) are not directly=20 > > connected, then the GMPLS control packets will need to be=20 > > forwarded by the 802.1 switch. Would the idea be to have say=20 > > a "control traffic VLAN" configured between the DSLAM and=20 > > edge router via the transit 802.1 switch so that control=20 > > packets could be forwarded transparently by the 802.1 switch?=20 > >=20 > > 4. If we assume that control messages can be transparently=20 > > forwarded by the intermediate 802.1 aggregation switch then=20 > > how are user frames forwarded in the data plane by the 802.1=20 > > switch? If the DSLAM sets up a L2-LSP to the edge router=20 > > using say VLAN ID 99, how does the switch know that any=20 > > packets coming in with a VLAN ID of 99 should be forwarded=20 > > out of the port connecting to the edge router?=20 > >=20 > > 5. During connection setup, I take it that the GMPLS peers=20 > > assume that resources are available within the intermediate=20 > > 802.1 switch to support the connection (as its transparent).=20 > > However, as there is no CAC/policing being performed by the=20 > > 802.1 switch (the L2-LSPs are transparent), and due to the=20 > > any-to-any and broadcast based nature of Ethernet (i.e. its=20 > > CL properties), then how can this assumption be made? This=20 > > comes back to my original question, how can end-to-end QoS be=20 > > guaranteed in an environment where there is a mixture of=20 > > 802.1 and GMPLS enabled switches?=20 > >=20 > > Thanks,=20 > > Richard=20 > >=20 > > > -----Original Message-----=20 > > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr]=20 > > > Sent: 12 July 2005 09:21=20 > > > To: Spencer,R,Richard,XDE73 R; ccamp@ops.ietf.org=20 > > > Subject: RE: End-to-end L2-LSP=20 > > >=20 > > >=20 > > >=20 > > >=20 > > > Dear Richard,=20 > > >=20 > > > Yes, I believe it is not only commercially profitable but it=20 > > > may also grow total=20 > > > revenue of service operators by introducing new services.=20 > > >=20 > > > For example, think the example of IP phone you referred.=20 > > > How do you think you can provide reliable VoIP service and=20 > > > safely authenticate phone users and charge fees as counted by=20 > > > packet usage=20 > > > using simple Ethernet based xDSL network ?=20 > > >=20 > > > There's hardly new service that providers may offer using=20 > > > Ethernet based=20 > > > architecture, and collect sufficient profit to investment.=20 > > >=20 > > > A sophisticated ATM switch might do the work if it is not=20 > > expensive.=20 > > > However, my intension is modifiying existing design of=20 > > > Ethernet hardware=20 > > > with a little bit additional cost, and provide the level of=20 > > > service that=20 > > > ATM switches might offer.=20 > > >=20 > > > We have initial design of the Ethernet-MPLS hardware. I=20 > don't think=20 > > > GMPLS enabled Ethernet switch will be so complex. The=20 > market price=20 > > > of GMPLS enabled switch will not be so expensive than normal=20 > > > Ethernet switches.=20 > > >=20 > > > Further, it only requires upgrade of switches in some=20 > > > bottleneck points of network.=20 > > > Read the framework documents of L2SC. It doesn't require=20 > > > entire replacement=20 > > > of switches to introduce new service.=20 > > >=20 > > > Every year, ISPs purchase new access switches and replaces=20 > > > old machines.=20 > > > With a bit additional cost, ISPs may use the GMPLS enabled=20 > > switch for=20 > > > normal Ethernet switching in initial deployment.=20 > > >=20 > > > When the GMPLS function is enabled in some upgraded part of=20 > > network,=20 > > > ISPs may provide high quality, reliable VoIP service to=20 > > xDSL customers=20 > > > using RSVP-TE enabled VoIP phone.=20 > > > The high quality video-phone service enabled by GMPLS will be=20 > > > competable to legacy voice only phone service.=20 > > >=20 > > >=20 > > > There are lots of things I can argue about commercial aspects=20 > > > of this technique.=20 > > > However I don't think it is not appropriate to discuss this=20 > > > issue in this mailing list.=20 > > > All I can say is,=20 > > > yes, I believe that it may be commercially viable.=20 > > >=20 > > >=20 > > > Thanks=20 > > >=20 > > > Jaihyung=20 > > >=20 > > > = =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=20 > > >=20 > > > Do you really think the benefits of extended GMPLS down to=20 > > > the end user could ever justify the costs involved in=20 > > > upgrading LAN switches and things like IP phones (simple low=20 > > > cost commodity items) to support GMPLS signalling/routing?=20 > > >=20 > > > Your proposal may be technically possible (though it would=20 > > > require new hardware and significant software upgrades), but=20 > > > commercially I would consider this to be a non-starter.=20 > > >=20 > > > Regards,=20 > > > Richard=20 > > >=20 > > > > -----Original Message-----=20 > > > > From: owner-ccamp@ops.ietf.org=20 > > [mailto:owner-ccamp@ops.ietf.org]On=20 > > > > Behalf Of CHO, JAI HYUNG=20 > > > > Sent: 12 July 2005 01:10=20 > > > > To: ccamp@ops.ietf.org=20 > > > > Subject: End-to-end L2-LSP=20 > > > >=20 > > > >=20 > > > > =20 > > > > Dear all,=20 > > > >=20 > > > > Greetings.=20 > > > >=20 > > > > I am a co-editor of the Ethernet-GMPLS framework document.=20 > > > > (draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt)=20 > > > >=20 > > > > I am writing this to explain the motivation behind the=20 > > > > proposal of scenario-1, and to raise the issue of=20 > > > > end-to-end LSP to be included in CCAMP charter.=20 > > > >=20 > > > > An important goal intended in scenario-1 is=20 > > > > QoS control over end-to-end packet delivery using LSP.=20 > > > > In other words, the purpose of scenario-1 is extending coverage=20 > > > > of GMPLS control including customer network and=20 > end-user device,=20 > > > > such as PC, VoIP phone, IP-TV, or even future 4G mobile phone.=20 > > > > End-user applications will have a means to request resources=20 > > > > to network using GMPLS signal.=20 > > > >=20 > > > > Unfortunately, this view is not included in the=20 > > > > framework because there was strong opinion that=20 > > > > L2-LSP must not span beyond provider network because=20 > > > > CCAMP is not chartered to work on customer=20 > > > > network. I am wondering if this is right argument.=20 > > > > Perhaps the charter can be amended to include this work.=20 > > > >=20 > > > > I believe there is little technical reason to preclude=20 > > > > LSP being used in customer network. It is a reasonable=20 > > > > assumption that private company may also use GMPLS=20 > > > > implemented Ethernet switch, once the L2SC work is successful.=20 > > > > Currently, CCAMP is the only place discussing the matter of=20 > > > > Ethernet and LSP. There is no other WG dealing similar issue.=20 > > > > If CCAMP wants to work on Ethernet,=20 > > > > I think the coverage of common GMPLS control must be=20 > > > > extended to include all area where Ethernet is deployed.=20 > > > > Currently, Ethernet is dominantly deployed in first-mile=20 > > > > network where access network as well as customer network=20 > > > > and user hosts are usually included. Therefore, user host must=20 > > > > be considered as initiation and termination point of L2-LSP.=20 > > > >=20 > > > > Another important reason to include user host in CCAMP=20 > > > > work scope is to achieve consistent end-to-end control of QoS.=20 > > > > End-to-end QoS has long been desired goal since RSVP was=20 > > > > first designed. This goal can be best achieved using=20 > > > > end-to-end GMPLS signal and L2-LSP.=20 > > > > Currently, RSVP (RFC2205) is available in many user hosts.=20 > > > > There is only a little difference between RSVP and RSVP-TE.=20 > > > > If the RSVP-TE signal of provider network can reach user host=20 > > > > across Ethernet based private network, an LSP can be=20 > established=20 > > > > from application to application via the provider backbone.=20 > > > > Service providers will be able to control application flows=20 > > > > in L2-LSP level as well as in aggregated LSP pipe of similar=20 > > > > application type. A variety of new services will be=20 > enabled using=20 > > > > the enhanced capability of distinguishing and controlling each=20 > > > > individual application flow.=20 > > > >=20 > > > > I personally believe that end-to-end L2-LSP may offer=20 > > > > innovative solutions for servicing IP-TV as well as fast=20 > > mobility.=20 > > > > There have been some experimental approaches combining=20 > > > > MPLS and IP mobility. However such potential of the new=20 > > > > application can only be explored when end-user host is=20 > > > > included in the scope of CCAMP work area.=20 > > > >=20 > > > > Therefore, I hereby sincerely request people's attention=20 > > > > and support on this perspective of end-to-end LSP.=20 > > > > If there is enough number of people agree on this view,=20 > > > > the WG can request the necessary update of the charter.=20 > > > >=20 > > > > Thank you.=20 > > > > =20 > > > > Jaihyung Cho=20 > > > > =20 > > > > =20 > > > > =20 > > > > ETRI, Korea=20 > > > > phone : 042) 860-5514=20 > > > > oversea: +82-42-860-5514=20 > > > > fax: +82-42-861-5550=20 > > > >=20 > > > >=20 > > > >=20 > > > >=20 > > > >=20 > > > >=20 > > > >=20 > > >=20 > > >=20 > > >=20 > > >=20 > >=20 > >=20 > >=20 > >=20 >=20 >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 14:50:19 +0000 Sensitivity: Subject: RE: Moving LSP ownership between control plane and management plane To: sdshew@nortel.com Cc: "<ccamp" <ccamp@ops.ietf.org> Message-ID: <OF1E9D9B19.0F4303DE-ONC1257044.00505E5C-C1257044.005153D8@uk.marconicomms.com> From: "Diego Caviglia" <Diego.Caviglia@marconi.com> Date: Wed, 20 Jul 2005 16:48:14 +0200 MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii In line Regards Diego "Stephen Shew" <sdshew@nortel.com> on 20/07/2005 14.25.18 To: "Diego Caviglia" <Diego.Caviglia@marconi.com> cc: <ccamp@ops.ietf.org> Subject: RE: Moving LSP ownership between control plane and management plane [CUT] I have no issue with the motivation. I just think there are more scenarios for creating control plane state for an existing cross connection. [dc] Yes I agree, do you have any suggestion on other scenarios? > If you do this then the bit should be renamed something like > "cross-connection exists" and if set, then the signaling > protocol does not tell the cross connection to set/release > it. [dc] That is how it works apart for the name of the flag. I realize that. What I'm suggesting is that the name reflect an action rather than one of several possible purposes. [dc] No problem with this, actually the original name of the flag was Fake flag indicating that creation and deletion are not for real. If there are no objection on this we can change the flag name. [CUT] The suggestion was that there are other scenarios in which establishment of new LSP state for an existing cross connection is useful. In the MP->CP case, the MP initially holds state for the connection and before the action is completed, both MP and CP hold state. In the two LSP example above, two entities in the CP hold state about the same connection. I see these two cases as similar in that the general case is about multiple entities holding state for the same resource. [dc] Yes agreed > > Stephen > > > > -----Original Message----- > > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On > > Behalf Of Diego Caviglia > > Sent: Wednesday, July 20, 2005 05:44 > > To: Diego.Caviglia@marconi.com > > Cc: ""Dimitri.Papadimitriou" <Dimitri.Papadimitriou"; > "ccamp" <ccamp; > > "gnewsome" <gnewsome; "adrian" <adrian > > Subject: RE: Moving LSP ownership between control plane and > management > > plane > > > > > > > > Sorry for this but may be not all have read the ID, I hope > that having > > the text help the discussion. > > > > Regards > > > > Diego > > > > > > > > > > Network Working Group > > Internet Draft > > Diego Caviglia > > Document: draft-caviglia-mp2cpcp2mp-02.txt Marconi > > Proposed Status: Updates RFC 3473 > > Dino Bramanti > > > > Marconi > > Expires: December 2005 > > Nicola Ciulli > > <snip> > > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 14:50:11 +0000 Content-Type: Multipart/Mixed; Boundary="NextPart" Mime-Version: 1.0 To: i-d-announce@ietf.org Cc: ccamp@ops.ietf.org From: Internet-Drafts@ietf.org Subject: I-D ACTION:draft-ietf-ccamp-gmpls-ason-routing-eval-01.txt Message-Id: <E1DvFtZ-0005tR-Ej@newodin.ietf.org> Date: Wed, 20 Jul 2005 10:50:01 -0400 --NextPart A New Internet-Draft is available from the on-line Internet-Drafts directories. This draft is a work item of the Common Control and Measurement Plane Working Group of the IETF. Title : Evaluation of existing Routing Protocols against ASON Routing Requirements Author(s) : D. Papadimitriou, et al. Filename : draft-ietf-ccamp-gmpls-ason-routing-eval-01.txt Pages : 0 Date : 2005-7-20 The Generalized MPLS (GMPLS) suite of protocols has been defined to control different switching technologies as well as different applications. These include support for requesting TDM connections including SONET/SDH and Optical Transport Networks (OTNs). This document provides an evaluation of the IETF Routing Protocols against the routing requirements for an Automatically Switched Optical Network (ASON) as defined by ITU-T. A URL for this Internet-Draft is: http://www.ietf.org/internet-drafts/draft-ietf-ccamp-gmpls-ason-routing-eval-01.txt To remove yourself from the I-D Announcement list, send a message to i-d-announce-request@ietf.org with the word unsubscribe in the body of the message. You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce to change your subscription settings. Internet-Drafts are also available by anonymous FTP. Login with the username "anonymous" and a password of your e-mail address. After logging in, type "cd internet-drafts" and then "get draft-ietf-ccamp-gmpls-ason-routing-eval-01.txt". A list of Internet-Drafts directories can be found in http://www.ietf.org/shadow.html or ftp://ftp.ietf.org/ietf/1shadow-sites.txt Internet-Drafts can also be obtained by e-mail. Send a message to: mailserv@ietf.org. In the body type: "FILE /internet-drafts/draft-ietf-ccamp-gmpls-ason-routing-eval-01.txt". NOTE: The mail server at ietf.org can return the document in MIME-encoded form by using the "mpack" utility. To use this feature, insert the command "ENCODING mime" before the "FILE" command. To decode the response(s), you will need "munpack" or a MIME-compliant mail reader. Different MIME-compliant mail readers exhibit different behavior, especially when dealing with "multipart" MIME messages (i.e. documents which have been split up into multiple messages), so check your local documentation on how to manipulate these messages. Below is the data which will enable a MIME compliant mail reader implementation to automatically retrieve the ASCII version of the Internet-Draft. --NextPart Content-Type: Multipart/Alternative; Boundary="OtherAccess" --OtherAccess Content-Type: Message/External-body; access-type="mail-server"; server="mailserv@ietf.org" Content-Type: text/plain Content-ID: <2005-7-20103529.I-D@ietf.org> ENCODING mime FILE /internet-drafts/draft-ietf-ccamp-gmpls-ason-routing-eval-01.txt --OtherAccess Content-Type: Message/External-body; name="draft-ietf-ccamp-gmpls-ason-routing-eval-01.txt"; site="ftp.ietf.org"; access-type="anon-ftp"; directory="internet-drafts" Content-Type: text/plain Content-ID: <2005-7-20103529.I-D@ietf.org> --OtherAccess-- --NextPart-- Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 12:26:45 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: RE: Moving LSP ownership between control plane and management plane Date: Wed, 20 Jul 2005 08:25:18 -0400 Message-ID: <29D15BBCA340DA4D8146D38B4924FC7A011223AA@zcarhxm0.corp.nortel.com> Thread-Topic: Moving LSP ownership between control plane and management plane Thread-Index: AcWNHrYFsDY/8jdvSoOTo+HceyEd8QABafkw From: "Stephen Shew" <sdshew@nortel.com> To: "Diego Caviglia" <Diego.Caviglia@marconi.com> Cc: <ccamp@ops.ietf.org> See below. > -----Original Message----- > From: Diego Caviglia [mailto:Diego.Caviglia@marconi.com]=20 > Sent: Wednesday, July 20, 2005 07:33 > To: Shew, Stephen [CAR:QT30:EXCH] > Cc: ccamp@ops.ietf.org > Subject: Re: >=20 >=20 >=20 > Hi Stephen, > my answers in line. >=20 > Regards >=20 > Diego >=20 >=20 >=20 > "Stephen Shew" <sdshew@nortel.com>@ops.ietf.org on 20/07/2005 12.22.49 >=20 > Sent by: owner-ccamp@ops.ietf.org >=20 >=20 > To: "\"ccamp\" <ccamp" > cc: >=20 > Subject: >=20 >=20 > Diego, my suggestion for this draft is that it separate the=20 > mechanism (not setting the cross connects specified in an=20 > ERO), from the reason for doing so. This is because the=20 > mechanism could be used for more than just CP<->MP handoff.=20 > [dc] Hmmm I've defined that application mainly for two reason: > 1 up to my knowledge that is the 'killer application' for=20 > this kind of > protocol modification > 2 my understanding of CCAMP is that a good motivation needs to be > provided to modify GRSVP-TE and this leads to point 1 I have no issue with the motivation. I just think there are more scenarios for creating control plane state for an existing cross connection. > If you do this then the bit should be renamed something like=20 > "cross-connection exists" and if set, then the signaling=20 > protocol does not tell the cross connection to set/release=20 > it. [dc] That is how it works apart for the name of the flag. I realize that. What I'm suggesting is that the name reflect an action rather than one of several possible purposes. > Did you consider a generalization of this where in the ERO,=20 > each element has this type of flag? If so, you could use the=20 > mechanism to say replace two LSPs where the tail of one is=20 > next to the head of the other, with a single LSP. The action=20 > would be to signal an ERO containing connection details of=20 > each LSP that have the flag on, and where they join, have the=20 > flag off. [dc] If I've understood corretly your point, what=20 > you are saying is a little bir different from what the ID=20 > proposes. I mean in your exaple you have two different LSPs=20 > that already have some control plane information associated=20 > with them and you want to splice them together. In this case=20 > you have to replace the CP information while the ID proposes=20 > to add CP information to a XC that don't have it. Do you=20 > agree with that? The suggestion was that there are other scenarios in which establishment of new LSP state for an existing cross connection is useful. In the MP->CP case, the MP initially holds state for the connection and before the action is completed, both MP and CP hold state. In the two LSP example above, two entities in the CP hold state about the same connection. I see these two cases as similar in that the general case is about multiple entities holding state for the same resource. >=20 > Stephen >=20 >=20 > > -----Original Message----- > > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On=20 > > Behalf Of Diego Caviglia > > Sent: Wednesday, July 20, 2005 05:44 > > To: Diego.Caviglia@marconi.com > > Cc: ""Dimitri.Papadimitriou" <Dimitri.Papadimitriou";=20 > "ccamp" <ccamp;=20 > > "gnewsome" <gnewsome; "adrian" <adrian > > Subject: RE: Moving LSP ownership between control plane and=20 > management=20 > > plane > > > > > > > > Sorry for this but may be not all have read the ID, I hope=20 > that having=20 > > the text help the discussion. > > > > Regards > > > > Diego > > > > > > > > > > Network Working Group > > Internet Draft > > Diego Caviglia > > Document: draft-caviglia-mp2cpcp2mp-02.txt Marconi > > Proposed Status: Updates RFC 3473 > > Dino Bramanti > > > > Marconi > > Expires: December 2005 > > Nicola Ciulli >=20 > <snip> >=20 >=20 >=20 >=20 >=20 >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 11:39:45 +0000 Message-ID: <154801c58d1f$f063c4a0$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: "Arthi Ayyangar" <arthi@juniper.net>, <Dimitri.Papadimitriou@alcatel.be> Cc: "Igor Bryskin" <ibryskin@movaz.com>, <ccamp@ops.ietf.org> Subject: Re: LSP stitching questions for discussion Date: Wed, 20 Jul 2005 12:28:48 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi all, > first a recap: this item on stitching was initiated in order to allow per > "domain" control of end-to-end (contiguous) LSP, as part of a separate > document, and not to extend its scope (unduly, in this case, see below) > but mainly in order to have the set of all related protocol mechanisms > within a specific document - I partially agree. I would say that stitching was first introduced in the context of per-domain control of end-to-end LSPs, but that it quickly became apparent that stitching was an entity in its own right. That is, stitching is a control plane and data plane technique. I wanted to separate the function from the applicability, and in particular I wanted to make sure that the function was not burried inside an inter-domain I-D that would hide the function and potentially limit its applicability. Thus, the chairs insisted that the work was pulled out into a separate I-D that, as you say, has the set of all related protocol mechanisms within a specific document. > from this initial working assumption, it is clear that the segment must have > the properties of the incoming end-to-end LSP which are determined among > other by its switching type e.g. PSC-3 or TDM, the first requirement of LSP > stitching is contiguity of the end-to-end LSP at the data plane level (after > stitching operation) - and this by definition of the initial working assumption - Well, you'll note that I did not quite agree with your initial working assumption. I am one of those people that fit into the category that Igor outlined: I believe that it is technically possible to stitch a segment of a lower layer between two segments of a higher layer. This is *far* from saying that I can see a good reason to do this. Indeed, it seems that the mechanisms we are developing say that when domains exist in distinct layers we would normally use hierarchical LSPs, and when domains exist in the same layer we would use stitched LSPs. We should think carefully, however, before specifying in the control plane that a specific hardware dpleoyment is illegal. > note: the only latitude which is left (for a given switching type e.g. PSC-2) is due > to the properties of packet (PSC) networks that allows you stitch a segment > having a larger Max Reservable Bandwidth and Unreserved Bandwidth (we > would have equivalent latitude for other attributes but this is a local policy > decision - which at the end is the sole purpose of this approach) > > > Here are some of the questions related to LSP stitching to start a > > discussion on the usage of LSP stitching. > > > >1) Are the control plane signaling procedures for LSP stitching as > >described in this ID, (explicit request for LSP stitching on > >LSP segment, different label allocation rules; etc) REQUIRED to stitch > >packet LSPs in data plane ? > > -> yes (beside manual config operation) > > >2) Are the control plane signaling procedures for LSP stitching as > >described in this ID, (explicit request for LSP stitching on > >LSP segment, different label allocation rules; etc) REQUIRED to stitch > >non-packet LSPs in data plane ? > > -> yes (beside manual config operation) > > >3) If an e2e LSP crosses a region boundary (based on different TE > >link switching capabilities), then can I use LSP stitching - > >a) control plane procedures for packet LSPs > >b) data plane procedures for packet LSPs > >c) control plane procedures for non-packet LSPs > >d) data plane procedures for non-packet LSPs > > > -> no, when crossing a region boundary, you simply use a document > named "draft-mpls-lsp-hierarchy-08.txt" which is in the RFC Editor > queue, but this does not prevent from creating an LSP segment on top > of an FA-LSP (see below) I must ask "why not?" If the answer is that it cannot be done in hardware, I think this is incorrect. If the answer is that it does not need to be done because hierarchical LSPs exist, then I can buy that argument, but only because I cannot think of a reason to use stitched segments in this way. > >4) If the switching type of an e2e LSP is different from that of an LSP > >segment, can I use LSP stitching - > >a) control plane procedures for packet LSPs > >b) data plane procedures for packet LSPs > >c) control plane procedures for non-packet LSPs > >d) data plane procedures for non-packet LSPs > > -> no, in this case progression of the end-to-end LSP establishment, implies > that the LSP segment must fulfil a set of constraints carried as part of the > incoming end-to-end LSP request, in this case one has to create an LSP > segment (with a switching type equal to the incoming end-to-end LSP > switching type) over that FA (i.e. referred here above to the LSP segment > with a different Switching Type) I don't see any difference between questions 3 and 4 notwithstanding the explanatory paragraph below. Please note that, given the appropriate adaptation capabilities I can advertise an LSP that crosses a lower layer as a TE link in a higher layer. When I do this, I advertise the switching capabilities of the TE link as those of the higher layer, not the lower layer. The same technique could be applied for stitching segments (provided suitable adaptation exists). Agreed this might represent a waste of bandwidth in the lower layer. Agreed there appears to be no reason to do this when hierarchical LSPs can be created. I just want to separate requirements/desires from technical possibilities. > >Note that although 3) and 4) appear to be the same, the reason I have > >them as different questions is that in one case the TE link switching > >capability is examined and in the other case the Generalized Label Request > >is examined. > > -> what it appears to me is that this discussion mixes the issue of the operation > describing the stitching operations at the domain boundaries from the LSP > segment creation itself i.e. an LSP segment can indeed be an FA link inherited > from the creation of an (intra-domain) FA-LSP - Yes. It is worth noting that if you create a hierarchical LSP *within* a domain, you have necessarily created a layer (or region) boundary. If you create a stitching segment within a domain, you do not do this. Layer boundaries may not be important in PSC networks since PSC1 and PSC2 are very similar. But layer boundaries are really important in TDM, LSC and FSC networks. In these cases stitching segments within the layer may play an important role for domain sub-division. Adrian Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 11:32:43 +0000 Sensitivity: Subject: Re: To: sdshew@nortel.com Cc: ccamp@ops.ietf.org Message-ID: <OF93C1D6B0.04A02BB2-ONC1257044.003E2F43-C1257044.003F6C38@uk.marconicomms.com> From: "Diego Caviglia" <Diego.Caviglia@marconi.com> Date: Wed, 20 Jul 2005 13:32:40 +0200 MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii Hi Stephen, my answers in line. Regards Diego "Stephen Shew" <sdshew@nortel.com>@ops.ietf.org on 20/07/2005 12.22.49 Sent by: owner-ccamp@ops.ietf.org To: "\"ccamp\" <ccamp" cc: Subject: Diego, my suggestion for this draft is that it separate the mechanism (not setting the cross connects specified in an ERO), from the reason for doing so. This is because the mechanism could be used for more than just CP<->MP handoff. [dc] Hmmm I've defined that application mainly for two reason: 1 up to my knowledge that is the 'killer application' for this kind of protocol modification 2 my understanding of CCAMP is that a good motivation needs to be provided to modify GRSVP-TE and this leads to point 1 If you do this then the bit should be renamed something like "cross-connection exists" and if set, then the signaling protocol does not tell the cross connection to set/release it. [dc] That is how it works apart for the name of the flag. Did you consider a generalization of this where in the ERO, each element has this type of flag? If so, you could use the mechanism to say replace two LSPs where the tail of one is next to the head of the other, with a single LSP. The action would be to signal an ERO containing connection details of each LSP that have the flag on, and where they join, have the flag off. [dc] If I've understood corretly your point, what you are saying is a little bir different from what the ID proposes. I mean in your exaple you have two different LSPs that already have some control plane information associated with them and you want to splice them together. In this case you have to replace the CP information while the ID proposes to add CP information to a XC that don't have it. Do you agree with that? Stephen > -----Original Message----- > From: owner-ccamp@ops.ietf.org > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Diego Caviglia > Sent: Wednesday, July 20, 2005 05:44 > To: Diego.Caviglia@marconi.com > Cc: ""Dimitri.Papadimitriou" <Dimitri.Papadimitriou"; "ccamp" > <ccamp; "gnewsome" <gnewsome; "adrian" <adrian > Subject: RE: Moving LSP ownership between control plane and > management plane > > > > Sorry for this but may be not all have read the ID, I hope > that having the text help the discussion. > > Regards > > Diego > > > > > Network Working Group > Internet Draft > Diego Caviglia > Document: draft-caviglia-mp2cpcp2mp-02.txt Marconi > Proposed Status: Updates RFC 3473 > Dino Bramanti > > Marconi > Expires: December 2005 > Nicola Ciulli <snip> Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 11:25:10 +0000 To: "Diego Caviglia" <Diego.Caviglia@marconi.com> Cc: "Dimitri.Papadimitriou" <Dimitri.Papadimitriou@alcatel.be>, "ccamp" <ccamp@ops.ietf.org>, "gnewsome" <gnewsome@ieee.org>, "adrian" <adrian@olddog.co.uk> MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: RE: Moving LSP ownership between control plane and management plane Date: Wed, 20 Jul 2005 13:23:42 +0200 Message-ID: <OF3EC0D474.82BF2A2D-ONC1257044.003E97A5-C1257044.003E9879@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmRpZWdvIDwvRk9OVD48L1A+PFA+PEZP TlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmFzIHBhcnQgb2YgdGhlIG1vdGl2YXRpb24gc2Vj dGlvbiBvZiB0aGlzIGRvY3VtZW50IGl0IHdvdWxkIGJlIGluZGVlZCBpbnRlcmVzdGluZyB0byBi ZSBtdWNoIG1vcmUgZXhwbGljaXQgYWJvdXQgdGhlIENQLSZndDtNUCBoYW5kb3ZlciBwdXJwb3Nl KHMpIGJlZm9yZSBwcm9ncmVzc2luZyBvbiB0aGlzIHBhcnQgb2YgdGhlIG1lY2hhbmlzbTwvRk9O VD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPi0gZC48L0ZPTlQ+PEJSPjxC Uj48Rk9OVCBTSVpFPTI+PEI+JnF1b3Q7RGllZ28gQ2F2aWdsaWEmcXVvdDsgJmx0O0RpZWdvLkNh dmlnbGlhQG1hcmNvbmkuY29tJmd0OzwvQj48L0ZPTlQ+PEJSPjxGT05UIFNJWkU9Mj5TZW50IGJ5 OiBvd25lci1jY2FtcEBvcHMuaWV0Zi5vcmc8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9Mj4wNy8yMC8y MDA1IDExOjM3IFpFMjwvRk9OVD48QlI+PEJSPiA8Rk9OVCBTSVpFPTI+VG86PC9GT05UPiA8Rk9O VCBTSVpFPTI+RGltaXRyaSBQQVBBRElNSVRSSU9VL0JFL0FMQ0FURUxAQUxDQVRFTDwvRk9OVD48 QlI+IDxGT05UIFNJWkU9Mj5jYzo8L0ZPTlQ+IDxGT05UIFNJWkU9Mj4mcXVvdDtjY2FtcCZxdW90 OyAmbHQ7Y2NhbXBAb3BzLmlldGYub3JnJmd0OywgJnF1b3Q7Z25ld3NvbWUmcXVvdDsgJmx0O2du ZXdzb21lQGllZWUub3JnJmd0OywgJnF1b3Q7YWRyaWFuJnF1b3Q7ICZsdDthZHJpYW5Ab2xkZG9n LmNvLnVrJmd0OzwvRk9OVD48QlI+IDxGT05UIFNJWkU9Mj5iY2M6PC9GT05UPiA8QlI+IDxGT05U IFNJWkU9Mj5TdWJqZWN0OjwvRk9OVD4gPEZPTlQgU0laRT0yPlJFOiBNb3ZpbmcgTFNQIG93bmVy c2hpcCBiZXR3ZWVuIGNvbnRyb2wgcGxhbmUgYW5kIG1hbmFnZW1lbnQgcGxhbmU8L0ZPTlQ+PEJS PiA8QlI+PEJSPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+RGltaXRyaSw8 QlI+eWVzIG9mIGNvdXJzZSB0aGUgbWFpbiBhcHBsaWNhdGlvbiBpcyBNUC0tJmd0O0NQIHRoZSBD UC0tJmd0O01QIGlzIGp1c3Q8L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVy Ij5mb3Igc2ltbWV0cnk7LSkpPEJSPjwvRk9OVD48QlI+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENv dXJpZXIiPkRpZWdvPEJSPjwvRk9OVD48QlI+PEJSPjxCUj48Rk9OVCBGQUNFPSJNb25vc3BhY2Us Q291cmllciI+RGltaXRyaS5QYXBhZGltaXRyaW91QGFsY2F0ZWwuYmUgb24gMjAvMDcvMjAwNSAx MS4zMC4zNTxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5Ubzog Jm5ic3A7ICZuYnNwOyZxdW90O0RpZWdvIENhdmlnbGlhJnF1b3Q7ICZsdDtEaWVnby5DYXZpZ2xp YUBtYXJjb25pLmNvbSZndDs8QlI+Y2M6ICZuYnNwOyAmbmJzcDtjY2FtcEBvcHMuaWV0Zi5vcmcs IGduZXdzb21lQGllZWUub3JnLCBhZHJpYW5Ab2xkZG9nLmNvLnVrPEJSPjwvRk9OVD48QlI+PEZP TlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPlN1YmplY3Q6ICZuYnNwOyAmbmJzcDtSRTogTW92 aW5nIExTUCBvd25lcnNoaXAgYmV0d2VlbiBjb250cm9sIHBsYW5lIGFuZCBtYW5hZ2VtZW50PEJS PnBsYW5lPC9GT05UPjxCUj48QlI+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmRpZWdv IC08QlI+PC9GT05UPjxCUj48QlI+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPiZxdW90 O015IGJhc2ljIHF1ZXN0aW9uIGlzOiBkbyBzb21lb25lIHRoaW5rIHRoYXQgaGF2aW5nIGEgbWVj aGFuaXNtIHRvIG1vdmU8QlI+TVAvQ1AgY3JlYXRlZCBMU1AgdW5kZXIgdGhlIENQL01QIGRvbWFp biBpcyB1c2VsZXNzPyZxdW90OzxCUj48L0ZPTlQ+PEJSPjxCUj48QlI+PEJSPjxCUj48Rk9OVCBG QUNFPSJNb25vc3BhY2UsQ291cmllciI+bm8gLTxCUj48L0ZPTlQ+PEJSPjxCUj48QlI+PEJSPjxC Uj48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+aW4gcGFydC4gdGhlIE1QLSZndDtDUCBo YW5kb3ZlciBpcyBhIGtleSBtZWNoYW5pc20gZm9yIG1pZ3JhdGlvbiB0b3dhcmQgR01QTFM8QlI+ Q1A8QlI+PC9GT05UPjxCUj48QlI+PEJSPjxCUj48QlI+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENv dXJpZXIiPi1kaW1pdHJpLjxCUj48L0ZPTlQ+PEJSPjxCUj48QlI+PEJSPjxCUj48QlI+PEJSPjxC Uj48QlI+PEJSPjxCUj48L1A+ Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 10:29:30 +0000 Message-ID: <146101c58d16$384833a0$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: "Tom Petch" <nwnetworks@dial.pipex.com>, <ccamp@ops.ietf.org> Subject: Re: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON Lexicography Date: Wed, 20 Jul 2005 10:54:04 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Tom, I don't think it is CCAMP's role to correct the English of another SDO :-) Thanks, Adrian ----- Original Message ----- From: "Tom Petch" <nwnetworks@dial.pipex.com> To: "Adrian Farrel" <adrian@olddog.co.uk>; <ccamp@ops.ietf.org> Sent: Tuesday, July 19, 2005 7:18 PM Subject: Re: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON Lexicography > Adrian > > I think it should be > " > - CCAMP has primary responsibility for the descriptions of GMPLS terms > > Correct " > > even if that is not what SG15 said. > > Primarily (sic) CCAMP has responsibility for the specification of GMPLS, not > just its terms:-) > > Tom Petch > > ----- Original Message ----- > From: "Adrian Farrel" <adrian@olddog.co.uk> > To: <ccamp@ops.ietf.org> > Cc: "'Kireeti Kompella'" <kireeti@juniper.net>; <zinin@psg.com>; "Bill Fenner" > <fenner@research.att.com>; "Scott Bradner" <sob@harvard.edu> > Sent: Tuesday, July 19, 2005 1:59 PM > Subject: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON > Lexicography > > > > Hi, > > > > Please comment on this draft response liaison which I intend to send on > > Monday 25th July. > > You can see the original incoming liaison at > > http://www.olddog.co.uk/incoming.htm > > > > Thanks, > > Adrian > > ===== > > To: ITU-T SG15 Questions 12 and 14 > > From: IETF CCAMP Working Group > > Subject: Response to your liaison on GMPLS/ASON Lexicography > > For: Action > > Deadline: August 31st 2005 > > > > The CCAMP Working Group would like to thank SG15 and especially Questions > > 12 and 14 for the time and effort they put into providing input and > > feedback on the GMPLS/ASON lexicography Internet-Draft that is being > > developed by CCAMP. Much of the material supplied as direct comments or as > > mark-ups to the Internet-Draft has been incorporated into the latest > > revision which is available at > > > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-gmpls-ason-lexicography-03. > txt > > > > In your liaison you state some assumptions which we would like to comment > > on as follows. > > > > - The purpose of the document is to enable those familiar with ASON > > to understand GMPLS terminology in an ASON context > > > > Correct. > > > > - GMPLS terms are described informally in the document in a way that > > is generally consistent with GMPLS (existing RFCs and work in progress) > > > > While the definition of GMPLS terms would be helpful for understanding > > GMPLS terms in a GMPLS context, that is not the purpose of this document. > > > > - Based on these descriptions, interpretations of these GMPLS terms > > in ASON terminology are provided > > > > It is intended that the interpretation of the GMPLS terms is not based on > > these descriptions, but based on the full meaning of the GMPLS terms. The > > descriptions of the GMPLS terms are provided in this Internet-Draft for > > context and a brief summary. > > > > - CCAMP has primarily responsibility for the descriptions of GMPLS terms > > > > Correct. > > > > - Q12/15 has responsibility to provide input to the GMPLS-ASON > > interpretations (based on the descriptions in this draft) > > > > We welcome this assumption of responsibility by Q12/15 to provide input to > > the Internet-Draft. While the Internet-Draft remains under the overall > > editorial control of the CCAMP working group, we hope to be able to use > > the text supplied by Q12/15 with only editorial changes. > > > > > > Additionally, we welcome the pledge to continue work on this > > Internet-Draft through correspondence on the Q12/15 mailing list, and > > thank Q12/15 for assigning Ben Mack-Crane as the SG15 representative on > > this matter. Ben has proved very helpful in clarifying and developing > > comments received from SG15. > > > > > > We have had several email exchanges with Ben about a few of the specific > > comments made in your review of the 02 revision of the Internet-Draft. > > Although he has taken these comments to the Q12/15 mailing list no > > response has been received, perhaps due to pressure of time. We are taking > > the opportunity to restate several questions here for your convenience and > > would appreciate your answers as part of the response to this liaison. > > > > 1. Definition of a Resource (section 3.2) > > You have supplied us with the simple text that says... > > In ASON applications: > > A resource refers only to a link connection (e.g. VC-11, > > VC-4 etc.). > > > > We would like to clarify this because we don't understand how > > the original text is wrong. It used to say... > > ITU-T [should be ASON] terms for resource: > > - Connection point (cp) in the context of link discovery > > and resource management (allocation, binding into > > cross-connects, etc.); > > - Link connection or trail termination in the context of > > routing, path computation and signaling. > > > > Recall that we are talking about the ASON terms for what GMPLS > > calls a resource. We think the following... > > In ASON a link connection is an association of output cp on one > > side of the link and input cp on the other side of the link. In > > GMPLS by resource we mean in most of the cases the local end of > > the resource (which is cp in ASON terminology). This is true in > > the context of link discovery. > > > > More importantly, this is also true in the context of a > > signaling application. Specifically, when a GMPLS signaling > > sub-system requests allocation of a resource, the GMPLS > > Resource Manager allocates only the local end of the > > resource. This happens on both sides of the link, that is, we > > allocate input and output resources. This contrasts with ASON, > > where there is a much more versatile LRM, and the connection > > manager (signaling application) allocates link connections > > only on the output side, leaving LRM to coordinate the > > allocation with the neighbor. > > > > However, in GMPLS TE advertisements we account only for > > outgoing resources, that is, we don't advertise incoming > > resources on the assumption that they match the outgoing > > resources for the discovered TE links. > > > > Bottom line: we believe that we should retain the previous > > definitions, and have not made this change in the new > > revision. We would like to discuss this point further with > > Q12/15. > > > > 2. Link ends and Data links (Section 3.5 - was section 3.4) > > > > 2a.Although we understand that ASON does not require the concept > > of a link end, GMPLS does. Therefore we should provide some > > form of language to help people do the mapping. > > > > The proposed new text from Q12/15 removes all reference to > > link ends from the ASON section, and we feel this is a > > mistake and propose the following text: > > In ASON, a GMPLS unidirectional data link end is a > > collection of connection points from the same client > > layer that are supported by a single trail termination > > (access point). > > > > This text depends on the fact that a link end is described as > > a set of resources that transfer traffic to a neighbor. > > > > 2b.The new text from Q12/15 gives us... > > A GMPLS data link is an ASON link supported by a single > > server trail. > > > > We are not sure that we understand this correctly. It > > appears to say that a unidirectional data link could be > > supported by multiple trail terminations (access points). > > This seems to suggest that a client would have to make a > > choice about where to send data, which seems to imply it > > really has to choose between data links! We are confused! > > > > We think some of this problem may come from the *need* in > > GMPLS to identify data link ends. There may be some value in > > stressing that a GMPLS link end is supported by exactly one > > trail termination. When several trail terminations are > > involved, we can still talk about single TE link (as a > > bundle), but each component link will be a distinct data > > link. We think that in ASON the SNPP (TE link in GMPLS > > language) could be supported by multiple terminations. > > > > 3. Link interfaces (section 3.6 - was 3.5) > > The new text from Q12/15 gives us... > > A GMPLS interface is "all the stuff between a physical > > interface and a matrix in an ASON network element." More > > precisely, it is a trail termination function and adaptation > > function for which adapted client layer connection points > > are bound to a matrix. > > > > We think "physical interface" may be ambiguous in the multi- > > layer context. We think the "stuff" should be between a link > > and a matrix, since we are talking about links in distinct > > layers. > > > > > > In addition to your feedback on the specific points above, we would > > welcome your continued comments and review of the Internet-Draft. > > > > Regards, > > Adrian Farrel and Kireeti Kompella > > IETF CCAMP Working Group Co-chairs > > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 10:29:24 +0000 Message-ID: <145201c58d16$30300ee0$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: "Scott Brim \(sbrim\)" <sbrim@cisco.com> Cc: <ccamp@ops.ietf.org> Subject: Re: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON Lexicography Date: Wed, 20 Jul 2005 10:45:16 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Scott, In general, providing a long list of RFCs and I-Ds would probably not be helpful. In any case, it can be obtained by looking at the CCAMP and MPLS charter pages. On several occasions someone has suggested compiling a compendium of GMPLS terms with cross-references to RFCs and I-Ds, while others have said this is not worth the effort because it would be necessary to read the referenced documents in their entirety anyway. The chairs have always said that they would welcome such a compendium if someone wrote it and did a thorough job, but no-one has stepped up to do this task. The main point of my original comment is that the Lexicography is not this compendium. Thanks, Adrian ----- Original Message ----- From: "Scott Brim (sbrim)" <sbrim@cisco.com> To: <ccamp@ops.ietf.org> Sent: Tuesday, July 19, 2005 8:30 PM Subject: RE: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON Lexicography On Tuesday, July 19, 2005 06:59 AM, Adrian Farrel allegedly wrote: > While the definition of GMPLS terms would be helpful for > understanding GMPLS terms in a GMPLS context, that is not the > purpose of this document. It would be polite to tell them where to find the definions :-) Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 10:29:16 +0000 Message-ID: <145f01c58d16$318b0b00$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <ccamp@ops.ietf.org>, <inoue.ichiro@lab.ntt.co.jp> Cc: "Thomas D. Nadeau" <tnadeau@cisco.com>, <zinin@psg.com>, "Bill Fenner" <fenner@research.att.com> Subject: Re: no more milestones? Date: Wed, 20 Jul 2005 10:49:07 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi Ichiro, In the agenda (now on-line at http://www.olddog.co.uk/63/ccamp.htm) you'll see that we have set aside just over 30 minutes to discuss the work plan for CCAMP. The chairs will prepare some material based on the discussions last November and the new work that has come forward since then. We will post this material in advance of the meeting to give people a chance to think and talk about it. Hopefully, in the meeting or before, we can discover the process we should follow and go on to clarify what the working group should be working on for the next 12 months or so. Thanks, Adrian > I agree that base signalling and routing GMPLS protocols need to be > worked out, and at the same time I would like to see GMPLS applications > in operational networks. > > I remember about the vote on specific items (to-be-rechartered items) which > included basic signalling and routing aspects as well as some extended > applications. > At that time, I felt we had a long and good discussion already about what > would be on the revised charter and it was very close to decide. > > Now, though (because?) I don't know what process is done and further needed > beyond what we already went through and saw, I appreciate if we can see new > milestones reflecting the vote and recent discussions so that we can determine in > Paris hopefully without, but with some more discussion if needed. > > Thanks. > > Ichiro Inoue > > > >> I would like to see: > > >> - a refinement of the charter to focus the next work items > > >> - a year's worth of milestones to ensure we stay focused. > > >> > > >> On the other hand, we should aim to close down CCAMP and move the > > >> work to > > >> other or new working groups. > > >> > > > > > > coupling both sentences does it mean in your view > > > > > > 1) define several short-term milestones for refining topics that > > > have been > > > already initiated and requires further well scoped attention > > > > > > 2) (brand) new topics should target new working groups - as already > > > initiated with L1VPN for inst. - or existing working groups > > > > > > nevertheless, i do not see why this automatically means CCAMP can not > > > remain the focal point concerning *base* GMPLS signaling and routing > > > protocols > > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 10:23:32 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: Date: Wed, 20 Jul 2005 06:22:49 -0400 Message-ID: <29D15BBCA340DA4D8146D38B4924FC7A011223A8@zcarhxm0.corp.nortel.com> Thread-Index: AcWNFPrDLlo2GHD0TOCErNAWoJnGlQ== From: "Stephen Shew" <sdshew@nortel.com> To: "\"ccamp\" <ccamp" <ccamp@ops.ietf.org> Diego, my suggestion for this draft is that it separate the mechanism (not setting the cross connects specified in an ERO), from the reason for doing so. This is because the mechanism could be used for more than just CP<->MP handoff. If you do this then the bit should be renamed something like "cross-connection exists" and if set, then the signaling protocol does not tell the cross connection to set/release it. Did you consider a generalization of this where in the ERO, each element has this type of flag? If so, you could use the mechanism to say replace two LSPs where the tail of one is next to the head of the other, with a single LSP. The action would be to signal an ERO containing connection details of each LSP that have the flag on, and where they join, have the flag off. Stephen > -----Original Message----- > From: owner-ccamp@ops.ietf.org=20 > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Diego Caviglia > Sent: Wednesday, July 20, 2005 05:44 > To: Diego.Caviglia@marconi.com > Cc: ""Dimitri.Papadimitriou" <Dimitri.Papadimitriou"; "ccamp"=20 > <ccamp; "gnewsome" <gnewsome; "adrian" <adrian > Subject: RE: Moving LSP ownership between control plane and=20 > management plane >=20 >=20 >=20 > Sorry for this but may be not all have read the ID, I hope=20 > that having the text help the discussion. >=20 > Regards >=20 > Diego >=20 >=20 >=20 >=20 > Network Working Group > Internet Draft =20 > Diego Caviglia > Document: draft-caviglia-mp2cpcp2mp-02.txt Marconi > Proposed Status: Updates RFC 3473 =20 > Dino Bramanti > =20 > Marconi > Expires: December 2005 =20 > Nicola Ciulli <snip>=20 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 09:54:35 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: Date: Wed, 20 Jul 2005 05:54:07 -0400 Message-ID: <29D15BBCA340DA4D8146D38B4924FC7A011223A7@zcarhxm0.corp.nortel.com> Thread-Index: AcWNEPgdeHoo0tewQNGL9l/vVMgTtQ== From: "Stephen Shew" <sdshew@nortel.com> To: <ccamp@ops.ietf.org> Adrian, see comment below. > -----Original Message----- > From: owner-ccamp@ops.ietf.org=20 > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Adrian Farrel > Sent: Tuesday, July 19, 2005 10:02 > To: George Newsome > Cc: ccamp@ops.ietf.org > Subject: Re: Moving LSP ownership between control plane and=20 > management plane >=20 >=20 > Thanks George, >=20 > Useful input. <snip> > > - Call awareness of migration vs. connection being unaware=20 > of migration. > > When connection state is being created to match an existing=20 > > connection, the connection controllers (CCs) do not require=20 > awareness=20 > > of why this is happening as the context could be migration or=20 > > recovery. A mechanism to create control state without affecting=20 > > transport plane state is needed in the CCs. Call=20 > controllers though=20 > > do need migration awareness as they need to obtain/derive call=20 > > identification from the MP. >=20 > I disagree. > Connection controllers must be able to distinguish between=20 > resources that are already in use and cannot be assigned to a=20 > new connection, and those that are already in use and can be=20 > assigned to a new connection because the connection is=20 > replacing an existing connection. If this were happening=20 > entirely within the control plane we would use resource=20 > sharing, but when the operation spans the control and=20 > management planes resource ownership/sharing is more complex=20 > and the conneciotn controller needs to be able to issue the=20 > approriate instrucitons for the assignment of the resources.=20 > On the other hand, call controllers do not need awareness of=20 > the details of the realisation of the connections. I'm not sure if the difference between migration awareness and the action of creating connection state for an existing transport connection is clear in the original point. I agree with your first sentence and think that signalling could be extended to include the case where connection details (in a complete ERO) are supplied with the attribute that the cross connections are already in place. Similarly, teardown of connection state can occur where cross connections are not undone. I still agree though with the original point and note that it does not say that call controllers are aware of connection details. Call controllers are aware that a migration action is the reason for triggering the connection controller to establish state. The point about awareness refers to an understanding of why LSP state is being created. If existing LSP state is being replaced by another one (e.g., a recovery scenario or maybe a change of the extent of the transport connection), then I think it's sufficient for the connection controller to be told that it should not change the cross connection (matrix level subnetwork connection). If the overall action fails for some reason, the call entity can then decide whether to try again, and/or tell the management plane what happened. I believe that this is consistent with what the existing protocol(s) know when they are creating LSPs. For example, a protocol instance is not aware whether the connection being created is a result of a "break before make" or "make before break" scenario, or that it is the 3rd connection being created for a VCAT group. This understanding should only be known by the initiator of the action. Stephen Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 09:44:43 +0000 Sensitivity: Subject: RE: Moving LSP ownership between control plane and management plane To: Diego.Caviglia@marconi.com Cc: "\"\"Dimitri.Papadimitriou\" <Dimitri.Papadimitriou\"" <Dimitri.Papadimitriou@alcatel.be>, ""ccamp" <ccamp" <ccamp@ops.ietf.org>, ""gnewsome" <gnewsome" <gnewsome@ieee.org>, ""adrian" <adrian" <adrian@olddog.co.uk> Message-ID: <OF1CB03843.890F3673-ONC1257044.0035637B-C1257044.00357E75@uk.marconicomms.com> From: "Diego Caviglia" <Diego.Caviglia@marconi.com> Date: Wed, 20 Jul 2005 11:44:13 +0200 MIME-Version: 1.0 Content-type: text/plain; charset=iso-8859-1 Content-transfer-encoding: quoted-printable Sorry for this but may be not all have read the ID, I hope that having = the text help the discussion. Regards Diego Network Working Group Internet Draft Diego Caviglia= Document: draft-caviglia-mp2cpcp2mp-02.txt Marconi Proposed Status: Updates RFC 3473 Dino Bramanti= Marconi= Expires: December 2005 Nicola Ciulli= NextWorks= June 2005= GMPLS Signaling Extensions for the Transfer of Ownership of Label Switched Paths Between the Management and Control Planes By submitting this Internet-Draft, each author represents that any applicable patent or other IPR claims of which he or she is aware have been or will be disclosed, and any of which he or she becomes aware will be disclosed, in accordance with Section 6 of BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet- Drafts. Internet-Drafts are draft documents valid for a maximum of six month= s and may be updated, replaced, or obsoleted by other documents at any= time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress=C6=C6. The list of current Internet-Drafts can be accessed at http://www.ietf.org/1id-abstracts.html The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html Abstract During migration scenarios it may be desirable to transfer the ownership of a Label Switched Path (LSP) from the Management Plane (MP) to the Control Plane (CP), or vice versa. If the LSP is carryin= g traffic this change needs to be made "in service," that is, without affecting traffic. This memo provides minor extensions to the Generalized Multi Protoco= l Label Switching (GMPLS) signaling protocol, GRSVP_TE (Generalized Resource Reservation Protocol with Traffic Engineering Extensions), Caviglia et al. Expires - December 2005 [Page 1] =0C draft-caviglia-mp2cpcp2mp-02 June 20= 05 to enable such transfer of ownership and describes the proposed procedures. Conventions used in this document The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in thi= s document are to be interpreted as described in RFC-2119 [1]. Table of Contents 1. Introduction...................................................2 2. Problem Explanation............................................3 3. Proposed solution..............................................3 4. LSP Ownership Handover Procedure Between Management And Control Planes............................................................5 4.1 "MP to CP handover" - LSP Ownership Transfer From Management Plane To Control Plane.........................................5 4.2 MP to CP Handover Procedure Failure Handling...............8 4.3 "CP to MP handover" - LSP Ownership Transfer From Control Plane To Management Plane..................................8 4.4 CP to MP Handover Procedure Failure Handling...............9 5. RSVP Message Formats..........................................10 5.1 Object Modification.......................................10 6. Security Considerations.......................................10 7. IANA Consideration............................................10 8. References....................................................11 9. Acknowledgments...............................................11 Author's Addresses...............................................11 1. Introduction In a typical, traditional transport network scenario, Data Plane connections between two endpoints are controlled basically by means of a Network Management System (NMS) operating within Management Plane (MP). NMS/MP is the owner of such transport connections, being= responsible of their set up, tear down and maintenance. The adoption= and use of a GMPLS Control Plane over networks that are controlled b= y NMS at Management Plane level and that are already in service could not be considered a green field application. In this context, let's indicate with the term Label Switched Path (LSP) the Data Plane forwarding path and Data Plane state associated to a connection, whose control is owned by either Management or Control Plane (via GRSVP-TE). In a mixed scenario, LSPs owned by Management Plane and LSPs owned by Control Plane have to coexist and a way to move their control and ownership between planes, while preserving corresponding= Caviglia et al. Expires - December 2005 [Page 2] =0C draft-caviglia-mp2cpcp2mp-02 June 20= 05 Data Plane traffic, is needed. It is in fact possible that a network= operator wants to move the control of a Management Plane owned transport connection to Control Plane and, in the same way, the opposite operation is also needed. In this memo let's indicate with "MP to CP handover" the procedure that is aimed at moving the contro= l and ownership of a transport connection owned by Management Plane to= Control Plane. Let's call "CP to MP handover" the opposite procedure= , aimed at moving the ownership of a LSP from Control Plane to Management Plane. 2. Problem Explanation Having the ownership of a LSP means basically having the access to the information associated to a physical Data Plane connection between two endpoints (ingress and egress of LSP) that traverses a list of nodes. The owner of a LSP has to store and use such information (LSP hop list, associated TE information etc.) to contro= l completely the LSP. Let's take the case of a Data Plane connection between an ingress node and an egress node, which has been set up by= means of NMS. The network resources allocated to this connection cannot be used or controlled by GRSVP-TE as it doesn't have any information record about it. If a standard GRSVP-TE setup request over the same resources was issued, the nodes will reject it, as the= y find the involved resources already allocated. A standard GRSVP-TE release request for that LSP wont be successful as well, because LSP= related information to be matched with the one sent in signaling messages is not available at Control Plane level. If a Data Plane connection between an ingress node and an egress has= been set up via GRSVP-TE signaling, all the info related to the LSP is present at Control Plane level which is the owner of that connection. If NMS wants to take over the LSP, the Control Plane has= to be informed that involved resources are no more under its control= . In both cases, the point is that the Data Plane connection has to stay up, but related information allowing control of the LSP has to be handed over to Control Plane or Management Plane. In other words,= a handover of LSP ownership between planes means that Data Plane circuit has to stay untouched in terms of topology and resources allocated, not only in terms of carried traffic. 3. Proposed solution A new flag in the Administrative Status Object (RFC 3471[4] and RFC 3473 [5]), named Handover flag, is proposed in this memo as a mean t= o make possible necessary information exchange among GMPLS enabled nodes, in order to implement and support the functionality introduce= d Caviglia et al. Expires - December 2005 [Page 3] =0C draft-caviglia-mp2cpcp2mp-02 June 20= 05 above. The idea is that standard GRSVP-TE signaling flow can be used= to inform nodes about the ownership handover procedure, where such flow has to be flagged in order to distinguish it from normal LSP setup/release procedure. When a LSP owned by Management Plane has to= be handed over to Control Plane, a signaling set-up with HANDOVER flag set has to be sent from ingress node. At ingress node all the information related to the LSP is passed to GRSVP-TE, which uses it to fill in the PATH message (Explicit Route Object - ERO, Traffic Engineering and context information). Such PATH message, sent with HANDOVER flag set, reaches the nodes along the LSP path and informs them that the referred LSP is already present at Data Plane level an= d that it has to be adopted by Control Plane. After a node has receive= d such special PATH, it becomes owner of the LSP and treat it like any= other LSP set up via GRSVP-TE. Let's call this procedure "MP to CP handover". The opposite procedure "CP to MP handover" works in a similar way. When a LSP owned by Control Plane has to be handed over to Managemen= t Plane, a signaling PATH DOWN with HANDOVER flag set has to be sent from ingress node along the LSP path, informing involved nodes that = a CP to MP handover is in progress for that LSP. The information about that LSP under control within GRSVP-TE scope i= s passed to Management Plane at ingress node, and in every node, which= receives flagged PATH DOWN. A node is able to recognize such special= PATH DOWN by reading the handover flag value. If a node finds that flag set, then it is aware that that the related LSP has to stay untouched at Data Plane level, but its ownership has to be passed to= Management Plane. It is worth stressing that, when the LSP is adopte= d either by CP or MP, i.e. at the end of signaling with Handover flag set, normal CP procedures or MP procedures have to take their place as usual when needed. This means that a LSP formerly owned by MP, signaled within CP with Handover flag set (i.e. handed over to CP) can be deleted by usual relevant Control Plane signaling flows (i.e.= with Handover flag not set). The same applies when considering the handover of a LSP from CP to MP when, at the end of procedure, the LSP belongs to Management Plane and can be fully controlled by NMS. In other words, after the LSP handover procedures have taken place, the LSP is not different from the other LSP owned by handover destination entity and it has to be treated with usual rules for tha= t entity. This is in some way similar to the Restart Procedure, (Section 4.3 RFC 3473 [5]), in the sense that the status of the physical resource= s at Data Plane has to stay unmodified but the associated information allowing its control has to be transferred. The modification propose= d in this document refers only to Administrative Status object, that is, the message flow is left unmodified for both set-up and deletion= . Following section gives detailed description of proposed "MP to CP handover" and "CP to MP handover" procedures. In the following the handover of a bidirectional LSP is assumed. The= case of unidirectional LSP can be easily derived from that. Caviglia et al. Expires - December 2005 [Page 4] =0C draft-caviglia-mp2cpcp2mp-02 June 20= 05 4. LSP Ownership Handover Procedure Between Management And Control Planes Resv Confirm message within LSP setup signaling flow SHOULD be supported in order to manage at best LSP Ownership Handover Procedur= e Between Control And Management Planes. If Resv Confirm is not presen= t in message set used for LSP setup, the handover procedure described below is still applicable using the simple Path/Resv sequence (with handover flag set as detailed in the following). In that case, handover related operations that in the following description are triggered by reception of Resv Confirm, MUST to be executed at the reception of Resv message. 4.1 "MP to CP handover" - LSP Ownership Transfer From Management Plane To Control Plane Let's consider the case of a Data Plane connection between two nodes= acting as ingress and egress for a LSP. Let's assume that Management= Plane has the ownership and control of the LSP and that we want to hand it over to Control Plane. At the ingress node NMS initiates the transfer of LSP related information residing within Management Plane to GRSVP-TE records within Control Plane. We assume that this happens under operator or management application control and in particular that: - Control requests are sent to the ingress LSR by the MP - The MP has some way of knowing when the CP has completed its task or has failed. Ingress node collects from MP all the LSP related information needed= at Control Plane level. The way this operation is done and where suc= h information is collected within MP is outside the scope of this memo= . A relevant part of such information is represented by the LSP path, which has to be handed over to CP to be used by signaling entity to fill the Explicit Route Object (ERO) during setup. In order to support the MP to CP handover of LSP, the ERO object in the Path message MUST be filled with all the LSP relevant informatio= n down to the Label level. That can be done by means of the object an= d procedures defined in [5]. The precise filling of the ERO object is needed as we are assuming that the LSP already exists in the network and that every signaling relevant info about it is available and accessible to MP in terms of= required LSP parameters to build a GRSVP-TE PATH message. After gathering all the LSP related information, the ingress node sends ou= t a GRSVP-TE PATH message including the Administrative Status Object with HANDOVER flag set. Caviglia et al. Expires - December 2005 [Page 5] =0C draft-caviglia-mp2cpcp2mp-02 June 20= 05 Upon reception of such GRSVP-TE PATH, a node MUST be able to understand that a MP to CP handover procedure is in progress by reading Handover flag. Either the ingress node of the LSP (upon request from MP) and intermediate and egress nodes (when receiving a Path message containing an Administrative Status object with the Handover flag set) is informed about the fact that a LSP adoption procedure is requested or ongoing. The node assumes that a Data Plane resource related to the info carried in Path Message is already allocated and= in place. The node SHOULD check however the consistence of the actua= l Data Plane status of such resource: - If the check goes OK, then a GRSVP-TE record for the LSP is created associating it to the corresponding Data Plane state. The node accepts all the LSP information carried in PATH (if the node is not ingress of the LSP, otherwise the information is sent from relevant MP entity) and stores it in Path State Block. After that,= the procedure goes on as described below. - If the check goes NOT OK, that is actual Data Plane state for the indicated resource is different from the one indicated in the Path= message, then: o A PathErr with Path State Removed flag set should be generate= d o GMPLS Control Plane state information about it is not accepte= d by the handover destination entity In both cases, no operation is done over Data Plane. In case of positive check, no change is required at that level since the connection is already set up and in service. In case of negative check, a mismatch or some other error has occurred and no LSP contro= l handover is possible. The procedure rolls back and information transfer process from MP to CP at ingress node of the LSP has to be fixed and reinitiated. A node participating in a MP to CP handover procedure MUST in fact keep track of the special 'handover' conditio= n of the LSP involved, by retaining handover flag status within GRSVP-= TE records. This is important because during handover procedure no other Data Plane, Control Plane or Management Plane action has to be taken on the LSP outside the control of the procedure itself. Such special state regarding the involved LSP has to be retained until the procedure itself has correctly ended. After propagating handover Path, a node MUST wait for a Resv message= including Administrative Status Object with handover flag set. After= receiving it, the actual migration of LSP information is complete. However, Handover flag is cleared in Path/Resv state block of the involved LSP, only by reception of ResvConfirm message (or Resv Caviglia et al. Expires - December 2005 [Page 6] =0C draft-caviglia-mp2cpcp2mp-02 June 20= 05 message, if ResvConfirm is not supported). After the flag is cleared= , the LSP is left completely under control of GRSVP- TE within Control= Plane. This means that any memory about the former MP ownership of the LSP is lost. The following example covers all possible MP to CP Handover scenarios, either in case of success or failure. In the example we assume that Data Plane connection, whose control and control information has to be handed over from MP to CP, is TDM based (eithe= r SDH or SONET). A more generic case, where the Data Plane resource making up the LSP is not tied to a specific technology, can easily b= e derived from this one. The table refers to possible cases when a nod= e at CP level has received LSP information provided by MP and verifies= if handover is feasible. Let's consider a LSP over the network, connecting a ingress node I with an egress node E. Let's call timeslot A and B the Data Plane resources referred by control information involved in Handover in a given node traversed by the LSP. This means that Handover flagged signaling refers to A-B cross-connection over Data Plane. The ingress node initiates the procedure upon request from Managemen= t Plane. The way LSP related information is passed from MP to ingress node is outside the scope of this procedure description. Intermediates nodes and egress node receive the request for LSP adoption and the information needed for the operation from Handover flagged GRSVP-TE signaling. The symbol <----> in table below indicates that the two Timeslots involved in Data Plane cross-connection are actually cross-connected= over Data Plane, hence Data Plane state corresponds to the indicatio= n provided by LSP data held by MP and in the process of being handed over to CP. |Actual Data|Control Plane |Management Plane|Data Plane |Plane State|LSP data record|LSP data record |Resources check Case 1| A<---->B |No info yet |MP expects A-B |OK to MP to CP | | | |LSP handover Case 2| A<---->C |No info yet |MP expects A-B |NOT OK for MP to= | | | |CP LSP handover Case 1: - LSP info from Management Plane to be used for LSP control hand over to GRSVP-TE matches Data Plane state in terms of involved resources - LSP data record is not owned yet by Control Plane, hence LSP control is still up to MP - Checks are OK, so GRSVP-TE state (related to involved LSP) is associated to Data Plane state after Handover flagged signaling flow (Path/Resv/Resv Confirm with Handover flag set) has ended. - At the end of signaling the LSP is completely under CP control. - No actions are taken in the Data Plane. Caviglia et al. Expires - December 2005 [Page 7] =0C draft-caviglia-mp2cpcp2mp-02 June 20= 05 Case 2: - LSP info from Management Plane to be used for LSP control hand over to GRSVP-TE doesn't match Data Plane state in terms of involved resources. - Control Plane does not own LSP data record yet; hence LSP control is still up to MP. - Checks are NOT OK. A-B connection is not actually present over Data Plane and indicated resources are used within other context (A is x-connected to C). - GRSVP-TE state (related to involved LSP) is not associated to the cross connection after Handover flagged Path message. - A PathErr with Path State Removed flag set MUST be sent Upstream. - LSP ownership remains completely under MP control. Handover has failed. - No actions are taken in the Data Plane. 4.2 MP to CP Handover Procedure Failure Handling When a mismatch is detected between LSP information owned by MP (and= going to be handed over to CP) and the actual Data Plane state corresponding to that LSP, actions have to be taken to roll back the= LSP ownership handover correctly. If such mismatch is detected at LSP ingress node, the issue has to b= e resolved directly between ingress node and MP designed entity and this lies outside the scope of this memo. No Control plane signaling= is involved yet at this stage. If the mismatch is detected at intermediate or egress nodes, when th= e LSP control information arrives at the node via handover flagged Pat= h message, the node MUST reject it by issuing PathErr with Path State Removed towards the ingress node. In such a way the procedure is interrupted at that node, upstream nodes are informed and no changes= are done over control and Data Plane. When a node receives PathErr with Path State Removed referred to a LSP, whose data record at CP has handover flag set (being in 'handover state'), it MUST clear suc= h LSP data record and propagate the PathErr message upstream. No Data Plane actions have to be taken in this case as well. The same applies to PathTear message. 4.3 "CP to MP handover" - LSP Ownership Transfer From Control Plane To Management Plane Let's now consider the case of LSP Ownership Transfer From Control Plane To Management Plane. The scenario is still a Data Plane connection between two nodes acting as ingress and egress for a Caviglia et al. Expires - December 2005 [Page 8] =0C draft-caviglia-mp2cpcp2mp-02 June 20= 05 LSP. But let's assume in this case that Control Plane has the ownership and control of the LSP and that we want to hand it over= to Management Plane. This means that at the end of such procedure= , the Data Plane state related to that connection is still untouched, but the LSP related information record is no more owned by GRSVP-TE over Control Plane. In other words, after LSP ownership transfer from CP to MP, the LSP is no more under control of GRSVP-TE, which is no more able to "see"= the LSP itself. This Section covers the procedure needed to manage this procedure as a dual, opposite procedure respect to the one described in previous section. The procedure is performed at a signaling level as described in Section 7.2.1 of the RFC 3473 [5]. At LSP ingress node, relevant MP entity requests the ownership of th= e LSP, How this is done is outside the scope of memo. Ingress node and= MP exchange the relevant information for this task and then propagates it over Control Plane by means of GRSVP-TE tear down signaling flow as detailed below. Ingress node MUST send out a Path Down message, with Handover and Reflect bits in Admin Status set. No action is taken over Data Plane= and Control Plane keeps track of special handover state the LSP is in. Transit and Egress nodes, upon reception of such handover Path Down,= propagate it without any Data Plane action, retaining the handover state information associated to the LSP. After that, every node wait= s until the Handover bit is received back in the Resv. Then a PathTear= is issued and the whole LSP information record is cleared from GRSVP= - TE data structures. In other words, a normal LSP tear down signaling= is exchanged between nodes traversed by the LSP, but handover flag set in Path Down message indicates that no Data Plane action has to correspond to Control Plane signaling. At the end of handover tear down signaling flow, the LSP is released from Control Plane point of= view, but its Data Plane state is still unmodified and it is now owned and controllable by MP. 4.4 CP to MP Handover Procedure Failure Handling Failures during CP to MP handover procedure MUST be managed at signaling level as in normal LSP tear down procedure. The only difference is the handover flag set in Administrative Status Object inside Path Down message which MUST be read by receiving node and imposes that no action has to be made over Data Plane resource whose= corresponding Control Plane record is involved in handover procedure= . Caviglia et al. Expires - December 2005 [Page 9] =0C draft-caviglia-mp2cpcp2mp-02 June 20= 05 5. RSVP Message Formats This memo does not introduce any modification in RSVP messages. 5.1 Object Modification This memo introduces a new flag into the Administrative Status object. The Admin_Status Object is defined in RFC 3473 [5]. This document uses the H-bit of the Admin_Status object. The bit is bit number (TBD by IANA). Handover signaling (H): 1 bit When set, indicates that a Handover procedure for the transfer of LS= P ownership between Management and Control Planes is ongoing . 6. Security Considerations The procedures described in this document rely completely on GRSVP-T= E messages and mechanism. The use of Handover Flag set in Admin Status= Object basically informs the receiving entity that no operations are= to be done over Data Plane as consequence of such special signaling flow. Using specially flagged signaling messages we want to limit th= e function of setup and tear down messages to Control Plane, making them not effective over related Data Plane resource usage. So, no additional or special issues are arisen by adopting this procedure, that aren't already brought up by the use of the same messages, without handover flag setting, for LSP control. For GRSVP-TE Securit= y please refer to [3]. 7. IANA Consideration IANA has been asked to manage the bit allocations for the Administrative Status object [6]. This document requires the allocation of the Handover bit: the H-bit= . IANA is requested to allocate a bit for this purpose. Caviglia et al. Expires - December 2005 [Page 10] =0C draft-caviglia-mp2cpcp2mp-02 June 20= 05 8. References [1] Bradner, S., "The Internet Standards Process -- Revision 3", BCP= 9, RFC 2026, October 1996. [2] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997 [3] Crocker, D. and Overell, P.(Editors), "Augmented BNF for Syntax Specifications: ABNF", RFC 2234, Internet Mail Consortium and Demon Internet Ltd., November 1997 [4] L. Berger (Ed.) =E6=E6Generalized Multi-Protocol Label Switching= (GMPLS) Signaling Functional Description=C6=C6, RFC 3471, January 20= 03 [5] L. Berger (Ed.) =E6=E6Generalized Multi-Protocol Label Switching= (GMPLS) Signaling Resource ReserVation Protocol-Traffic Engineering (RSVP-TE) Extensions=C6=C6, RFC 3473, January 2003 [6] Zamfir, A., " Component Link Recording and Resource Control for GMPLS Link Bundles", draft-zamfir-explicit-resource-control-bundle- 03, February 2004 - work in progress. [7] L. Berger (Ed.) "GMPLS - Communication of Alarm Information", draft-ietf-ccamp-gmpls-alarm-spec-02.txt, November 2004, work in progress. 9. Acknowledgments Adrian Farrel provided editorial assistance to prepare this draft fo= r publication. Author's Addresses Diego Caviglia Marconi Via A. Negrone 1/A Genova-Sestri Ponente, Italy Phone: +390106003738 Email: diego.caviglia@marconi.com Dino Bramanti Marconi Via Moruzzi 1 Caviglia et al. Expires - December 2005 [Page 11] =0C draft-caviglia-mp2cpcp2mp-02 June 20= 05 C/O Area Ricerca CNR Pisa, Italy Email: dino.bramanti@marconi.com Nicola Ciulli NextWorks Corso Italia 116 56125 Pisa, Italy Email: n.ciulli@nextworks.it Intellectual Property Rights Notices The IETF takes no position regarding the validity or scope of any Intellectual Property Rights or other rights that might be claimed t= o pertain to the implementation or use of the technology described in this document or the extent to which any license under such rights might or might not be available; nor does it represent that it has made any independent effort to identify any such rights. Information= on the procedures with respect to rights in RFC documents can be found in BCP 78 and BCP 79. Copies of IPR disclosures made to the IETF Secretariat and any assurances of licenses to be made available, or the result of an attempt made to obtain a general license or permission for the use o= f such proprietary rights by implementers or users of this specification can be obtained from the IETF on-line IPR repository a= t http://www.ietf.org/ipr. The IETF invites any interested party to bring to its attention any copyrights, patents or patent applications, or other proprietary rights that may cover technology that may be required to implement this standard. Please address the information to the IETF at ietf- ipr@ietf.org. Full Copyright Statement Copyright (C) The Internet Society (2005). This document is subject to the rights, licenses and restrictions contained in BCP 78, and except as set forth therein, the authors retain all their rights. This document and the information contained herein are provided on a= n "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENT= S OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Caviglia et al. Expires - December 2005 [Page 12] =0C draft-caviglia-mp2cpcp2mp-02 June 20= 05 Caviglia et al. Expires - December 2005 [Page 13] = Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 09:37:14 +0000 Sensitivity: Subject: RE: Moving LSP ownership between control plane and management plane To: "Dimitri.Papadimitriou" <Dimitri.Papadimitriou@alcatel.be> Cc: "ccamp" <ccamp@ops.ietf.org>, "gnewsome" <gnewsome@ieee.org>, "adrian" <adrian@olddog.co.uk> Message-ID: <OFFB8978C4.F69E790A-ONC1257044.00348866-C1257044.0034D8C7@uk.marconicomms.com> From: "Diego Caviglia" <Diego.Caviglia@marconi.com> Date: Wed, 20 Jul 2005 11:37:09 +0200 MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii Dimitri, yes of course the main application is MP-->CP the CP-->MP is just for simmetry;-)) Diego Dimitri.Papadimitriou@alcatel.be on 20/07/2005 11.30.35 To: "Diego Caviglia" <Diego.Caviglia@marconi.com> cc: ccamp@ops.ietf.org, gnewsome@ieee.org, adrian@olddog.co.uk Subject: RE: Moving LSP ownership between control plane and management plane diego - "My basic question is: do someone think that having a mechanism to move MP/CP created LSP under the CP/MP domain is useless?" no - in part. the MP->CP handover is a key mechanism for migration toward GMPLS CP -dimitri. Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 09:31:58 +0000 Sensitivity: Subject: Re: Moving LSP ownership between control plane and management plane To: "George Newsome <gnewsome" <gnewsome@ieee.org> Cc: "Adrian Farrel <adrian" <adrian@olddog.co.uk>, "ccamp" <ccamp@ops.ietf.org>, "Neil Harrison <neil.2.harrison" <neil.2.harrison@bt.com>, "Dino Bramanti <Dino.Bramanti" <Dino.Bramanti@marconi.com> Message-ID: <OF51FD6EC8.C30CE76C-ONC1257044.0033C9C8-C1257044.003461D7@uk.marconicomms.com> From: "Diego Caviglia" <Diego.Caviglia@marconi.com> Date: Wed, 20 Jul 2005 11:32:04 +0200 MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii Hi Geroge, sorry but I don't see how the management effort is something related to this ID and more in general with CCAMP. There are already RFC ID that provide tools to signal LSP with ERO filled up to the Label level and of course this can be done only via management, I don't remeber any objections on that due to NMS impact. I'd like to discuss here (I mean in the CCAMP ML) about the ID, that is, if it useful and if could work and not about the difficulties it implies for the NMS. I'm open to discuss, and actually I did, NMS related issues on the ITU-T ML. Regards Diego George Newsome <gnewsome@ieee.org>@ops.ietf.org on 20/07/2005 11.03.00 Sent by: owner-ccamp@ops.ietf.org To: Adrian Farrel <adrian@olddog.co.uk> cc: ccamp@ops.ietf.org, Neil Harrison <neil.2.harrison@bt.com>, Dino Bramanti <Dino.Bramanti@marconi.com> Subject: Re: Moving LSP ownership between control plane and management plane Adrian, Dino, etc., This thread has rapidly got too long, and I don't have the time at the moment to argue the toss, so I'll reply as a note rather than intersperse remarks. On Adrian's question of which operators were at Holmdel - AT&T and T-Systems. On do operators think they want this - some do some don't. BT clearly don't see much value in fiddling with signaling. You ask whether this might be done in service, and there are certainly those who would like to see just that. Related questions are how many times in the life of the system might this be done for the same connection? Is it cost effective to develop a highly reliable solution that is used just once in the life of the system? However this is not so much a discussion of whether this could be useful, but a technical discussion of whether changing a bit (and a slew of procedures) in a signaling protocol actually provides the claimed facility. Because of the effort the management system needs to provide before signaling might be involved, changing the signaling protocol gives the illusion of success, yet does not provide the claimed service at all. Once the management systems have got as far as assigning control plane names, we can ask what happens to the cross connect object when it is deleted as part of the hand off procedure? The behavior of this object is that its deletion results in dropping the matrix connection, which is exactly not required during the change over. It is my considered opinion that fiddling with signaling merely obfuscates the significant difficulties, which are glossed over by stating that those difficulties occur at a different time. Regards George Adrian Farrel wrote: > Thanks George, > > Useful input. > > >>You asked for comments on this proposal, which was presented to ITU-T >>Q14 experts in January of this year. > > > Can you confirm that there were plenty of providers represented at the > meeting. > > >>There are so many issues involved in moving a connection from the >>management plane to the control plane that involve intensive management >>action, that it is not evident that signaling need be involved at all. > > > Interesting. It is certainly valuable to have a list of these issues out > in the open. I have made some comments in the appended meeting report. > > If the control plane is to take over the connection then we assume that > signaling state must be built so that the control plane could then modify > or tear down the connection. how would you propose that this be done > without involving signaling? > > >>As a personal opinion, it makes little sense to consider moving the >>connection. It makes more sense to replace the connection with one set >>up via the control plane, and to then delete the original management set >>up connection. > > > The question arrises of how this would be done in a network with limited > resources. That is, if it is possible to set up a parallel conneciton then > the method you describe would be fine, but if there are not enough > parallel resources then would you propose that a lambda (for example) > could be used for the management-established connection and for the > control-plane-owned connection at the same time? This might present some > difficulties as well. > > Do you have an opinion on my main question which is whether this tyoe of > change-over would ever be done in service? > > Thanks, > Adrian > > >>Excerpt of the meeting report of Q14 Experts, January 2005 >> >>WD06 (Marconi) "G.7713 Modification in order to support Circuit >>Migration" contains a modified version of draft G.7713 Revision as was >>presented in the Berlin meeting (WD 11, i.e., TD50/3 Dec.2004). This >>contribution addresses SP24 (PC ? SPC) of G.7713. It proposes to extend >>the concept of Call/Connection setup to setup/adopt and release to >>release/de-adopt. In Connection Adoption both the SNC are not actually >>created and the LC are not actually established due to the fact that the >>underlying physical resources are already in place. In the >>Call/Connection dis-adoption SNC-LC-SNC is not actually de-allocated, >>only the Control Plane information associated with then is removed and >>the ownership is moved to Management Plane. Attributes for indicating >>Adoption/Dis-adoption are proposed for the INNI Connection messages. >> >>The discussion was mainly on the Management plane and Control plane >>actions required before signalling is involved. The group noted that the >>following issues need to be addressed: >> >>- Naming of transport resources to the control plane. Before a call can >>be placed under signalling control, links that are involved in the >>connection need to be given control plane names. Without these, no >>explicit route can be formed to match the resources of the connection. > > > This is pretty obvious. But actually, this is an issue with moving the > resources from one plane to another (or having them present in both > planes). It is certainly a prerequisite for moving the connection, but it > is not functionally an element of moving the connection. > > The draft, as I read it, assumes that these operations have been completed > and then asks "how do I move the connection from one plane to the other?" > > >>- Resource state while under dual CP and MP views. After resources are >>given control plane names, the resources are still viewed by the >>management plane. It may be necessary to create a new state for the >>resource to indicate that the management plane cannot perform some >>actions on the connection points as the control plane is about to take > > over. > > Yes. There is certainly scope for a race condition here. I believe that > the I-D proposes to (try to) make the transfer an atomic action from the > point of view of the change-over. > > Note, however, that it is wrong to assume that a resource belongs only to > the control or to the management plane at any moment in time. This may be > true with respect to the ability to provision/control connections, but > resources that are available for use by the control plane can still be > managed through the management plane. Thus, you take a laser out of > service through the management plane, not through the control plane. > > >>- Role of discovery functions (esp. CELA). After control plane names >>are allocated, distributed control plane functions need to be associated >>and communication adjacencies formed. This too is a precursor to any >>signalling procedures in migrating from the MP to CP. > > > Note to CCAMP: CELA is Control Entity Logical Adjacency. > This too is "obvious". > But this discussion seems to be wide of the mark. Before deploying a > control plane, it is clear that this type of operation needs to take > place. Since (within the context of CCAMP) a control plane *is* going to > be deployed, we can assume this operation and move on to the discussion in > the draft. > > >>- Similarity of migration to synchronization after CP failure and >>subsequent recovery. Connection control procedures might be the same in >>migrating a call from the MP to the CP as the situation when the CP has >>failed and is recovering. Here, the network connections are already in >>place, but connection and call state need to be created to match it. >>Knowledge of the call and what the connection should be is obtained from >>the MP for migration, and from some reliable database in the recovery > > case. > > This is a valid point. > Control plane failure, however, is unlikely to be total. That is, it would > be unusual for more than one control plane link or node to fail. Thus, in > the control plane, the "reliable database" is distributed and recovery can > follow the procedures outlined in draft-ietf-ccamp-rsvp-restart-ext-03.txt > > It would certainly be worth considering whether the procedures used for > control plane recovery could be applied to transfer of ownership from the > management to control plane. at the moment, I suspect that they could not > because the restart extensions assume a specific period during which > signaling is associated with recovery. > > We would also need to look at the reverse process (transfer from CP to MP) > that is not possible to cover by CP restart procedures. > > >>- Call awareness of migration vs. connection being unaware of migration. >> When connection state is being created to match an existing >>connection, the connection controllers (CCs) do not require awareness of >>why this is happening as the context could be migration or recovery. A >>mechanism to create control state without affecting transport plane >>state is needed in the CCs. Call controllers though do need migration >>awareness as they need to obtain/derive call identification from the MP. > > > I disagree. > Connection controllers must be able to distinguish between resources that > are already in use and cannot be assigned to a new connection, and those > that are already in use and can be assigned to a new connection because > the connection is replacing an existing connection. If this were happening > entirely within the control plane we would use resource sharing, but when > the operation spans the control and management planes resource > ownership/sharing is more complex and the conneciotn controller needs to > be able to issue the approriate instrucitons for the assignment of the > resources. > On the other hand, call controllers do not need awareness of the details > of the realisation of the connections. > > >>In summary, the group decided that it is premature to consider >>signalling procedure before the above issues, amongst other, have been >>studied. However the contribution does enable the group to expose in a >>larger context the interaction between CP and MP for connection >>migration. The above issues were included in SP24 of the G.7713 Living >>List (see WD22). > > > Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 09:31:22 +0000 To: "Diego Caviglia" <Diego.Caviglia@marconi.com> Cc: ccamp@ops.ietf.org, gnewsome@ieee.org, adrian@olddog.co.uk MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: RE: Moving LSP ownership between control plane and management plane Date: Wed, 20 Jul 2005 11:30:35 +0200 Message-ID: <OFF7D70928.9E607288-ONC1257044.00343C93-C1257044.00343D21@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmRpZWdvIC08QlI+PEJSPjxCUj4mcXVv dDtNeSBiYXNpYyBxdWVzdGlvbiBpczogZG8gc29tZW9uZSB0aGluayB0aGF0IGhhdmluZyBhIG1l Y2hhbmlzbSB0byBtb3ZlPEJSPk1QL0NQIGNyZWF0ZWQgTFNQIHVuZGVyIHRoZSBDUC9NUCBkb21h aW4gaXMgdXNlbGVzcz8mcXVvdDs8L0ZPTlQ+PC9QPjxQPiZuYnNwOzwvUD48UD48Rk9OVCBGQUNF PSJNb25vc3BhY2UsQ291cmllciI+bm8gLSA8L0ZPTlQ+PC9QPjxQPiZuYnNwOzwvUD48UD48Rk9O VCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+aW4gcGFydC4gdGhlIE1QLSZndDtDUCBoYW5kb3Zl ciBpcyBhIGtleSBtZWNoYW5pc20gZm9yIG1pZ3JhdGlvbiB0b3dhcmQgR01QTFMgQ1A8L0ZPTlQ+ PC9QPjxQPiZuYnNwOzwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+LWRpbWl0 cmkuPEJSPjwvRk9OVD48QlI+PEJSPjxCUj48QlI+PEJSPjwvUD4= Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 09:04:12 +0000 Message-ID: <42DE1344.7060403@ieee.org> Date: Wed, 20 Jul 2005 05:03:00 -0400 From: George Newsome <gnewsome@ieee.org> User-Agent: Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.4) Gecko/20030624 Netscape/7.1 (ax) MIME-Version: 1.0 To: Adrian Farrel <adrian@olddog.co.uk> Cc: ccamp@ops.ietf.org, Neil Harrison <neil.2.harrison@bt.com>, Dino Bramanti <Dino.Bramanti@marconi.com> Subject: Re: Moving LSP ownership between control plane and management plane Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Adrian, Dino, etc., This thread has rapidly got too long, and I don't have the time at the moment to argue the toss, so I'll reply as a note rather than intersperse remarks. On Adrian's question of which operators were at Holmdel - AT&T and T-Systems. On do operators think they want this - some do some don't. BT clearly don't see much value in fiddling with signaling. You ask whether this might be done in service, and there are certainly those who would like to see just that. Related questions are how many times in the life of the system might this be done for the same connection? Is it cost effective to develop a highly reliable solution that is used just once in the life of the system? However this is not so much a discussion of whether this could be useful, but a technical discussion of whether changing a bit (and a slew of procedures) in a signaling protocol actually provides the claimed facility. Because of the effort the management system needs to provide before signaling might be involved, changing the signaling protocol gives the illusion of success, yet does not provide the claimed service at all. Once the management systems have got as far as assigning control plane names, we can ask what happens to the cross connect object when it is deleted as part of the hand off procedure? The behavior of this object is that its deletion results in dropping the matrix connection, which is exactly not required during the change over. It is my considered opinion that fiddling with signaling merely obfuscates the significant difficulties, which are glossed over by stating that those difficulties occur at a different time. Regards George Adrian Farrel wrote: > Thanks George, > > Useful input. > > >>You asked for comments on this proposal, which was presented to ITU-T >>Q14 experts in January of this year. > > > Can you confirm that there were plenty of providers represented at the > meeting. > > >>There are so many issues involved in moving a connection from the >>management plane to the control plane that involve intensive management >>action, that it is not evident that signaling need be involved at all. > > > Interesting. It is certainly valuable to have a list of these issues out > in the open. I have made some comments in the appended meeting report. > > If the control plane is to take over the connection then we assume that > signaling state must be built so that the control plane could then modify > or tear down the connection. how would you propose that this be done > without involving signaling? > > >>As a personal opinion, it makes little sense to consider moving the >>connection. It makes more sense to replace the connection with one set >>up via the control plane, and to then delete the original management set >>up connection. > > > The question arrises of how this would be done in a network with limited > resources. That is, if it is possible to set up a parallel conneciton then > the method you describe would be fine, but if there are not enough > parallel resources then would you propose that a lambda (for example) > could be used for the management-established connection and for the > control-plane-owned connection at the same time? This might present some > difficulties as well. > > Do you have an opinion on my main question which is whether this tyoe of > change-over would ever be done in service? > > Thanks, > Adrian > > >>Excerpt of the meeting report of Q14 Experts, January 2005 >> >>WD06 (Marconi) "G.7713 Modification in order to support Circuit >>Migration" contains a modified version of draft G.7713 Revision as was >>presented in the Berlin meeting (WD 11, i.e., TD50/3 Dec.2004). This >>contribution addresses SP24 (PC ? SPC) of G.7713. It proposes to extend >>the concept of Call/Connection setup to setup/adopt and release to >>release/de-adopt. In Connection Adoption both the SNC are not actually >>created and the LC are not actually established due to the fact that the >>underlying physical resources are already in place. In the >>Call/Connection dis-adoption SNC-LC-SNC is not actually de-allocated, >>only the Control Plane information associated with then is removed and >>the ownership is moved to Management Plane. Attributes for indicating >>Adoption/Dis-adoption are proposed for the INNI Connection messages. >> >>The discussion was mainly on the Management plane and Control plane >>actions required before signalling is involved. The group noted that the >>following issues need to be addressed: >> >>- Naming of transport resources to the control plane. Before a call can >>be placed under signalling control, links that are involved in the >>connection need to be given control plane names. Without these, no >>explicit route can be formed to match the resources of the connection. > > > This is pretty obvious. But actually, this is an issue with moving the > resources from one plane to another (or having them present in both > planes). It is certainly a prerequisite for moving the connection, but it > is not functionally an element of moving the connection. > > The draft, as I read it, assumes that these operations have been completed > and then asks "how do I move the connection from one plane to the other?" > > >>- Resource state while under dual CP and MP views. After resources are >>given control plane names, the resources are still viewed by the >>management plane. It may be necessary to create a new state for the >>resource to indicate that the management plane cannot perform some >>actions on the connection points as the control plane is about to take > > over. > > Yes. There is certainly scope for a race condition here. I believe that > the I-D proposes to (try to) make the transfer an atomic action from the > point of view of the change-over. > > Note, however, that it is wrong to assume that a resource belongs only to > the control or to the management plane at any moment in time. This may be > true with respect to the ability to provision/control connections, but > resources that are available for use by the control plane can still be > managed through the management plane. Thus, you take a laser out of > service through the management plane, not through the control plane. > > >>- Role of discovery functions (esp. CELA). After control plane names >>are allocated, distributed control plane functions need to be associated >>and communication adjacencies formed. This too is a precursor to any >>signalling procedures in migrating from the MP to CP. > > > Note to CCAMP: CELA is Control Entity Logical Adjacency. > This too is "obvious". > But this discussion seems to be wide of the mark. Before deploying a > control plane, it is clear that this type of operation needs to take > place. Since (within the context of CCAMP) a control plane *is* going to > be deployed, we can assume this operation and move on to the discussion in > the draft. > > >>- Similarity of migration to synchronization after CP failure and >>subsequent recovery. Connection control procedures might be the same in >>migrating a call from the MP to the CP as the situation when the CP has >>failed and is recovering. Here, the network connections are already in >>place, but connection and call state need to be created to match it. >>Knowledge of the call and what the connection should be is obtained from >>the MP for migration, and from some reliable database in the recovery > > case. > > This is a valid point. > Control plane failure, however, is unlikely to be total. That is, it would > be unusual for more than one control plane link or node to fail. Thus, in > the control plane, the "reliable database" is distributed and recovery can > follow the procedures outlined in draft-ietf-ccamp-rsvp-restart-ext-03.txt > > It would certainly be worth considering whether the procedures used for > control plane recovery could be applied to transfer of ownership from the > management to control plane. at the moment, I suspect that they could not > because the restart extensions assume a specific period during which > signaling is associated with recovery. > > We would also need to look at the reverse process (transfer from CP to MP) > that is not possible to cover by CP restart procedures. > > >>- Call awareness of migration vs. connection being unaware of migration. >> When connection state is being created to match an existing >>connection, the connection controllers (CCs) do not require awareness of >>why this is happening as the context could be migration or recovery. A >>mechanism to create control state without affecting transport plane >>state is needed in the CCs. Call controllers though do need migration >>awareness as they need to obtain/derive call identification from the MP. > > > I disagree. > Connection controllers must be able to distinguish between resources that > are already in use and cannot be assigned to a new connection, and those > that are already in use and can be assigned to a new connection because > the connection is replacing an existing connection. If this were happening > entirely within the control plane we would use resource sharing, but when > the operation spans the control and management planes resource > ownership/sharing is more complex and the conneciotn controller needs to > be able to issue the approriate instrucitons for the assignment of the > resources. > On the other hand, call controllers do not need awareness of the details > of the realisation of the connections. > > >>In summary, the group decided that it is premature to consider >>signalling procedure before the above issues, amongst other, have been >>studied. However the contribution does enable the group to expose in a >>larger context the interaction between CP and MP for connection >>migration. The above issues were included in SP24 of the G.7713 Living >>List (see WD22). > > > Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 08:56:52 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: RE: Moving LSP ownership between control plane and management plane Date: Wed, 20 Jul 2005 09:56:04 +0100 Message-ID: <0536FC9B908BEC4597EE721BE6A353890A9F1E04@i2km07-ukbr.domain1.systemhost.net> Thread-Topic: Moving LSP ownership between control plane and management plane Thread-Index: AcWM/9UjPIT0x57mRMyOU2Wlm2MZtQABrR8A From: <neil.2.harrison@bt.com> To: <Diego.Caviglia@marconi.com> Cc: <gnewsome@ieee.org>, <adrian@olddog.co.uk>, <ccamp@ops.ietf.org> Hi Diego.....No I don't think what you are saying is useless, and I'm sorry if that was the impression I conveyed .......in fact you MUST consider these aspects (eg the ones George flagged) when you add a CP. What I was trying to point out is the IMPORTANCE of a CP here relative to other stuff. Let me just play out the logic again as briefly as I can. As we move towards the duct the importance/value of the control-plane diminishes. So for addressing S-PVCs and resilience issues its fine......where I have grave doubts are those who think they can create 'topology on the fly' (for other client layer networks). A key factor here is acceptable GoS/business models....our studies indicate that to be viable one would need a waiting time (post demand epoch) of at least the same order as the expected holding time.....now figure out what that really means. And of course this still assumes everything BELOW this SVC level is already provisioned. =20 Then there is the issue of 'who owns which layer?'.....and this means the overriding requirement that must be addressed 1st is client/server functional decoupling. Closely related to this 'who owns which layer' issue is how the 3 planes of one layer network are provided over the *traffic plane* of another/lower layer network....and I gave a very detailed response to Igor on this recently. And once you piece it all together then you realise we can't have full convergence from IP to Optics....and a smart observer at this point might ask 'so why are folks being prescriptive about the choice of functional components, eg addressing, signalling, routing?' For example I could postulate that PNNI is a better signalling protocol than RSVP ...and why would anyone in their right minds start with v4 addressing when they already know its problems? Should at least use v6 here IMO. Maybe I only need static routing and this should be a centralised function? =20 And finally...even when you add the CP it is fact of life, with which I see you agree, that the MP is still king here.....a CP without a good MP is rather useless. regards, neil > -----Original Message----- > From: Diego Caviglia [mailto:Diego.Caviglia@marconi.com] > Sent: 20 July 2005 08:52 > To: Harrison,N,Neil,IKM1 R > Cc: <gnewsome; <adrian; <ccamp > Subject: RE: Moving LSP ownership between control plane and=20 > management plane >=20 >=20 >=20 > Hi Neil, > it always a pleasure read your e-mail but I think > I'm missing something. >=20 > The basic purpose of the ID is to provide a simple mechanism > to move LSPs from a domain to another, I don't see any=20 > implication, with layering. Better to say this ID have the=20 > same implication with layering as 'normal' signalling. >=20 > Moreover I agree with you that NMS is the 'king' of the > network and in fact the ID forseen that is MNS to provide the=20 > right information to CP to move the LSP from a domain to another one. >=20 > My basic question is: do someone think that having a > mechanism to move MP/CP created LSP under the CP/MP domain is useless? >=20 > My experience tell me that the answer is no, but I'm open to > discuss this. >=20 > Regards >=20 > Diego >=20 >=20 >=20 > <neil.2.harrison@bt.com>@ops.ietf.org on 19/07/2005 23.36.27 >=20 > Sent by: owner-ccamp@ops.ietf.org >=20 >=20 > To: <gnewsome@ieee.org>, <adrian@olddog.co.uk> > cc: <ccamp@ops.ietf.org> >=20 > Subject: RE: Moving LSP ownership between control plane=20 > and management > plane >=20 > Hi George....nice to hear from you...hope all is well with you, >=20 > This all sounds very reasonable to me. At the end of the > day, all a signalling protocol is is a distributed agent of=20 > network management.....in the types of large BW granularity=20 > networks being > discussed the NMS will always be king. Sorry if that's a problem to > some folks, its just a fact for the way we see/implement it. >=20 > This also follows from a good understanding of the > commercials at play here. That is, there is a major=20 > difference between: > - a public service context (cf PSTN), or extending=20 > a similar > capability out to end-systems in the co-ps mode say, where=20 > end-systems/applications drive trail set-up (ie SVCs across a=20 > UNI). Here one hopes there could be a large population of users (think > peer-peer/BB) and that the trail holding-times are not 'too=20 > long'. In this environment there could be a viable business=20 > case based on a providing a given GoS at some cost.....and it=20 > might solve some > (otherwise) hard QoS/pricing problems for operators....but I digress; > - a private network builder service context where one is a > providing a (relatively long time-constant)=20 > traffic-engineering function to a higher layer network, ie=20 > creating its topology. This is a client/server relationship=20 > and THE starting point for this must be functional decoupling=20 > of the client and server layer networks because they will=20 > often be owned by different parties. We are talking=20 > relatively long forecast/planning cycles here and long trail=20 > holding times....esp when it comes to real physical build. =20 > Further, the potential population is rather small compared to=20 > the previous case. If folks look at whether there is a=20 > viable business case here based on a providing a given GoS at=20 > some cost in an acceptable time period I suspect they will be=20 > disappointed. >=20 > What this all means is that the control-plane is not such an > important topic as we move down the layered network stack=20 > towards the duct....increasingly other factors like strong=20 > NMS and OAM assume a far greater importance. This is not to=20 > say a control-plane is not useful > but: > - we should be clear there cannot be any common=20 > addressing and > routing functions across a set of nested layer networks,=20 > which is very obvious when different parties own different=20 > layer networks....one can (and should) re-use components (if=20 > they are the right ones), but they will need to operate in=20 > different layer network spaces > - signalling is a useful agent for functions like=20 > S-PVCs and > resilience. >=20 > Well that's our analysis anyway....if anyone knows different > I'd be happy to review their commercial assumptions with folks here. >=20 >=20 > regards, Neil >=20 > > -----Original Message----- > > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On > > Behalf Of George Newsome > > Sent: 19 July 2005 12:44 > > To: Adrian Farrel > > Cc: ccamp@ops.ietf.org > > Subject: Re: Moving LSP ownership between control plane and=20 > management > > plane > > > > > > Adrian, > > > > You asked for comments on this proposal, which was > presented to ITU-T > > Q14 experts in January of this year. > > > > There are so many issues involved in moving a connection from the > > management plane to the control plane that involve intensive=20 > > management action, that it is not evident that signaling need be=20 > > involved at all. > > > > As a personal opinion, it makes little sense to consider moving the > > connection. It makes more sense to replace the connection=20 > with one set > > up via the control plane, and to then delete the original > management > > set up connection. > > > > Appended is an excerpt of the meeting report detailing some of the > > issues that need to be resolved before signaling need be considered. > > > > Regards > > > > George > > > > Excerpt of the meeting report of Q14 Experts, January 2005 > > > > WD06 (Marconi) "G.7713 Modification in order to support Circuit > > Migration" contains a modified version of draft G.7713=20 > Revision as was > > presented in the Berlin meeting (WD 11, i.e., TD50/3 > Dec.2004). This > > contribution addresses SP24 (PC ? SPC) of G.7713. It proposes to > > extend the concept of Call/Connection setup to setup/adopt=20 > and release > > to release/de-adopt. In Connection Adoption both the SNC are not=20 > > actually created and the LC are not actually established due to the > > fact that the > > underlying physical resources are already in place. In the > > Call/Connection dis-adoption SNC-LC-SNC is not actually=20 > de-allocated, > > only the Control Plane information associated with then is removed=20 > > and the ownership is moved to Management Plane. Attributes for > indicating > > Adoption/Dis-adoption are proposed for the INNI Connection messages. > > > > The discussion was mainly on the Management plane and Control plane > > actions required before signalling is involved. The group=20 > noted that > > the following issues need to be addressed: > > > > - Naming of transport resources to the control plane. > Before a call > > can be placed under signalling control, links that are > involved in the > > connection need to be given control plane names. Without these, no > > explicit route can be formed to match the resources of the=20 > connection. > > > > - Resource state while under dual CP and MP views. After resources > > are given control plane names, the resources are still=20 > viewed by the > > management plane. It may be necessary to create a new > state for the > > resource to indicate that the management plane cannot perform some > > actions on the connection points as the control plane is=20 > about to take > > over. > > > > - Role of discovery functions (esp. CELA). After control > plane names > > are allocated, distributed control plane functions need to be > > associated and communication adjacencies formed. This too is a=20 > > precursor to any signalling procedures in migrating from=20 > the MP to CP. > > > > - Similarity of migration to synchronization after CP failure and > > subsequent recovery. Connection control procedures might=20 > be the same > > in migrating a call from the MP to the CP as the situation > when the CP > > has failed and is recovering. Here, the network connections are=20 > > already in place, but connection and call state need to be created=20 > > to match it. Knowledge of the call and what the connection should be > > is obtained from > > the MP for migration, and from some reliable database in the > > recovery case. > > > > - Call awareness of migration vs. connection being unaware of > > migration. > > When connection state is being created to match an existing=20 > > connection, the connection controllers (CCs) do not require=20 > awareness > > of why this is happening as the context could be migration or=20 > > recovery. A mechanism to create control state without affecting=20 > > transport plane state is needed in the CCs. Call controllers though > > do need migration > > awareness as they need to obtain/derive call identification > > from the MP. > > > > In summary, the group decided that it is premature to consider > > signalling procedure before the above issues, amongst=20 > other, have been > > studied. However the contribution does enable the group to > expose in a > > larger context the interaction between CP and MP for connection=20 > > migration. The above issues were included in SP24 of the G.7713=20 > > Living List (see WD22). > > > > > > > > >=20 >=20 >=20 >=20 >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 08:47:44 +0000 To: Arthi Ayyangar <arthi@juniper.net> Cc: Igor Bryskin <ibryskin@movaz.com>, ccamp@ops.ietf.org MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: Re: LSP stitching questions for discussion Date: Wed, 20 Jul 2005 10:46:49 +0200 Message-ID: <OF75709CD6.6639BC60-ONC1257044.00303A8B-C1257044.00303B54@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmFydGhpLCA8L0ZPTlQ+PC9QPjxQPiZu YnNwOzwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+Zmlyc3QgYSByZWNhcDog dGhpcyBpdGVtIG9uIHN0aXRjaGluZyB3YXMgaW5pdGlhdGVkIGluIG9yZGVyIHRvIGFsbG93IHBl ciAmcXVvdDtkb21haW4mcXVvdDsgY29udHJvbCBvZiBlbmQtdG8tZW5kIChjb250aWd1b3VzKSBM U1AsIGFzIHBhcnQgb2YgYSBzZXBhcmF0ZSBkb2N1bWVudCwgYW5kIG5vdCB0byBleHRlbmQgaXRz IHNjb3BlICh1bmR1bHksIGluIHRoaXMgY2FzZSwgc2VlIGJlbG93KSBidXQgbWFpbmx5IGluIG9y ZGVyIHRvIGhhdmUgdGhlIHNldCBvZiBhbGwgcmVsYXRlZCBwcm90b2NvbCBtZWNoYW5pc21zIHdp dGhpbiBhIHNwZWNpZmljIGRvY3VtZW50IC08L0ZPTlQ+PC9QPjxQPiZuYnNwOzwvUD48UD48Rk9O VCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+ZnJvbSB0aGlzIGluaXRpYWwgd29ya2luZyBhc3N1 bXB0aW9uLCBpdCBpcyBjbGVhciB0aGF0IHRoZSBzZWdtZW50IG11c3QgaGF2ZSB0aGUgcHJvcGVy dGllcyBvZiB0aGUgaW5jb21pbmcgZW5kLXRvLWVuZCBMU1Agd2hpY2ggYXJlIGRldGVybWluZWQg YW1vbmcgb3RoZXIgYnkgaXRzIHN3aXRjaGluZyB0eXBlIGUuZy4gUFNDLTMgb3IgVERNLCB0aGUg Zmlyc3QgcmVxdWlyZW1lbnQgb2YgTFNQIHN0aXRjaGluZyBpcyBjb250aWd1aXR5IG9mIHRoZSBl bmQtdG8tZW5kIExTUCBhdCB0aGUgZGF0YSBwbGFuZSBsZXZlbCAoYWZ0ZXIgc3RpdGNoaW5nIG9w ZXJhdGlvbikgLSBhbmQgdGhpcyBieSBkZWZpbml0aW9uIG9mIHRoZSBpbml0aWFsIHdvcmtpbmcg YXNzdW1wdGlvbiAtPC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+ bm90ZTogdGhlIG9ubHkgbGF0aXR1ZGUgd2hpY2ggaXMgbGVmdCAoZm9yIGEgZ2l2ZW4gc3dpdGNo aW5nIHR5cGUgZS5nLiBQU0MtMikgaXMgZHVlIHRvIHRoZSBwcm9wZXJ0aWVzIG9mIHBhY2tldCAo UFNDKSBuZXR3b3JrcyB0aGF0IGFsbG93cyB5b3Ugc3RpdGNoIGEgc2VnbWVudCBoYXZpbmcgYSBs YXJnZXIgTWF4IFJlc2VydmFibGUgQmFuZHdpZHRoIGFuZCBVbnJlc2VydmVkIEJhbmR3aWR0aCAo d2Ugd291bGQgaGF2ZSBlcXVpdmFsZW50IGxhdGl0dWRlIGZvciBvdGhlciBhdHRyaWJ1dGVzIGJ1 dCB0aGlzIGlzIGEgbG9jYWwgcG9saWN5IGRlY2lzaW9uIC0gd2hpY2ggYXQgdGhlIGVuZCBpcyB0 aGUgc29sZSBwdXJwb3NlIG9mIHRoaXMgYXBwcm9hY2gpPC9GT05UPjwvUD48UD48Rk9OVCBGQUNF PSJNb25vc3BhY2UsQ291cmllciI+PEJSPjxCUj5IZXJlIGFyZSBzb21lIG9mIHRoZSBxdWVzdGlv bnMgcmVsYXRlZCB0byBMU1Agc3RpdGNoaW5nIHRvIHN0YXJ0IGE8QlI+ZGlzY3Vzc2lvbiBvbiB0 aGUgdXNhZ2Ugb2YgTFNQIHN0aXRjaGluZy48QlI+PEJSPiZndDsxKSBBcmUgdGhlIGNvbnRyb2wg cGxhbmUgc2lnbmFsaW5nIHByb2NlZHVyZXMgZm9yIExTUCBzdGl0Y2hpbmcgYXM8QlI+Jmd0O2Rl c2NyaWJlZCBpbiB0aGlzIElELCAoZXhwbGljaXQgcmVxdWVzdCBmb3IgTFNQIHN0aXRjaGluZyBv bjxCUj4mZ3Q7TFNQIHNlZ21lbnQsIGRpZmZlcmVudCBsYWJlbCBhbGxvY2F0aW9uIHJ1bGVzOyBl dGMpIFJFUVVJUkVEIHRvIHN0aXRjaDxCUj4mZ3Q7cGFja2V0IExTUHMgaW4gZGF0YSBwbGFuZSA/ PEJSPjwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPi0mZ3Q7IHll cyAoYmVzaWRlIG1hbnVhbCBjb25maWcgb3BlcmF0aW9uKTwvRk9OVD48L1A+PFA+PEZPTlQgRkFD RT0iTW9ub3NwYWNlLENvdXJpZXIiPjxCUj4mZ3Q7MikgQXJlIHRoZSBjb250cm9sIHBsYW5lIHNp Z25hbGluZyBwcm9jZWR1cmVzIGZvciBMU1Agc3RpdGNoaW5nIGFzPEJSPiZndDtkZXNjcmliZWQg aW4gdGhpcyBJRCwgKGV4cGxpY2l0IHJlcXVlc3QgZm9yIExTUCBzdGl0Y2hpbmcgb248QlI+Jmd0 O0xTUCBzZWdtZW50LCBkaWZmZXJlbnQgbGFiZWwgYWxsb2NhdGlvbiBydWxlczsgZXRjKSBSRVFV SVJFRCB0byBzdGl0Y2g8QlI+Jmd0O25vbi1wYWNrZXQgTFNQcyBpbiBkYXRhIHBsYW5lID88QlI+ PC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+LSZndDsgeWVzIChi ZXNpZGUgbWFudWFsIGNvbmZpZyBvcGVyYXRpb24pPC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJN b25vc3BhY2UsQ291cmllciI+PEJSPiZndDszKSBJZiBhbiBlMmUgTFNQIGNyb3NzZXMgYSByZWdp b24gYm91bmRhcnkgKGJhc2VkIG9uIGRpZmZlcmVudCBURTxCUj4mZ3Q7bGluayBzd2l0Y2hpbmcg Y2FwYWJpbGl0aWVzKSwgdGhlbiBjYW4gSSB1c2UgTFNQIHN0aXRjaGluZyAtPEJSPiZndDthKSBj b250cm9sIHBsYW5lIHByb2NlZHVyZXMgZm9yIHBhY2tldCBMU1BzPEJSPiZndDtiKSBkYXRhIHBs YW5lIHByb2NlZHVyZXMgZm9yIHBhY2tldCBMU1BzPEJSPiZndDtjKSBjb250cm9sIHBsYW5lIHBy b2NlZHVyZXMgZm9yIG5vbi1wYWNrZXQgTFNQczxCUj4mZ3Q7ZCkgZGF0YSBwbGFuZSBwcm9jZWR1 cmVzIGZvciBub24tcGFja2V0IExTUHM8QlI+PC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25v c3BhY2UsQ291cmllciI+LSZndDsgbm8sIHdoZW4gY3Jvc3NpbmcgYSByZWdpb24gYm91bmRhcnks IHlvdSBzaW1wbHkgdXNlIGEgZG9jdW1lbnQgbmFtZWQgJnF1b3Q7ZHJhZnQtbXBscy1sc3AtaGll cmFyY2h5LTA4LnR4dCZxdW90OyB3aGljaCBpcyBpbiB0aGUgUkZDIEVkaXRvciBxdWV1ZSwgYnV0 IHRoaXMgZG9lcyBub3QgcHJldmVudCBmcm9tIGNyZWF0aW5nIGFuIExTUCBzZWdtZW50IG9uIHRv cCBvZiBhbiBGQS1MU1AgKHNlZSBiZWxvdyk8L0ZPTlQ+PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9z cGFjZSxDb3VyaWVyIj48QlI+Jmd0OzQpIElmIHRoZSBzd2l0Y2hpbmcgdHlwZSBvZiBhbiBlMmUg TFNQIGlzIGRpZmZlcmVudCBmcm9tIHRoYXQgb2YgYW4gTFNQPEJSPiZndDtzZWdtZW50LCBjYW4g SSB1c2UgTFNQIHN0aXRjaGluZyAtPEJSPiZndDthKSBjb250cm9sIHBsYW5lIHByb2NlZHVyZXMg Zm9yIHBhY2tldCBMU1BzPEJSPiZndDtiKSBkYXRhIHBsYW5lIHByb2NlZHVyZXMgZm9yIHBhY2tl dCBMU1BzPEJSPiZndDtjKSBjb250cm9sIHBsYW5lIHByb2NlZHVyZXMgZm9yIG5vbi1wYWNrZXQg TFNQczxCUj4mZ3Q7ZCkgZGF0YSBwbGFuZSBwcm9jZWR1cmVzIGZvciBub24tcGFja2V0IExTUHM8 QlI+PC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+LSZndDsgbm8s IGluIHRoaXMgY2FzZSBwcm9ncmVzc2lvbiBvZiB0aGUgZW5kLXRvLWVuZCBMU1AgZXN0YWJsaXNo bWVudCwgaW1wbGllcyB0aGF0IHRoZSBMU1Agc2VnbWVudCBtdXN0IGZ1bGZpbCBhIHNldCBvZiBj b25zdHJhaW50cyBjYXJyaWVkIGFzIHBhcnQgb2YgdGhlIGluY29taW5nIGVuZC10by1lbmQgTFNQ IHJlcXVlc3QsIGluIHRoaXMgY2FzZSBvbmUgaGFzIHRvIGNyZWF0ZSBhbiBMU1Agc2VnbWVudCAo d2l0aCBhIHN3aXRjaGluZyB0eXBlIGVxdWFsIHRvIHRoZSBpbmNvbWluZyBlbmQtdG8tZW5kIExT UCBzd2l0Y2hpbmcgdHlwZSkgb3ZlciB0aGF0IEZBIChpLmUuIHJlZmVycmVkIGhlcmUgYWJvdmUg dG8gdGhlIExTUCBzZWdtZW50IHdpdGggYSBkaWZmZXJlbnQgU3dpdGNoaW5nIFR5cGUpPC9GT05U PjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+PEJSPiZndDtOb3RlIHRoYXQg YWx0aG91Z2ggMykgYW5kIDQpIGFwcGVhciB0byBiZSB0aGUgc2FtZSwgdGhlIHJlYXNvbiBJIGhh dmU8QlI+Jmd0O3RoZW0gYXMgZGlmZmVyZW50IHF1ZXN0aW9ucyBpcyB0aGF0IGluIG9uZSBjYXNl IHRoZSBURSBsaW5rIHN3aXRjaGluZzxCUj4mZ3Q7Y2FwYWJpbGl0eSBpcyBleGFtaW5lZCBhbmQg aW4gdGhlIG90aGVyIGNhc2UgdGhlIEdlbmVyYWxpemVkIExhYmVsIFJlcXVlc3Q8QlI+Jmd0O2lz IGV4YW1pbmVkLjxCUj48QlI+PC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291 cmllciI+LSZndDsgd2hhdCBpdCBhcHBlYXJzIHRvIG1lIGlzIHRoYXQgdGhpcyBkaXNjdXNzaW9u IG1peGVzIHRoZSBpc3N1ZSBvZiB0aGUgb3BlcmF0aW9uIGRlc2NyaWJpbmcgdGhlIHN0aXRjaGlu ZyBvcGVyYXRpb25zIGF0IHRoZSBkb21haW4gYm91bmRhcmllcyBmcm9tIHRoZSBMU1Agc2VnbWVu dCBjcmVhdGlvbiBpdHNlbGYgaS5lLiBhbiBMU1Agc2VnbWVudCBjYW4gaW5kZWVkIGJlIGFuIEZB IGxpbmsgaW5oZXJpdGVkIGZyb20gdGhlIGNyZWF0aW9uIG9mIGFuIChpbnRyYS1kb21haW4pIEZB LUxTUCAtIDwvRk9OVD48L1A+ Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 08:14:06 +0000 Sensitivity: Subject: Re: Moving LSP ownership between control plane and management plane To: "George Newsome <gnewsome" <gnewsome@ieee.org> Cc: "Adrian Farrel <adrian" <adrian@olddog.co.uk>, "ccamp" <ccamp@ops.ietf.org> Message-ID: <OFB6212453.3AB37B0A-ONC1257044.00298283-C1257044.002D39EE@uk.marconicomms.com> From: "Diego Caviglia" <Diego.Caviglia@marconi.com> Date: Wed, 20 Jul 2005 10:13:55 +0200 MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii Hi Geroge, I remember quite well that discussuion, but I also remember that the common feeling was that the proposed feature is useful. I can agree that this implies some discussion on the naming of the resource and so on but the ID assumes that all this stuff has been resolved. Some more comment in line. Regards Diego George Newsome <gnewsome@ieee.org>@ops.ietf.org on 19/07/2005 13.44.29 Sent by: owner-ccamp@ops.ietf.org To: Adrian Farrel <adrian@olddog.co.uk> cc: ccamp@ops.ietf.org Subject: Re: Moving LSP ownership between control plane and management plane Adrian, You asked for comments on this proposal, which was presented to ITU-T Q14 experts in January of this year. There are so many issues involved in moving a connection from the management plane to the control plane that involve intensive management action, that it is not evident that signaling need be involved at all. [dc] I agree with the fisrt part of your sentence but not with the second. >From a CP point of view 'adopt' a LSP or create a new one is more or less the same (in fact the ID basically adds only one bit to the already existing signalling protocol) so saying that '...is not evident that signaling need be involved at all' for an adoption can be extended to signaling in general and that is why I disagree. As a personal opinion, it makes little sense to consider moving the connection. It makes more sense to replace the connection with one set up via the control plane, and to then delete the original management set up connection. [dc] Sorry but seems that this is traffic affecting that is something that usually operators wants to avoid. Appended is an excerpt of the meeting report detailing some of the issues that need to be resolved before signaling need be considered. Regards George Excerpt of the meeting report of Q14 Experts, January 2005 WD06 (Marconi) "G.7713 Modification in order to support Circuit Migration" contains a modified version of draft G.7713 Revision as was presented in the Berlin meeting (WD 11, i.e., TD50/3 Dec.2004). This contribution addresses SP24 (PC ? SPC) of G.7713. It proposes to extend the concept of Call/Connection setup to setup/adopt and release to release/de-adopt. In Connection Adoption both the SNC are not actually created and the LC are not actually established due to the fact that the underlying physical resources are already in place. In the Call/Connection dis-adoption SNC-LC-SNC is not actually de-allocated, only the Control Plane information associated with then is removed and the ownership is moved to Management Plane. Attributes for indicating Adoption/Dis-adoption are proposed for the INNI Connection messages. The discussion was mainly on the Management plane and Control plane actions required before signalling is involved. The group noted that the following issues need to be addressed: - Naming of transport resources to the control plane. Before a call can be placed under signalling control, links that are involved in the connection need to be given control plane names. Without these, no explicit route can be formed to match the resources of the connection. [dc] This is outside the scope of the ID and probabily also from the scope of the CCAMP/IETF. We assume that this point has been resolved in the ID. - Resource state while under dual CP and MP views. After resources are given control plane names, the resources are still viewed by the management plane. It may be necessary to create a new state for the resource to indicate that the management plane cannot perform some actions on the connection points as the control plane is about to take over. [dc] Again this is a NMS related issue that is uotside the scope of the ID/CCAMP and may be IETF. - Role of discovery functions (esp. CELA). After control plane names are allocated, distributed control plane functions need to be associated and communication adjacencies formed. This too is a precursor to any signalling procedures in migrating from the MP to CP. [dc] Yes but the ID assumes that this has been done. - Similarity of migration to synchronization after CP failure and subsequent recovery. Connection control procedures might be the same in migrating a call from the MP to the CP as the situation when the CP has failed and is recovering. Here, the network connections are already in place, but connection and call state need to be created to match it. Knowledge of the call and what the connection should be is obtained from the MP for migration, and from some reliable database in the recovery case. [dc] Adrian comment is also my comment - Call awareness of migration vs. connection being unaware of migration. When connection state is being created to match an existing connection, the connection controllers (CCs) do not require awareness of why this is happening as the context could be migration or recovery. A mechanism to create control state without affecting transport plane state is needed in the CCs. Call controllers though do need migration awareness as they need to obtain/derive call identification from the MP. [dc] As above. In summary, the group decided that it is premature to consider signalling procedure before the above issues, amongst other, have been studied. However the contribution does enable the group to expose in a larger context the interaction between CP and MP for connection migration. The above issues were included in SP24 of the G.7713 Living List (see WD22). Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 07:52:41 +0000 Sensitivity: Subject: RE: Moving LSP ownership between control plane and management plane To: "<neil.2.harrison" <neil.2.harrison@bt.com> Cc: "<gnewsome" <gnewsome@ieee.org>, "<adrian" <adrian@olddog.co.uk>, "<ccamp" <ccamp@ops.ietf.org> Message-ID: <OF0542E111.CA3D3ABB-ONC1257044.00296EEA-C1257044.002B31DE@uk.marconicomms.com> From: "Diego Caviglia" <Diego.Caviglia@marconi.com> Date: Wed, 20 Jul 2005 09:51:43 +0200 MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii Hi Neil, it always a pleasure read your e-mail but I think I'm missing something. The basic purpose of the ID is to provide a simple mechanism to move LSPs from a domain to another, I don't see any implication, with layering. Better to say this ID have the same implication with layering as 'normal' signalling. Moreover I agree with you that NMS is the 'king' of the network and in fact the ID forseen that is MNS to provide the right information to CP to move the LSP from a domain to another one. My basic question is: do someone think that having a mechanism to move MP/CP created LSP under the CP/MP domain is useless? My experience tell me that the answer is no, but I'm open to discuss this. Regards Diego <neil.2.harrison@bt.com>@ops.ietf.org on 19/07/2005 23.36.27 Sent by: owner-ccamp@ops.ietf.org To: <gnewsome@ieee.org>, <adrian@olddog.co.uk> cc: <ccamp@ops.ietf.org> Subject: RE: Moving LSP ownership between control plane and management plane Hi George....nice to hear from you...hope all is well with you, This all sounds very reasonable to me. At the end of the day, all a signalling protocol is is a distributed agent of network management.....in the types of large BW granularity networks being discussed the NMS will always be king. Sorry if that's a problem to some folks, its just a fact for the way we see/implement it. This also follows from a good understanding of the commercials at play here. That is, there is a major difference between: - a public service context (cf PSTN), or extending a similar capability out to end-systems in the co-ps mode say, where end-systems/applications drive trail set-up (ie SVCs across a UNI). Here one hopes there could be a large population of users (think peer-peer/BB) and that the trail holding-times are not 'too long'. In this environment there could be a viable business case based on a providing a given GoS at some cost.....and it might solve some (otherwise) hard QoS/pricing problems for operators....but I digress; - a private network builder service context where one is a providing a (relatively long time-constant) traffic-engineering function to a higher layer network, ie creating its topology. This is a client/server relationship and THE starting point for this must be functional decoupling of the client and server layer networks because they will often be owned by different parties. We are talking relatively long forecast/planning cycles here and long trail holding times....esp when it comes to real physical build. Further, the potential population is rather small compared to the previous case. If folks look at whether there is a viable business case here based on a providing a given GoS at some cost in an acceptable time period I suspect they will be disappointed. What this all means is that the control-plane is not such an important topic as we move down the layered network stack towards the duct....increasingly other factors like strong NMS and OAM assume a far greater importance. This is not to say a control-plane is not useful but: - we should be clear there cannot be any common addressing and routing functions across a set of nested layer networks, which is very obvious when different parties own different layer networks....one can (and should) re-use components (if they are the right ones), but they will need to operate in different layer network spaces - signalling is a useful agent for functions like S-PVCs and resilience. Well that's our analysis anyway....if anyone knows different I'd be happy to review their commercial assumptions with folks here. regards, Neil > -----Original Message----- > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On > Behalf Of George Newsome > Sent: 19 July 2005 12:44 > To: Adrian Farrel > Cc: ccamp@ops.ietf.org > Subject: Re: Moving LSP ownership between control plane and > management plane > > > Adrian, > > You asked for comments on this proposal, which was presented to ITU-T > Q14 experts in January of this year. > > There are so many issues involved in moving a connection from the > management plane to the control plane that involve intensive > management action, that it is not evident that signaling need be > involved at all. > > As a personal opinion, it makes little sense to consider moving the > connection. It makes more sense to replace the connection with one set > up via the control plane, and to then delete the original > management set > up connection. > > Appended is an excerpt of the meeting report detailing some of the > issues that need to be resolved before signaling need be considered. > > Regards > > George > > Excerpt of the meeting report of Q14 Experts, January 2005 > > WD06 (Marconi) "G.7713 Modification in order to support Circuit > Migration" contains a modified version of draft G.7713 Revision as was > presented in the Berlin meeting (WD 11, i.e., TD50/3 Dec.2004). This > contribution addresses SP24 (PC ? SPC) of G.7713. It proposes > to extend > the concept of Call/Connection setup to setup/adopt and release to > release/de-adopt. In Connection Adoption both the SNC are not > actually > created and the LC are not actually established due to the > fact that the > underlying physical resources are already in place. In the > Call/Connection dis-adoption SNC-LC-SNC is not actually de-allocated, > only the Control Plane information associated with then is > removed and > the ownership is moved to Management Plane. Attributes for indicating > Adoption/Dis-adoption are proposed for the INNI Connection messages. > > The discussion was mainly on the Management plane and Control plane > actions required before signalling is involved. The group noted that > the following issues need to be addressed: > > - Naming of transport resources to the control plane. Before a call > can be placed under signalling control, links that are involved in the > connection need to be given control plane names. Without these, no > explicit route can be formed to match the resources of the connection. > > - Resource state while under dual CP and MP views. After resources > are given control plane names, the resources are still viewed by the > management plane. It may be necessary to create a new state for the > resource to indicate that the management plane cannot perform some > actions on the connection points as the control plane is > about to take over. > > - Role of discovery functions (esp. CELA). After control plane names > are allocated, distributed control plane functions need to be > associated and communication adjacencies formed. This too is a > precursor to any signalling procedures in migrating from the MP to CP. > > - Similarity of migration to synchronization after CP failure and > subsequent recovery. Connection control procedures might be the same > in migrating a call from the MP to the CP as the situation when > the CP has > failed and is recovering. Here, the network connections are > already in > place, but connection and call state need to be created to match it. > Knowledge of the call and what the connection should be is > obtained from > the MP for migration, and from some reliable database in the > recovery case. > > - Call awareness of migration vs. connection being unaware of > migration. > When connection state is being created to match an existing > connection, the connection controllers (CCs) do not require > awareness of > why this is happening as the context could be migration or > recovery. A > mechanism to create control state without affecting transport plane > state is needed in the CCs. Call controllers though do need > migration > awareness as they need to obtain/derive call identification > from the MP. > > In summary, the group decided that it is premature to consider > signalling procedure before the above issues, amongst other, have been > studied. However the contribution does enable the group to > expose in a > larger context the interaction between CP and MP for connection > migration. The above issues were included in SP24 of the > G.7713 Living > List (see WD22). > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 07:16:13 +0000 Sensitivity: Subject: Re: Moving LSP ownership between control plane and management plane To: ccamp@ops.ietf.org Message-ID: <OFB25FEFC1.F1286C8E-ONC1257044.0027D50B-C1257044.0027EF13@uk.marconicomms.com> From: "Dino Bramanti" <Dino.Bramanti@marconi.com> Date: Wed, 20 Jul 2005 09:15:56 +0200 MIME-Version: 1.0 Content-type: text/plain; charset=us-ascii Thanks Adrian for commenting on proposed ID. I do agree basically with everything you are saying in your post. I confirm that our intention was trying to define a simple control plane mechanism, with (very) low impact on existing GRSVPTE procedures and message set, but still able to control the handover of LSP ownership between planes. In fact I agree that the definition of procedures within management plane to be done prior to LSP signaling within control plane are quite a complicated work, requiring probably a bigger effort. However, the proposed handover procedure has to take place conceptually and temporally after the information collection phase (and address translation / parameter mapping between management plane and control plane "language") and it is not dependent on how this is done. If we accept that signaling mechanism has to be involved "somehow" in a LSP handover operation (and I think so, for the very same reasons as Adrian in his post), then we can try to resolve the whole problem of MP/CP migration of LSP looking separately at its different phases. The phase covering the collection of the whole information required to signal a LSP over control plane is outside the scope of the ID, which takes it as a starting point for the definition of proposed mechanism. That is, we assume that MP has to provide CP with all the information needed to format a valid GRSVPTE path message, this being possible because the dataplane of the LSP to be handed over is already in place and in service and visible from its current owning plane. dino Thanks George, Useful input. > You asked for comments on this proposal, which was presented to ITU-T > Q14 experts in January of this year. Can you confirm that there were plenty of providers represented at the meeting. > There are so many issues involved in moving a connection from the > management plane to the control plane that involve intensive management > action, that it is not evident that signaling need be involved at all. Interesting. It is certainly valuable to have a list of these issues out in the open. I have made some comments in the appended meeting report. If the control plane is to take over the connection then we assume that signaling state must be built so that the control plane could then modify or tear down the connection. how would you propose that this be done without involving signaling? > As a personal opinion, it makes little sense to consider moving the > connection. It makes more sense to replace the connection with one set > up via the control plane, and to then delete the original management set > up connection. The question arrises of how this would be done in a network with limited resources. That is, if it is possible to set up a parallel conneciton then the method you describe would be fine, but if there are not enough parallel resources then would you propose that a lambda (for example) could be used for the management-established connection and for the control-plane-owned connection at the same time? This might present some difficulties as well. Do you have an opinion on my main question which is whether this tyoe of change-over would ever be done in service? Thanks, Adrian > Excerpt of the meeting report of Q14 Experts, January 2005 > > WD06 (Marconi) "G.7713 Modification in order to support Circuit > Migration" contains a modified version of draft G.7713 Revision as was > presented in the Berlin meeting (WD 11, i.e., TD50/3 Dec.2004). This > contribution addresses SP24 (PC ? SPC) of G.7713. It proposes to extend > the concept of Call/Connection setup to setup/adopt and release to > release/de-adopt. In Connection Adoption both the SNC are not actually > created and the LC are not actually established due to the fact that the > underlying physical resources are already in place. In the > Call/Connection dis-adoption SNC-LC-SNC is not actually de-allocated, > only the Control Plane information associated with then is removed and > the ownership is moved to Management Plane. Attributes for indicating > Adoption/Dis-adoption are proposed for the INNI Connection messages. > > The discussion was mainly on the Management plane and Control plane > actions required before signalling is involved. The group noted that the > following issues need to be addressed: > > - Naming of transport resources to the control plane. Before a call can > be placed under signalling control, links that are involved in the > connection need to be given control plane names. Without these, no > explicit route can be formed to match the resources of the connection. This is pretty obvious. But actually, this is an issue with moving the resources from one plane to another (or having them present in both planes). It is certainly a prerequisite for moving the connection, but it is not functionally an element of moving the connection. The draft, as I read it, assumes that these operations have been completed and then asks "how do I move the connection from one plane to the other?" > - Resource state while under dual CP and MP views. After resources are > given control plane names, the resources are still viewed by the > management plane. It may be necessary to create a new state for the > resource to indicate that the management plane cannot perform some > actions on the connection points as the control plane is about to take over. Yes. There is certainly scope for a race condition here. I believe that the I-D proposes to (try to) make the transfer an atomic action from the point of view of the change-over. Note, however, that it is wrong to assume that a resource belongs only to the control or to the management plane at any moment in time. This may be true with respect to the ability to provision/control connections, but resources that are available for use by the control plane can still be managed through the management plane. Thus, you take a laser out of service through the management plane, not through the control plane. > - Role of discovery functions (esp. CELA). After control plane names > are allocated, distributed control plane functions need to be associated > and communication adjacencies formed. This too is a precursor to any > signalling procedures in migrating from the MP to CP. Note to CCAMP: CELA is Control Entity Logical Adjacency. This too is "obvious". But this discussion seems to be wide of the mark. Before deploying a control plane, it is clear that this type of operation needs to take place. Since (within the context of CCAMP) a control plane *is* going to be deployed, we can assume this operation and move on to the discussion in the draft. > - Similarity of migration to synchronization after CP failure and > subsequent recovery. Connection control procedures might be the same in > migrating a call from the MP to the CP as the situation when the CP has > failed and is recovering. Here, the network connections are already in > place, but connection and call state need to be created to match it. > Knowledge of the call and what the connection should be is obtained from > the MP for migration, and from some reliable database in the recovery case. This is a valid point. Control plane failure, however, is unlikely to be total. That is, it would be unusual for more than one control plane link or node to fail. Thus, in the control plane, the "reliable database" is distributed and recovery can follow the procedures outlined in draft-ietf-ccamp-rsvp-restart-ext-03.txt It would certainly be worth considering whether the procedures used for control plane recovery could be applied to transfer of ownership from the management to control plane. at the moment, I suspect that they could not because the restart extensions assume a specific period during which signaling is associated with recovery. We would also need to look at the reverse process (transfer from CP to MP) that is not possible to cover by CP restart procedures. > - Call awareness of migration vs. connection being unaware of migration. > When connection state is being created to match an existing > connection, the connection controllers (CCs) do not require awareness of > why this is happening as the context could be migration or recovery. A > mechanism to create control state without affecting transport plane > state is needed in the CCs. Call controllers though do need migration > awareness as they need to obtain/derive call identification from the MP. I disagree. Connection controllers must be able to distinguish between resources that are already in use and cannot be assigned to a new connection, and those that are already in use and can be assigned to a new connection because the connection is replacing an existing connection. If this were happening entirely within the control plane we would use resource sharing, but when the operation spans the control and management planes resource ownership/sharing is more complex and the conneciotn controller needs to be able to issue the approriate instrucitons for the assignment of the resources. On the other hand, call controllers do not need awareness of the details of the realisation of the connections. > In summary, the group decided that it is premature to consider > signalling procedure before the above issues, amongst other, have been > studied. However the contribution does enable the group to expose in a > larger context the interaction between CP and MP for connection > migration. The above issues were included in SP24 of the G.7713 Living > List (see WD22). Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 06:43:12 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01C58CF6.0EC950A5" Subject: TR : I-D ACTION:draft-leroux-ccamp-gmpls-mrn-eval-01.txt Date: Wed, 20 Jul 2005 08:41:28 +0200 Message-ID: <D109C8C97C15294495117745780657AE02E55351@ftrdmel1.rd.francetelecom.fr> Thread-Topic: I-D ACTION:draft-leroux-ccamp-gmpls-mrn-eval-01.txt Thread-Index: AcWM86smKmrBtpT6Sbiy7zzQIn40xQAAAfMw From: "LE ROUX Jean-Louis RD-CORE-LAN" <jeanlouis.leroux@francetelecom.com> To: <ccamp@ops.ietf.org> This is a multi-part message in MIME format. ------_=_NextPart_001_01C58CF6.0EC950A5 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Hi all, Here is a new revision of the MLN/MRN evaluation draft. It takes into account comments/suggestions received on the lists. Here are the major changes: -Terminology MLN/MRN aligned with the requirement draft. -Concept of "Virtual FA" removed. Replaced by "Virtual TE-link": This requires an association between end-points so as to exchange several parameters (interface ids...). This may rely on extensions of the Call procedure. -Some rewording for clarification. Please comment on this new version, Regards, JL et. al. > A New Internet-Draft is available from the on-line Internet-Drafts > directories. >=20 >=20 > Title : Evaluation of existing GMPLS Protocols against > Multi Layer and Multi Region Networks > MLN/MRN) > Author(s) : J. Le Roux, et al. > Filename : draft-leroux-ccamp-gmpls-mrn-eval-01.txt > Pages : 13 > Date : 2005-7-19 > =09 > This document provides an evaluation of Generalized Multi-Protocol=20 > Label Switching (GMPLS) protocols and mechanisms against the=20 > requirements for Multi-Layer Networks (MLN) and Multi Region > Networks=20 > (MRN). In addition, this document identifies areas where additional >=20 > protocol extensions or procedures are needed to satisfy these=20 > requirements, and provides guidelines for potential extensions. >=20 > A URL for this Internet-Draft is: > http://www.ietf.org/internet-drafts/draft-leroux-ccamp-gmpls-mrn-eval- > 01.txt >=20 > To remove yourself from the I-D Announcement list, send a message to=20 > i-d-announce-request at ietf.org with the word unsubscribe in the body > of the message. =20 > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce >=20 > to change your subscription settings. >=20 >=20 > Internet-Drafts are also available by anonymous FTP. Login with the > username > "anonymous" and a password of your e-mail address. After logging in, > type "cd internet-drafts" and then > "get draft-leroux-ccamp-gmpls-mrn-eval-01.txt". >=20 > A list of Internet-Drafts directories can be found in > http://www.ietf.org/shadow.html=20 > or ftp://ftp.ietf.org/ietf/1shadow-sites.txt >=20 >=20 > Internet-Drafts can also be obtained by e-mail. >=20 > Send a message to: > mailserv at ietf.org. > In the body type: > "FILE > /internet-drafts/draft-leroux-ccamp-gmpls-mrn-eval-01.txt". > =09 > NOTE: The mail server at ietf.org can return the document in > MIME-encoded form by using the "mpack" utility. To use this > feature, insert the command "ENCODING mime" before the "FILE" > command. To decode the response(s), you will need "munpack" or > a MIME-compliant mail reader. Different MIME-compliant mail > readers > exhibit different behavior, especially when dealing with > "multipart" MIME messages (i.e. documents which have been split > up into multiple messages), so check your local documentation on > how to manipulate these messages. > =09 > =09 > Below is the data which will enable a MIME compliant mail reader > implementation to automatically retrieve the ASCII version of the > Internet-Draft. > <ftp://ftp.ietf.org/internet-drafts/draft-leroux-ccamp-gmpls-mrn-eval- > 01.txt> > <ftp://ftp.ietf.org/internet-drafts/draft-leroux-ccamp-gmpls-mrn-eval- > 01.txt> =20 > _______________________________________________ > I-D-Announce mailing list > I-D-Announce at ietf.org > https://www1.ietf.org/mailman/listinfo/i-d-announce >=20 >=20 ------_=_NextPart_001_01C58CF6.0EC950A5 Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN"> <HTML> <HEAD> <META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; = charset=3Dus-ascii"> <META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version = 6.5.7226.0"> <TITLE>TR : I-D ACTION:draft-leroux-ccamp-gmpls-mrn-eval-01.txt </TITLE> </HEAD> <BODY> <!-- Converted from text/rtf format --> <P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">Hi all,</FONT> </P> <P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">Here is a new = revision of the MLN/MRN evaluation draft.</FONT> <BR><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">It takes into = account comments/suggestions received on the lists.</FONT> </P> <P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">Here are the major = changes:</FONT> </P> <P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">-Terminology MLN/MRN = aligned with the requirement draft.</FONT> <BR><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">-Concept of = "Virtual FA" removed. Replaced by "Virtual = TE-link":</FONT> <BR><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">This requires an = association between end-points so as to exchange several parameters = (interface ids…).</FONT> <BR><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">This may rely on = extensions of the Call procedure.</FONT> <BR><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">-Some rewording for = clarification.</FONT> </P> <P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">Please comment on = this new version,</FONT> </P> <P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">Regards,</FONT> </P> <P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">JL et. al.</FONT> </P> <BR> <BR> <UL> <P><FONT SIZE=3D2 FACE=3D"Courier New">A New Internet-Draft is available = from the on-line Internet-Drafts directories.</FONT> </P> <BR> <P> <FONT SIZE=3D2 = FACE=3D"Courier New">Title = : Evaluation of existing = GMPLS Protocols against</FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier = New"> &n= bsp; &nb= sp; Multi Layer and Multi Region Networks MLN/MRN)</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">Author(s) : J. = Le Roux, et al.</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">Filename = : draft-leroux-ccamp-gmpls-mrn-eval-01.txt</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">Pages = : 13</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">Date = : 2005-7-19</FONT> <BR> =20 <BR><FONT SIZE=3D2 FACE=3D"Courier New">This document provides an = evaluation of Generalized Multi-Protocol </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New"> Label Switching = (GMPLS) protocols and mechanisms against the </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New"> requirements for = Multi-Layer Networks (MLN) and Multi Region Networks </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New"> (MRN). In addition, = this document identifies areas where additional </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New"> protocol extensions = or procedures are needed to satisfy these </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New"> requirements, and = provides guidelines for potential extensions.</FONT> </P> <P><FONT SIZE=3D2 FACE=3D"Courier New">A URL for this Internet-Draft = is:</FONT> <BR><A = HREF=3D"http://www.ietf.org/internet-drafts/draft-leroux-ccamp-gmpls-mrn-= eval-01.txt"><U><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Courier = New">http://www.ietf.org/internet-drafts/draft-leroux-ccamp-gmpls-mrn-eva= l-01.txt</FONT></U></A> </P> <P><FONT SIZE=3D2 FACE=3D"Courier New">To remove yourself from the I-D = Announcement list, send a message to </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">i-d-announce-request at ietf.org = with the word unsubscribe in the body of the message. </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">You can also visit </FONT><A = HREF=3D"https://www1.ietf.org/mailman/listinfo/I-D-announce"><U><FONT = COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Courier = New">https://www1.ietf.org/mailman/listinfo/I-D-announce</FONT></U></A><F= ONT SIZE=3D2 FACE=3D"Courier New"> </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">to change your subscription = settings.</FONT> </P> <BR> <P><FONT SIZE=3D2 FACE=3D"Courier New">Internet-Drafts are also = available by anonymous FTP. Login with the username</FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">"anonymous" and a = password of your e-mail address. After logging in,</FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">type "cd = internet-drafts" and then</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">"get = draft-leroux-ccamp-gmpls-mrn-eval-01.txt".</FONT> </P> <P><FONT SIZE=3D2 FACE=3D"Courier New">A list of Internet-Drafts = directories can be found in</FONT> <BR><A HREF=3D"http://www.ietf.org/shadow.html"><U><FONT = COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Courier = New">http://www.ietf.org/shadow.html</FONT></U></A><FONT SIZE=3D2 = FACE=3D"Courier New"> </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">or </FONT><A = HREF=3D"ftp://ftp.ietf.org/ietf/1shadow-sites.txt"><U><FONT = COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Courier = New">ftp://ftp.ietf.org/ietf/1shadow-sites.txt</FONT></U></A> </P> <BR> <P><FONT SIZE=3D2 FACE=3D"Courier New">Internet-Drafts can also be = obtained by e-mail.</FONT> </P> <P><FONT SIZE=3D2 FACE=3D"Courier New">Send a message to:</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">mailserv at ietf.org.</FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">In the body type:</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">"FILE = /internet-drafts/draft-leroux-ccamp-gmpls-mrn-eval-01.txt".</FONT> <BR> =20 <BR><FONT SIZE=3D2 FACE=3D"Courier New">NOTE: The mail = server at ietf.org can return the document in</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">MIME-encoded form by using the "mpack" = utility. To use this</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">feature, insert the command "ENCODING = mime" before the "FILE"</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">command. To decode the response(s), you will = need "munpack" or</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">a MIME-compliant mail reader. Different = MIME-compliant mail readers</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">exhibit different behavior, especially when dealing = with</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">"multipart" MIME messages (i.e. documents = which have been split</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">up into multiple messages), so check your local = documentation on</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">how to manipulate these messages.</FONT> <BR> = =20 <BR> = =20 <BR><FONT SIZE=3D2 FACE=3D"Courier New">Below is the data which will = enable a MIME compliant mail reader</FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">implementation to automatically = retrieve the ASCII version of the</FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">Internet-Draft.</FONT> <BR><A = HREF=3D"ftp://ftp.ietf.org/internet-drafts/draft-leroux-ccamp-gmpls-mrn-e= val-01.txt"><U><FONT COLOR=3D"#0000FF" FACE=3D"Times New = Roman"><ftp://ftp.ietf.org/internet-drafts/draft-leroux-ccamp-gmpls-mr= n-eval-01.txt></FONT></U></A><FONT FACE=3D"Times New Roman"> </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier = New">_______________________________________________</FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">I-D-Announce mailing list</FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">I-D-Announce at ietf.org</FONT> <BR><A = HREF=3D"https://www1.ietf.org/mailman/listinfo/i-d-announce"><U><FONT = COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Courier = New">https://www1.ietf.org/mailman/listinfo/i-d-announce</FONT></U></A> </P> <BR> </UL> </BODY> </HTML> ------_=_NextPart_001_01C58CF6.0EC950A5-- Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 01:50:45 +0000 Message-ID: <42DDADFB.50901@kddilabs.jp> Date: Wed, 20 Jul 2005 10:50:51 +0900 From: Tomohiro Otani <otani@kddilabs.jp> User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; ja-JP; rv:1.7.2) Gecko/20040803 MIME-Version: 1.0 To: Dimitri.Papadimitriou@alcatel.be Cc: ccamp@ops.ietf.org Subject: Re: http://www.ietf.org/internet-drafts/draft-otani-ccamp-gmpls-cspf-constraints-00 Content-Type: text/plain; charset=ISO-2022-JP Content-Transfer-Encoding: 7bit Hi Dimitri, Ok, I understand what you said. We need more feedback about this. Regards, tomo Dimitri.Papadimitriou@alcatel.be wrote: >hi, > >i would say that 1) should be considered iff there have been several identified >misbehaviour detected and inconsistencies observed such that a set of >prescriptive guidelines have to be put together (but this would translate some >"common" mis-interpretation of the set of documents we currently have) > >if this is not the case, then beside some base description/clarification, i >don't see how we could turn this into 2) or 3) without obtaining a larger sample >of CSPF/routing practices - the difference between 2) and 3) being that in the >former case, information gathered is processed to describe practices and devise >operational guidelines while in the latter the document appears as a collection >of practices from individual practices > >in any case, and in order to progress - we need more feedback - > > > >thanks, > >-dimitri. > > >*Tomohiro Otani <otani@kddilabs.jp>* >Sent by: owner-ccamp@ops.ietf.org >07/19/2005 09:48 ZE9 > >To: Dimitri PAPADIMITRIOU/BE/ALCATEL@ALCATEL >cc: ccamp@ops.ietf.org >bcc: >Subject: Re: >http://www.ietf.org/internet-drafts/draft-otani-ccamp-gmpls-cspf-constraints-00 > > >Hi Dimitri, > >Per our initial discussion, we would like to formulate >the unified view on CSPF consideration, and we simply >thought it is BCP. > >Considering more deeply, our draft does not define the >protocol itself but the protocol related specification > >In that sense, you are right, it is prescriptive more >than descriptive for guiding operations. Moreover, I came >up my mind that it may be more standard track than informational. >But I am not sure. What's your thought on this ? > >With best regards, > >tomo > >PS: We posted 01 version. > > > > >Dimitri.Papadimitriou@alcatel.be wrote: > > >authors, is the purpose of the above mentioned draft: > > > > 1. prescriptive (even if informational) > > > > 2. descriptive for guiding operations (BCP) > > > > 3. informative (survey-like) > > > >from what i read, it is more 1 like - is my understanding correct ? > > > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 20 Jul 2005 00:22:29 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: RE: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON Lexicography Date: Tue, 19 Jul 2005 19:19:50 -0500 Message-ID: <A1A52203CA93634BA1748887B9993AEAFF2195@USNVEX1.tellabs-west.tellabsinc.net> Thread-Topic: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON Lexicography Thread-Index: AcWMXShS3XIKS6HkT1SIyh9DjwERlgAXG04A From: "Mack-Crane, T. Benjamin" <Ben.Mack-Crane@tellabs.com> To: "Adrian Farrel" <adrian@olddog.co.uk>, <ccamp@ops.ietf.org> Cc: "Kireeti Kompella" <kireeti@juniper.net>, <zinin@psg.com>, "Bill Fenner" <fenner@research.att.com>, "Scott Bradner" <sob@harvard.edu> Hello Adrian, I have provided some clarification of the assumptions in-line below, which may be helpful in determining whether the CCAMP comments are appropriate. Regards, Ben > -----Original Message----- > From: owner-ccamp@ops.ietf.org=20 > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of Adrian Farrel > Sent: Tuesday, July 19, 2005 6:59 AM > To: ccamp@ops.ietf.org > Cc: 'Kireeti Kompella'; zinin@psg.com; Bill Fenner; Scott Bradner > Subject: Proposed response to ITU-T SG15 Questions 12 and 14=20 > on ASON Lexicography >=20 > Hi, >=20 > Please comment on this draft response liaison which I intend=20 > to send on > Monday 25th July. > You can see the original incoming liaison at > http://www.olddog.co.uk/incoming.htm >=20 > Thanks, > Adrian > =3D=3D=3D=3D=3D > To: ITU-T SG15 Questions 12 and 14 > From: IETF CCAMP Working Group > Subject: Response to your liaison on GMPLS/ASON Lexicography > For: Action > Deadline: August 31st 2005 >=20 > The CCAMP Working Group would like to thank SG15 and=20 > especially Questions > 12 and 14 for the time and effort they put into providing input and > feedback on the GMPLS/ASON lexicography Internet-Draft that is being > developed by CCAMP. Much of the material supplied as direct=20 > comments or as > mark-ups to the Internet-Draft has been incorporated into the latest > revision which is available at > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-gmpls-aso > n-lexicography-03.txt >=20 > In your liaison you state some assumptions which we would=20 > like to comment > on as follows. >=20 > - The purpose of the document is to enable those familiar with ASON > to understand GMPLS terminology in an ASON context >=20 > Correct. >=20 > - GMPLS terms are described informally in the document in a way that > is generally consistent with GMPLS (existing RFCs and work=20 > in progress) >=20 > While the definition of GMPLS terms would be helpful for understanding > GMPLS terms in a GMPLS context, that is not the purpose of=20 > this document. This assumption is intended to convey that the descriptions in this draft are not definitional (authoritative), but that they can be used as a basis for the ASON interpretation (since concise defintions of GMPLS terms are not available, in part due to the fact that some GMPLS terms in the Lex draft are not yet used in GMPLS RFCs or internet drafts). It was not intended to convey anything about their use in a GMPLS context. >=20 > - Based on these descriptions, interpretations of these GMPLS terms > in ASON terminology are provided >=20 > It is intended that the interpretation of the GMPLS terms is=20 > not based on > these descriptions, but based on the full meaning of the=20 > GMPLS terms. The > descriptions of the GMPLS terms are provided in this=20 > Internet-Draft for > context and a brief summary. This presents a difficult problem. It is not feasible for the ITU experts to review the ASON interpretations without some basis. If the intended basis is existing GMPLS RFCs and drafts, a proper review would require reading and understanding all the uses of GMPLS terms in those documents. I have attempted this, for the normative references, and I believe it is unreasonable to expect this to be an effective method of review. For example, as noted above, some terms in the Lex draft are not used in the normative references. In other cases, terms are not used consistently and it would be difficult for reviewers to determine which uses are "correct." As a result, a review on this basis may not be feasible given the time available to ITU experts. This assumption was included to make the review process tractable. The ITU experts provided comments based on the descriptions of GMPLS terms in the Lex draft. If this is not what CCAMP is expecting, some clarification of the expectations would be helpful. >=20 > - CCAMP has primarily responsibility for the descriptions of=20 > GMPLS terms >=20 > Correct. >=20 > - Q12/15 has responsibility to provide input to the GMPLS-ASON > interpretations (based on the descriptions in this draft) >=20 > We welcome this assumption of responsibility by Q12/15 to=20 > provide input to > the Internet-Draft. While the Internet-Draft remains under the overall > editorial control of the CCAMP working group, we hope to be=20 > able to use > the text supplied by Q12/15 with only editorial changes. This is the understanding of SG15 as well (that SG15 would provide comments, not finished text). The comments were provided in the liaison as edited text as a convenience, not intended as final text. >=20 >=20 Snip... =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D The information contained in this message may be privileged and confidential and protected from disclosure. If the reader of this message is not the intended recipient, or an employee or agent responsible for delivering this message to the intended recipient, you are hereby notified that any reproduction, dissemination or distribution of this communication is strictly prohibited. If you have received this communication in error, please notify us immediately by replying to the message and deleting it from your computer. Thank you. Tellabs =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 23:21:53 +0000 Date: Tue, 19 Jul 2005 16:21:30 -0700 (PDT) From: Arthi Ayyangar <arthi@juniper.net> To: Igor Bryskin <ibryskin@movaz.com> cc: ccamp@ops.ietf.org Subject: LSP stitching questions for discussion Message-ID: <20050719160923.X71817@zircon.juniper.net> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi, Here are some of the questions related to LSP stitching to start a discussion on the usage of LSP stitching. 1) Are the control plane signaling procedures for LSP stitching as described in this ID, (explicit request for LSP stitching on LSP segment, different label allocation rules; etc) REQUIRED to stitch packet LSPs in data plane ? 2) Are the control plane signaling procedures for LSP stitching as described in this ID, (explicit request for LSP stitching on LSP segment, different label allocation rules; etc) REQUIRED to stitch non-packet LSPs in data plane ? 3) If an e2e LSP crosses a region boundary (based on different TE link switching capabilities), then can I use LSP stitching - a) control plane procedures for packet LSPs b) data plane procedures for packet LSPs c) control plane procedures for non-packet LSPs d) data plane procedures for non-packet LSPs 4) If the switching type of an e2e LSP is different from that of an LSP segment, can I use LSP stitching - a) control plane procedures for packet LSPs b) data plane procedures for packet LSPs c) control plane procedures for non-packet LSPs d) data plane procedures for non-packet LSPs Note that although 3) and 4) appear to be the same, the reason I have them as different questions is that in one case the TE link switching capability is examined and in the other case the Generalized Label Request is examined. thanks, -arthi Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 23:08:22 +0000 Date: Tue, 19 Jul 2005 16:07:49 -0700 (PDT) From: Arthi Ayyangar <arthi@juniper.net> To: Igor Bryskin <ibryskin@movaz.com> cc: ccamp@ops.ietf.org Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Message-ID: <20050719115057.Q71817@zircon.juniper.net> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi Igor, > One more comment. I think the draft should give a clear and unambiguous > answer to how e2e LSP and a stitching LSP (S-LSP) relate from the switching > capability point of view. Specifically, should they belong to the same > network layer, should they share the same switching type or maybe there are > no limitations at all, and you can stitch your e2e LSP pretty much with > anything you got, provided that this "anything" has enough bandwidth. > > Let me give you a specific example. Is it possible to stitch a PSC1 e2e LSP > with a PSC2 S-LSP? You probably remember our discussion with Kireeti in > Minneapolis (and you were also part of it), and he said (and I 100% agree > with him) that although it is doable (you can do anything you want with the > label stack), it is not right and this should be logically prohibited. What > you should do in this case is to nest the e2e LSP within the PSC2 LSP (that > is, use it as H-LSP) even in case it consumes the entire PSC2 LSP bandwidth. <AA>---> I do remember this discussion. > It is more clear in non-PSC cases. Although it is possible to "stitch" an > e2e VC12 LSP with a VC4 S-LSP, I could not imagine that a network operator > would ever agree with such solution. He is likely to say " What ? If the > service is supposed to stay for 12 month, than for entire year I pretty much > burry my 62 VC12 channels - I cannot use them even conceptually in case > when, say, on the next day I'll have to provide another 62 VC12 services > that could've used these wasted resources" > > On the other hand I know plenty GMPLS people who would say: "Sure, no > problem. Of course we can stitch a PSC e2e LSP with lambda S-LSP: If > somebody wants to do a trick stupid as it is, why should we prevent him from > doing so?" <AA>----> I know some of these people too. :-) > The bottom line is that a discussion like this in the draft and consensus > on how we see the stitching IMO would be extremely valuable for the GMPLS > multi-layer story, and I don't see a reason why this could not be achieved > in the frame of this work. <AA>------> I agree that we should have a discussion on this, but I would like folks to have this discussion on the list first, which I thank you for initiating. Yes, the draft has to state clearly whether LSP stitching between different switching capabilities is allowed or not. But it's more than that. Currently, the draft talks about this under sections 3.1 and 4.1 and 4.1.2. But, let us have this discussion on the list and we will be glad to add what's missing (with required details) to the ID, depending on the consensus reached. I will send out the questions to start the discussion in the next email. thanks, -arthi > > ----- Original Message ----- > From: "Arthi Ayyangar" <arthi@juniper.net> > To: "Igor Bryskin" <ibryskin@movaz.com> > Cc: <i-d-announce@ietf.org>; <Internet-Drafts@ietf.org>; > <ccamp@ops.ietf.org> > Sent: Monday, July 18, 2005 2:51 PM > Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > > > > Hi Igor, > > > > Please see inline. > > > > > 1. Why are saying that LSP Stitching is a private case of LSP Hierarchy? > > > > > there is more differences than similarities: > > > The differences are: > > > 1) In case of H-LSP there is a data plane adjacency, while in case of > S-LSP > > > there is none (as you correctly pointed out); > > > 2) In case of H-LSP there is an adaptation in data plane (label push/pop > for > > > PSC), while in case of S-LSP there is none - just simple > cross-connecting > > > ( label swap) as in case of two "native " e2e LSP adjacent segments ; > > > 3) H-LSP could be used by many e2e LSPs, while S-LSP could be used by > > > exactly one e2e LSP > > > 4) Signaling is different - there is no label negotiation in stitching > > > 5) H-LSP is used as a "true" data link, specifically there is a resource > > > allocation on the H-LSP edges, while in case of S-LSP there is none > > > 6) From MLN point of view, H-LSP is created in a server (lower) layer, > while > > > the S-LSP is created in the client (same as e2e LSP) layer. > > > > > > There are two similarities that I can think of: > > > 1) There is a signaling and possibly routing (see below) adjacencies > between > > > the ends; > > > 2) Both H-LSP and S-LSP could be advertised as separate TE links or as > TE > > > bundles > > > > > > I would recommend to dedicate a paragraph and enlist there similarities > and > > > differencies > > -------> Based on the discussions on the list over the last revision, the > > message that the authors received was that we do not want to go about such > > a lengthy discussion of similarities and differences (NOTE that the draft > > already does clearly highlight them where applicable). Instead we want > > this ID to simply explain how LSP stitching functions. In other words the > > idea was to make this ID a complete document by itself. But since it does > > borrow concepts from the LSP hierarchy ID, just state the concepts that > > are applicable or inapplicable. > > > > > 2. Why are you saying that a TE Link based on S-LSP can be used for > exactly > > > one e2e LSP? Parallel S-LSPs could be advertised as a single TE link > > > (bundle, see above) and hence can accomadate several e2e LSPs. > > ----> Igor, the draft does talk about this Bundling case. > > > > > 3. You are saying that S-LSP does not have a routing peering. Actually, > in > > > this respect it is no different from H-LSP: if it is advertised as a TE > link > > > into the same TE domain that was used for S-LSP creation (unlikely IMO > > > scenario) than it does not require the routing adjacency (in other > words, it > > > is an FA according to LSP-HIER definition), otherwise, it IS NOT and FA > and > > > does require the direct routing peering in the domain it is advertised > to > > > make it useful as a TE link in this domain > > -----> I am sorry, but I am missing your point here completely. Are you > > saying that the statement "the end points of an LSP segment do not have a > > routing adjacency", is incorrect or are you saying that "this is > > obvious" ? It is unclear to me what exactly your argument is. Please > > clarify. > > > > thanks, > > -arthi > > > > > > > > Cheers, > > > Igor > > > > > > > > > > > > ----- Original Message ----- > > > From: <Internet-Drafts@ietf.org> > > > To: <i-d-announce@ietf.org> > > > Cc: <ccamp@ops.ietf.org> > > > Sent: Friday, July 15, 2005 3:50 PM > > > Subject: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > > > > > > > > > > A New Internet-Draft is available from the on-line Internet-Drafts > > > directories. > > > > This draft is a work item of the Common Control and Measurement Plane > > > Working Group of the IETF. > > > > > > > > Title : Label Switched Path Stitching with Generalized MPLS Traffic > > > Engineering > > > > Author(s) : A. Ayyangar, J. Vasseur > > > > Filename : draft-ietf-ccamp-lsp-stitching-01.txt > > > > Pages : 19 > > > > Date : 2005-7-15 > > > > > > > > In certain scenarios, there may be a need to combine together two > > > > different Generalized Multi-Protocol Label Switching (GMPLS) Label > > > > Switched Paths (LSPs) such that in the data plane, a single end-to- > > > > end (e2e) LSP is achieved and all traffic from one LSP is switched > > > > onto the other LSP. We will refer to this as "LSP stitching". > This > > > > document covers cases where: a) the node performing the stitching > > > > does not require configuration of every LSP pair to be stitched > > > > together b) the node performing the stitching is not the egress of > > > > any of the LSPs c) LSP stitching not only results in an end-to-end > > > > LSP in the data plane, but there is also a corresponding end-to-end > > > > LSP (RSVP session) in the control plane. It might be possible to > > > > configure a GMPLS node to switch the traffic from an LSP for which > it > > > > is the egress, to another LSP for which it is the ingress, without > > > > requiring any signaling or routing extensions whatsoever, > completely > > > > transparent to other nodes. This will also result in LSP stitching > > > > in the data plane. However, this document does not cover this > > > > scenario of LSP stitching. > > > > > > > > A URL for this Internet-Draft is: > > > > > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt > > > > > > > > To remove yourself from the I-D Announcement list, send a message to > > > > i-d-announce-request@ietf.org with the word unsubscribe in the body of > the > > > message. > > > > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > > > > to change your subscription settings. > > > > > > > > > > > > Internet-Drafts are also available by anonymous FTP. Login with the > > > username > > > > "anonymous" and a password of your e-mail address. After logging in, > > > > type "cd internet-drafts" and then > > > > "get draft-ietf-ccamp-lsp-stitching-01.txt". > > > > > > > > A list of Internet-Drafts directories can be found in > > > > http://www.ietf.org/shadow.html > > > > or ftp://ftp.ietf.org/ietf/1shadow-sites.txt > > > > > > > > > > > > Internet-Drafts can also be obtained by e-mail. > > > > > > > > Send a message to: > > > > mailserv@ietf.org. > > > > In the body type: > > > > "FILE /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt". > > > > > > > > NOTE: The mail server at ietf.org can return the document in > > > > MIME-encoded form by using the "mpack" utility. To use this > > > > feature, insert the command "ENCODING mime" before the "FILE" > > > > command. To decode the response(s), you will need "munpack" or > > > > a MIME-compliant mail reader. Different MIME-compliant mail readers > > > > exhibit different behavior, especially when dealing with > > > > "multipart" MIME messages (i.e. documents which have been split > > > > up into multiple messages), so check your local documentation on > > > > how to manipulate these messages. > > > > > > > > > > > > Below is the data which will enable a MIME compliant mail reader > > > > implementation to automatically retrieve the ASCII version of the > > > > Internet-Draft. > > > > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 22:57:31 +0000 To: Tomohiro Otani <otani@kddilabs.jp> Cc: Dimitri.Papadimitriou@alcatel.be, ccamp@ops.ietf.org MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: Re: http://www.ietf.org/internet-drafts/draft-otani-ccamp-gmpls-cspf-constraints-00 Date: Wed, 20 Jul 2005 00:56:09 +0200 Message-ID: <OF48901CFD.5E264937-ONC1257043.007DFD40-C1257043.007DFDD6@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmhpLCA8L0ZPTlQ+PC9QPjxQPjxGT05U IEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5pIHdvdWxkIHNheSB0aGF0IDEpIHNob3VsZCBiZSBj b25zaWRlcmVkIGlmZiB0aGVyZSBoYXZlIGJlZW4gc2V2ZXJhbCBpZGVudGlmaWVkIG1pc2JlaGF2 aW91ciBkZXRlY3RlZCBhbmQgaW5jb25zaXN0ZW5jaWVzIG9ic2VydmVkIHN1Y2ggdGhhdCBhIHNl dCBvZiBwcmVzY3JpcHRpdmUgZ3VpZGVsaW5lcyBoYXZlIHRvIGJlIHB1dCB0b2dldGhlciAoYnV0 IHRoaXMgd291bGQgdHJhbnNsYXRlIHNvbWUgJnF1b3Q7Y29tbW9uJnF1b3Q7IG1pcy1pbnRlcnBy ZXRhdGlvbiBvZiB0aGUgc2V0IG9mIGRvY3VtZW50cyB3ZSBjdXJyZW50bHkgaGF2ZSkgPC9GT05U PjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+aWYgdGhpcyBpcyBub3QgdGhl IGNhc2UsIHRoZW4gYmVzaWRlIHNvbWUgYmFzZSBkZXNjcmlwdGlvbi9jbGFyaWZpY2F0aW9uLCBp IGRvbid0IHNlZSBob3cgd2UgY291bGQgdHVybiB0aGlzIGludG8gMikgb3IgMykgd2l0aG91dCBv YnRhaW5pbmcgYSBsYXJnZXIgc2FtcGxlIG9mIENTUEYvcm91dGluZyBwcmFjdGljZXMgLSB0aGUg ZGlmZmVyZW5jZSBiZXR3ZWVuIDIpIGFuZCAzKSBiZWluZyB0aGF0IGluIHRoZSBmb3JtZXIgY2Fz ZSwgaW5mb3JtYXRpb24gZ2F0aGVyZWQgaXMgcHJvY2Vzc2VkIHRvIGRlc2NyaWJlIHByYWN0aWNl cyBhbmQgZGV2aXNlIG9wZXJhdGlvbmFsIGd1aWRlbGluZXMgd2hpbGUgaW4gdGhlIGxhdHRlciB0 aGUgZG9jdW1lbnQgYXBwZWFycyBhcyBhIGNvbGxlY3Rpb24gb2YgcHJhY3RpY2VzIGZyb20gaW5k aXZpZHVhbCBwcmFjdGljZXM8L0ZPTlQ+PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3Vy aWVyIj5pbiBhbnkgY2FzZSwgYW5kIGluIG9yZGVyIHRvIHByb2dyZXNzIC0gd2UgbmVlZCBtb3Jl IGZlZWRiYWNrIC08L0ZPTlQ+PC9QPjxQPiZuYnNwOzwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3Bh Y2UsQ291cmllciI+dGhhbmtzLDwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENv dXJpZXIiPi1kaW1pdHJpLjwvRk9OVD48L1A+PFA+PEJSPjxGT05UIFNJWkU9Mj48Qj5Ub21vaGly byBPdGFuaSAmbHQ7b3RhbmlAa2RkaWxhYnMuanAmZ3Q7PC9CPjwvRk9OVD48QlI+PEZPTlQgU0la RT0yPlNlbnQgYnk6IG93bmVyLWNjYW1wQG9wcy5pZXRmLm9yZzwvRk9OVD48QlI+PEZPTlQgU0la RT0yPjA3LzE5LzIwMDUgMDk6NDggWkU5PC9GT05UPjxCUj48QlI+IDxGT05UIFNJWkU9Mj5Ubzo8 L0ZPTlQ+IDxGT05UIFNJWkU9Mj5EaW1pdHJpIFBBUEFESU1JVFJJT1UvQkUvQUxDQVRFTEBBTENB VEVMPC9GT05UPjxCUj4gPEZPTlQgU0laRT0yPmNjOjwvRk9OVD4gPEZPTlQgU0laRT0yPmNjYW1w QG9wcy5pZXRmLm9yZzwvRk9OVD48QlI+IDxGT05UIFNJWkU9Mj5iY2M6PC9GT05UPiA8QlI+IDxG T05UIFNJWkU9Mj5TdWJqZWN0OjwvRk9OVD4gPEZPTlQgU0laRT0yPlJlOiA8QSBIUkVGPWh0dHA6 Ly93d3cuaWV0Zi5vcmcvaW50ZXJuZXQtZHJhZnRzL2RyYWZ0LW90YW5pLWNjYW1wLWdtcGxzLWNz cGYtY29uc3RyYWludHMtMDA+aHR0cDovL3d3dy5pZXRmLm9yZy9pbnRlcm5ldC1kcmFmdHMvZHJh ZnQtb3RhbmktY2NhbXAtZ21wbHMtY3NwZi1jb25zdHJhaW50cy0wMDwvQT48L0ZPTlQ+PEJSPiA8 QlI+PEJSPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+SGkgRGltaXRyaSw8 QlI+PC9GT05UPjxCUj48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+UGVyIG91ciBpbml0 aWFsIGRpc2N1c3Npb24sIHdlIHdvdWxkIGxpa2UgdG8gZm9ybXVsYXRlPEJSPnRoZSB1bmlmaWVk IHZpZXcgb24gQ1NQRiBjb25zaWRlcmF0aW9uLCBhbmQgd2Ugc2ltcGx5PEJSPnRob3VnaHQgaXQg aXMgQkNQLjxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5Db25z aWRlcmluZyBtb3JlIGRlZXBseSwgb3VyIGRyYWZ0IGRvZXMgbm90IGRlZmluZSB0aGU8QlI+cHJv dG9jb2wgaXRzZWxmIGJ1dCB0aGUgcHJvdG9jb2wgcmVsYXRlZCBzcGVjaWZpY2F0aW9uPEJSPjwv Rk9OVD48QlI+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPkluIHRoYXQgc2Vuc2UsIHlv dSBhcmUgcmlnaHQsIGl0IGlzIHByZXNjcmlwdGl2ZSBtb3JlPEJSPnRoYW4gZGVzY3JpcHRpdmUg Zm9yIGd1aWRpbmcgb3BlcmF0aW9ucy4gTW9yZW92ZXIsIEkgY2FtZTxCUj51cCBteSBtaW5kIHRo YXQgaXQgbWF5IGJlIG1vcmUgc3RhbmRhcmQgdHJhY2sgdGhhbiBpbmZvcm1hdGlvbmFsLjxCUj5C dXQgSSBhbSBub3Qgc3VyZS4gV2hhdCdzIHlvdXIgdGhvdWdodCBvbiB0aGlzID88QlI+PC9GT05U PjxCUj48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+V2l0aCBiZXN0IHJlZ2FyZHMsPEJS PjwvRk9OVD48QlI+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPnRvbW88QlI+PC9GT05U PjxCUj48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+UFM6IFdlIHBvc3RlZCAwMSB2ZXJz aW9uLjxCUj48L0ZPTlQ+PEJSPjxCUj48QlI+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3Vy aWVyIj5EaW1pdHJpLlBhcGFkaW1pdHJpb3VAYWxjYXRlbC5iZSB3cm90ZTo8QlI+PC9GT05UPjxC Uj48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+Jmd0O2F1dGhvcnMsIGlzIHRoZSBwdXJw b3NlIG9mIHRoZSBhYm92ZSBtZW50aW9uZWQgZHJhZnQ6PEJSPiZndDs8QlI+Jmd0OyAxLiBwcmVz Y3JpcHRpdmUgKGV2ZW4gaWYgaW5mb3JtYXRpb25hbCk8QlI+Jmd0OzxCUj4mZ3Q7IDIuIGRlc2Ny aXB0aXZlIGZvciBndWlkaW5nIG9wZXJhdGlvbnMgKEJDUCk8QlI+Jmd0OzxCUj4mZ3Q7IDMuIGlu Zm9ybWF0aXZlIChzdXJ2ZXktbGlrZSk8QlI+Jmd0OzxCUj4mZ3Q7ZnJvbSB3aGF0IGkgcmVhZCwg aXQgaXMgbW9yZSAxIGxpa2UgLSBpcyBteSB1bmRlcnN0YW5kaW5nIGNvcnJlY3QgPzxCUj4mZ3Q7 PEJSPiZndDs8QlI+Jmd0OzxCUj48L0ZPTlQ+PC9QPg== Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 21:37:52 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: RE: Moving LSP ownership between control plane and management plane Date: Tue, 19 Jul 2005 22:36:27 +0100 Message-ID: <0536FC9B908BEC4597EE721BE6A353890A9F1DFF@i2km07-ukbr.domain1.systemhost.net> Thread-Topic: Moving LSP ownership between control plane and management plane Thread-Index: AcWMV6fjxUeSjrtZR7+MDzCUlMc0GQACZr5Q From: <neil.2.harrison@bt.com> To: <gnewsome@ieee.org>, <adrian@olddog.co.uk> Cc: <ccamp@ops.ietf.org> Hi George....nice to hear from you...hope all is well with you, This all sounds very reasonable to me. At the end of the day, all a signalling protocol is is a distributed agent of network management.....in the types of large BW granularity networks being discussed the NMS will always be king. Sorry if that's a problem to some folks, its just a fact for the way we see/implement it. This also follows from a good understanding of the commercials at play here. That is, there is a major difference between: - a public service context (cf PSTN), or extending a similar capability out to end-systems in the co-ps mode say, where end-systems/applications drive trail set-up (ie SVCs across a UNI). Here one hopes there could be a large population of users (think peer-peer/BB) and that the trail holding-times are not 'too long'. In this environment there could be a viable business case based on a providing a given GoS at some cost.....and it might solve some (otherwise) hard QoS/pricing problems for operators....but I digress; - a private network builder service context where one is a providing a (relatively long time-constant) traffic-engineering function to a higher layer network, ie creating its topology. This is a client/server relationship and THE starting point for this must be functional decoupling of the client and server layer networks because they will often be owned by different parties. We are talking relatively long forecast/planning cycles here and long trail holding times....esp when it comes to real physical build. Further, the potential population is rather small compared to the previous case. If folks look at whether there is a viable business case here based on a providing a given GoS at some cost in an acceptable time period I suspect they will be disappointed. What this all means is that the control-plane is not such an important topic as we move down the layered network stack towards the duct....increasingly other factors like strong NMS and OAM assume a far greater importance. This is not to say a control-plane is not useful but: - we should be clear there cannot be any common addressing and routing functions across a set of nested layer networks, which is very obvious when different parties own different layer networks....one can (and should) re-use components (if they are the right ones), but they will need to operate in different layer network spaces - signalling is a useful agent for functions like S-PVCs and resilience. Well that's our analysis anyway....if anyone knows different I'd be happy to review their commercial assumptions with folks here. regards, Neil > -----Original Message----- > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org] On=20 > Behalf Of George Newsome > Sent: 19 July 2005 12:44 > To: Adrian Farrel > Cc: ccamp@ops.ietf.org > Subject: Re: Moving LSP ownership between control plane and > management plane >=20 >=20 > Adrian, >=20 > You asked for comments on this proposal, which was presented to ITU-T=20 > Q14 experts in January of this year. >=20 > There are so many issues involved in moving a connection from the=20 > management plane to the control plane that involve intensive=20 > management action, that it is not evident that signaling need be=20 > involved at all. >=20 > As a personal opinion, it makes little sense to consider moving the=20 > connection. It makes more sense to replace the connection with one set > up via the control plane, and to then delete the original=20 > management set=20 > up connection. >=20 > Appended is an excerpt of the meeting report detailing some of the=20 > issues that need to be resolved before signaling need be considered. >=20 > Regards >=20 > George >=20 > Excerpt of the meeting report of Q14 Experts, January 2005 >=20 > WD06 (Marconi) "G.7713 Modification in order to support Circuit=20 > Migration" contains a modified version of draft G.7713 Revision as was > presented in the Berlin meeting (WD 11, i.e., TD50/3 Dec.2004). This=20 > contribution addresses SP24 (PC ? SPC) of G.7713. It proposes=20 > to extend=20 > the concept of Call/Connection setup to setup/adopt and release to=20 > release/de-adopt. In Connection Adoption both the SNC are not=20 > actually=20 > created and the LC are not actually established due to the=20 > fact that the=20 > underlying physical resources are already in place. In the=20 > Call/Connection dis-adoption SNC-LC-SNC is not actually de-allocated,=20 > only the Control Plane information associated with then is=20 > removed and=20 > the ownership is moved to Management Plane. Attributes for indicating=20 > Adoption/Dis-adoption are proposed for the INNI Connection messages. >=20 > The discussion was mainly on the Management plane and Control plane=20 > actions required before signalling is involved. The group noted that=20 > the following issues need to be addressed: >=20 > - Naming of transport resources to the control plane. Before a call=20 > can be placed under signalling control, links that are involved in the > connection need to be given control plane names. Without these, no=20 > explicit route can be formed to match the resources of the connection. >=20 > - Resource state while under dual CP and MP views. After resources=20 > are given control plane names, the resources are still viewed by the > management plane. It may be necessary to create a new state for the=20 > resource to indicate that the management plane cannot perform some=20 > actions on the connection points as the control plane is=20 > about to take over. >=20 > - Role of discovery functions (esp. CELA). After control plane names=20 > are allocated, distributed control plane functions need to be=20 > associated and communication adjacencies formed. This too is a=20 > precursor to any signalling procedures in migrating from the MP to CP. >=20 > - Similarity of migration to synchronization after CP failure and=20 > subsequent recovery. Connection control procedures might be the same=20 > in migrating a call from the MP to the CP as the situation when > the CP has=20 > failed and is recovering. Here, the network connections are=20 > already in=20 > place, but connection and call state need to be created to match it.=20 > Knowledge of the call and what the connection should be is=20 > obtained from=20 > the MP for migration, and from some reliable database in the=20 > recovery case. >=20 > - Call awareness of migration vs. connection being unaware of=20 > migration. > When connection state is being created to match an existing > connection, the connection controllers (CCs) do not require=20 > awareness of=20 > why this is happening as the context could be migration or=20 > recovery. A=20 > mechanism to create control state without affecting transport plane=20 > state is needed in the CCs. Call controllers though do need=20 > migration=20 > awareness as they need to obtain/derive call identification=20 > from the MP. >=20 > In summary, the group decided that it is premature to consider=20 > signalling procedure before the above issues, amongst other, have been > studied. However the contribution does enable the group to=20 > expose in a=20 > larger context the interaction between CP and MP for connection=20 > migration. The above issues were included in SP24 of the=20 > G.7713 Living=20 > List (see WD22). >=20 >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 19:51:04 +0000 Content-Type: Multipart/Mixed; Boundary="NextPart" Mime-Version: 1.0 To: i-d-announce@ietf.org Cc: ccamp@ops.ietf.org From: Internet-Drafts@ietf.org Subject: I-D ACTION:draft-ietf-ccamp-rsvp-te-exclude-route-04.txt Message-Id: <E1Duy6M-0002Jd-0S@newodin.ietf.org> Date: Tue, 19 Jul 2005 15:50:02 -0400 --NextPart A New Internet-Draft is available from the on-line Internet-Drafts directories. This draft is a work item of the Common Control and Measurement Plane Working Group of the IETF. Title : Exclude Routes - Extension to RSVP-TE Author(s) : A. Farrel, et al. Filename : draft-ietf-ccamp-rsvp-te-exclude-route-04.txt Pages : 34 Date : 2005-7-19 The RSVP-TE specification, "RSVP-TE: Extensions to RSVP for LSP Tunnels" (RFC 3209) and GMPLS extensions to RSVP-TE, "Generalized Multi-Protocol Label Switching (GMPLS) Signaling Resource ReserVation Protocol-Traffic Engineering (RSVP-TE) Extensions" (RFC 3473) allow abstract nodes and resources to be explicitly included in a path setup, but not to be explicitly excluded. In some networks where precise explicit paths are not computed at the head end it may be useful to specify and signal abstract nodes and resources that are to be explicitly excluded from routes. These exclusions may apply to the whole path, or to parts of a path between two abstract nodes specified in an explicit path. How Shared Risk Link Groups (SLRGs) can be excluded is also specified in this document. This document specifies ways to communicate route exclusions during path setup using RSVP-TE. A URL for this Internet-Draft is: http://www.ietf.org/internet-drafts/draft-ietf-ccamp-rsvp-te-exclude-route-04.txt To remove yourself from the I-D Announcement list, send a message to i-d-announce-request@ietf.org with the word unsubscribe in the body of the message. You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce to change your subscription settings. Internet-Drafts are also available by anonymous FTP. Login with the username "anonymous" and a password of your e-mail address. After logging in, type "cd internet-drafts" and then "get draft-ietf-ccamp-rsvp-te-exclude-route-04.txt". A list of Internet-Drafts directories can be found in http://www.ietf.org/shadow.html or ftp://ftp.ietf.org/ietf/1shadow-sites.txt Internet-Drafts can also be obtained by e-mail. Send a message to: mailserv@ietf.org. In the body type: "FILE /internet-drafts/draft-ietf-ccamp-rsvp-te-exclude-route-04.txt". NOTE: The mail server at ietf.org can return the document in MIME-encoded form by using the "mpack" utility. To use this feature, insert the command "ENCODING mime" before the "FILE" command. To decode the response(s), you will need "munpack" or a MIME-compliant mail reader. Different MIME-compliant mail readers exhibit different behavior, especially when dealing with "multipart" MIME messages (i.e. documents which have been split up into multiple messages), so check your local documentation on how to manipulate these messages. Below is the data which will enable a MIME compliant mail reader implementation to automatically retrieve the ASCII version of the Internet-Draft. --NextPart Content-Type: Multipart/Alternative; Boundary="OtherAccess" --OtherAccess Content-Type: Message/External-body; access-type="mail-server"; server="mailserv@ietf.org" Content-Type: text/plain Content-ID: <2005-7-19111430.I-D@ietf.org> ENCODING mime FILE /internet-drafts/draft-ietf-ccamp-rsvp-te-exclude-route-04.txt --OtherAccess Content-Type: Message/External-body; name="draft-ietf-ccamp-rsvp-te-exclude-route-04.txt"; site="ftp.ietf.org"; access-type="anon-ftp"; directory="internet-drafts" Content-Type: text/plain Content-ID: <2005-7-19111430.I-D@ietf.org> --OtherAccess-- --NextPart-- Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 19:50:56 +0000 Content-Type: Multipart/Mixed; Boundary="NextPart" Mime-Version: 1.0 To: i-d-announce@ietf.org Cc: ccamp@ops.ietf.org From: Internet-Drafts@ietf.org Subject: I-D ACTION:draft-ietf-ccamp-inter-domain-rsvp-te-01.txt Message-Id: <E1Duy6M-0002KY-Ag@newodin.ietf.org> Date: Tue, 19 Jul 2005 15:50:02 -0400 --NextPart A New Internet-Draft is available from the on-line Internet-Drafts directories. This draft is a work item of the Common Control and Measurement Plane Working Group of the IETF. Title : Inter domain GMPLS Traffic Engineering - RSVP-TE extensions Author(s) : A. Ayyangar, J. Vasseur Filename : draft-ietf-ccamp-inter-domain-rsvp-te-01.txt Pages : 17 Date : 2005-7-19 This document describes extensions to Generalized Multi-Protocol Label Switching (GMPLS) Resource ReserVation Protocol - Traffic Engineering (RSVP-TE) signaling required to support mechanisms for the establishment and maintenance of GMPLS Traffic Engineering (TE) Label Switched Paths (LSPs), both packet and non-packet, that traverse multiple domains. For the purpose of this document, a domain is considered to be any collection of network elements within a common realm of address space or path computation responsibility. Examples of such domains include Autonomous Systems, IGP areas and GMPLS overlay networks. A URL for this Internet-Draft is: http://www.ietf.org/internet-drafts/draft-ietf-ccamp-inter-domain-rsvp-te-01.txt To remove yourself from the I-D Announcement list, send a message to i-d-announce-request@ietf.org with the word unsubscribe in the body of the message. You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce to change your subscription settings. Internet-Drafts are also available by anonymous FTP. Login with the username "anonymous" and a password of your e-mail address. After logging in, type "cd internet-drafts" and then "get draft-ietf-ccamp-inter-domain-rsvp-te-01.txt". A list of Internet-Drafts directories can be found in http://www.ietf.org/shadow.html or ftp://ftp.ietf.org/ietf/1shadow-sites.txt Internet-Drafts can also be obtained by e-mail. Send a message to: mailserv@ietf.org. In the body type: "FILE /internet-drafts/draft-ietf-ccamp-inter-domain-rsvp-te-01.txt". NOTE: The mail server at ietf.org can return the document in MIME-encoded form by using the "mpack" utility. To use this feature, insert the command "ENCODING mime" before the "FILE" command. To decode the response(s), you will need "munpack" or a MIME-compliant mail reader. Different MIME-compliant mail readers exhibit different behavior, especially when dealing with "multipart" MIME messages (i.e. documents which have been split up into multiple messages), so check your local documentation on how to manipulate these messages. Below is the data which will enable a MIME compliant mail reader implementation to automatically retrieve the ASCII version of the Internet-Draft. --NextPart Content-Type: Multipart/Alternative; Boundary="OtherAccess" --OtherAccess Content-Type: Message/External-body; access-type="mail-server"; server="mailserv@ietf.org" Content-Type: text/plain Content-ID: <2005-7-19120530.I-D@ietf.org> ENCODING mime FILE /internet-drafts/draft-ietf-ccamp-inter-domain-rsvp-te-01.txt --OtherAccess Content-Type: Message/External-body; name="draft-ietf-ccamp-inter-domain-rsvp-te-01.txt"; site="ftp.ietf.org"; access-type="anon-ftp"; directory="internet-drafts" Content-Type: text/plain Content-ID: <2005-7-19120530.I-D@ietf.org> --OtherAccess-- --NextPart-- Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 19:31:44 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: RE: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON Lexicography Date: Tue, 19 Jul 2005 15:30:44 -0400 Message-ID: <CA8EB93B28BD0F4C88392699600BD65566730F@xmb-rtp-211.amer.cisco.com> Thread-Topic: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON Lexicography Thread-Index: AcWMmFsziUnUgjpdQwiSfQ/jbIa2KQ== From: "Scott Brim \(sbrim\)" <sbrim@cisco.com> To: <ccamp@ops.ietf.org> On Tuesday, July 19, 2005 06:59 AM, Adrian Farrel allegedly wrote: > While the definition of GMPLS terms would be helpful for > understanding GMPLS terms in a GMPLS context, that is not the > purpose of this document. =20 It would be polite to tell them where to find the definions :-) Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 19:21:18 +0000 Message-ID: <011d01c58c8e$57a1e6c0$0601a8c0@pc6> Reply-To: "Tom Petch" <nwnetworks@dial.pipex.com> From: "Tom Petch" <nwnetworks@dial.pipex.com> To: "Adrian Farrel" <adrian@olddog.co.uk>, <ccamp@ops.ietf.org> Subject: Re: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON Lexicography Date: Tue, 19 Jul 2005 20:18:16 +0200 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Adrian I think it should be " - CCAMP has primary responsibility for the descriptions of GMPLS terms Correct " even if that is not what SG15 said. Primarily (sic) CCAMP has responsibility for the specification of GMPLS, not just its terms:-) Tom Petch ----- Original Message ----- From: "Adrian Farrel" <adrian@olddog.co.uk> To: <ccamp@ops.ietf.org> Cc: "'Kireeti Kompella'" <kireeti@juniper.net>; <zinin@psg.com>; "Bill Fenner" <fenner@research.att.com>; "Scott Bradner" <sob@harvard.edu> Sent: Tuesday, July 19, 2005 1:59 PM Subject: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON Lexicography > Hi, > > Please comment on this draft response liaison which I intend to send on > Monday 25th July. > You can see the original incoming liaison at > http://www.olddog.co.uk/incoming.htm > > Thanks, > Adrian > ===== > To: ITU-T SG15 Questions 12 and 14 > From: IETF CCAMP Working Group > Subject: Response to your liaison on GMPLS/ASON Lexicography > For: Action > Deadline: August 31st 2005 > > The CCAMP Working Group would like to thank SG15 and especially Questions > 12 and 14 for the time and effort they put into providing input and > feedback on the GMPLS/ASON lexicography Internet-Draft that is being > developed by CCAMP. Much of the material supplied as direct comments or as > mark-ups to the Internet-Draft has been incorporated into the latest > revision which is available at > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-gmpls-ason-lexicography-03. txt > > In your liaison you state some assumptions which we would like to comment > on as follows. > > - The purpose of the document is to enable those familiar with ASON > to understand GMPLS terminology in an ASON context > > Correct. > > - GMPLS terms are described informally in the document in a way that > is generally consistent with GMPLS (existing RFCs and work in progress) > > While the definition of GMPLS terms would be helpful for understanding > GMPLS terms in a GMPLS context, that is not the purpose of this document. > > - Based on these descriptions, interpretations of these GMPLS terms > in ASON terminology are provided > > It is intended that the interpretation of the GMPLS terms is not based on > these descriptions, but based on the full meaning of the GMPLS terms. The > descriptions of the GMPLS terms are provided in this Internet-Draft for > context and a brief summary. > > - CCAMP has primarily responsibility for the descriptions of GMPLS terms > > Correct. > > - Q12/15 has responsibility to provide input to the GMPLS-ASON > interpretations (based on the descriptions in this draft) > > We welcome this assumption of responsibility by Q12/15 to provide input to > the Internet-Draft. While the Internet-Draft remains under the overall > editorial control of the CCAMP working group, we hope to be able to use > the text supplied by Q12/15 with only editorial changes. > > > Additionally, we welcome the pledge to continue work on this > Internet-Draft through correspondence on the Q12/15 mailing list, and > thank Q12/15 for assigning Ben Mack-Crane as the SG15 representative on > this matter. Ben has proved very helpful in clarifying and developing > comments received from SG15. > > > We have had several email exchanges with Ben about a few of the specific > comments made in your review of the 02 revision of the Internet-Draft. > Although he has taken these comments to the Q12/15 mailing list no > response has been received, perhaps due to pressure of time. We are taking > the opportunity to restate several questions here for your convenience and > would appreciate your answers as part of the response to this liaison. > > 1. Definition of a Resource (section 3.2) > You have supplied us with the simple text that says... > In ASON applications: > A resource refers only to a link connection (e.g. VC-11, > VC-4 etc.). > > We would like to clarify this because we don't understand how > the original text is wrong. It used to say... > ITU-T [should be ASON] terms for resource: > - Connection point (cp) in the context of link discovery > and resource management (allocation, binding into > cross-connects, etc.); > - Link connection or trail termination in the context of > routing, path computation and signaling. > > Recall that we are talking about the ASON terms for what GMPLS > calls a resource. We think the following... > In ASON a link connection is an association of output cp on one > side of the link and input cp on the other side of the link. In > GMPLS by resource we mean in most of the cases the local end of > the resource (which is cp in ASON terminology). This is true in > the context of link discovery. > > More importantly, this is also true in the context of a > signaling application. Specifically, when a GMPLS signaling > sub-system requests allocation of a resource, the GMPLS > Resource Manager allocates only the local end of the > resource. This happens on both sides of the link, that is, we > allocate input and output resources. This contrasts with ASON, > where there is a much more versatile LRM, and the connection > manager (signaling application) allocates link connections > only on the output side, leaving LRM to coordinate the > allocation with the neighbor. > > However, in GMPLS TE advertisements we account only for > outgoing resources, that is, we don't advertise incoming > resources on the assumption that they match the outgoing > resources for the discovered TE links. > > Bottom line: we believe that we should retain the previous > definitions, and have not made this change in the new > revision. We would like to discuss this point further with > Q12/15. > > 2. Link ends and Data links (Section 3.5 - was section 3.4) > > 2a.Although we understand that ASON does not require the concept > of a link end, GMPLS does. Therefore we should provide some > form of language to help people do the mapping. > > The proposed new text from Q12/15 removes all reference to > link ends from the ASON section, and we feel this is a > mistake and propose the following text: > In ASON, a GMPLS unidirectional data link end is a > collection of connection points from the same client > layer that are supported by a single trail termination > (access point). > > This text depends on the fact that a link end is described as > a set of resources that transfer traffic to a neighbor. > > 2b.The new text from Q12/15 gives us... > A GMPLS data link is an ASON link supported by a single > server trail. > > We are not sure that we understand this correctly. It > appears to say that a unidirectional data link could be > supported by multiple trail terminations (access points). > This seems to suggest that a client would have to make a > choice about where to send data, which seems to imply it > really has to choose between data links! We are confused! > > We think some of this problem may come from the *need* in > GMPLS to identify data link ends. There may be some value in > stressing that a GMPLS link end is supported by exactly one > trail termination. When several trail terminations are > involved, we can still talk about single TE link (as a > bundle), but each component link will be a distinct data > link. We think that in ASON the SNPP (TE link in GMPLS > language) could be supported by multiple terminations. > > 3. Link interfaces (section 3.6 - was 3.5) > The new text from Q12/15 gives us... > A GMPLS interface is "all the stuff between a physical > interface and a matrix in an ASON network element." More > precisely, it is a trail termination function and adaptation > function for which adapted client layer connection points > are bound to a matrix. > > We think "physical interface" may be ambiguous in the multi- > layer context. We think the "stuff" should be between a link > and a matrix, since we are talking about links in distinct > layers. > > > In addition to your feedback on the specific points above, we would > welcome your continued comments and review of the Internet-Draft. > > Regards, > Adrian Farrel and Kireeti Kompella > IETF CCAMP Working Group Co-chairs > > Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 18:43:29 +0000 Date: Tue, 19 Jul 2005 11:41:39 -0700 (PDT) From: Arthi Ayyangar <arthi@juniper.net> To: Adrian Farrel <adrian@olddog.co.uk> cc: Dimitri.Papadimitriou@alcatel.be, Igor Bryskin <ibryskin@movaz.com>, ccamp@ops.ietf.org Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Message-ID: <20050719114004.L71817@zircon.juniper.net> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII > Say rather, "can be admitted over a TE link supported by LSP segments". > That is, a segment can only admit one LSP, but a bundle of segments > advertised as a TE link, can admit multiple LSPs. -----> Yes, this more accurate. thanks, -arthi Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 16:57:29 +0000 Message-Id: <6.0.0.20.2.20050720013441.03d8b600@129.60.5.226> Date: Wed, 20 Jul 2005 01:58:21 +0900 To: <ccamp@ops.ietf.org> From: "inoue.ichiro@lab.ntt.co.jp" <inoue.ichiro@lab.ntt.co.jp> Subject: Re: no more milestones? Cc: "Thomas D. Nadeau" <tnadeau@cisco.com>, <zinin@psg.com>, "Bill Fenner" <fenner@research.att.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Content-Transfer-Encoding: 7bit Hi, I agree that base signalling and routing GMPLS protocols need to be worked out, and at the same time I would like to see GMPLS applications in operational networks. I remember about the vote on specific items (to-be-rechartered items) which included basic signalling and routing aspects as well as some extended applications. At that time, I felt we had a long and good discussion already about what would be on the revised charter and it was very close to decide. Now, though (because?) I don't know what process is done and further needed beyond what we already went through and saw, I appreciate if we can see new milestones reflecting the vote and recent discussions so that we can determine in Paris hopefully without, but with some more discussion if needed. Thanks. Ichiro Inoue > >> I would like to see: > >> - a refinement of the charter to focus the next work items > >> - a year's worth of milestones to ensure we stay focused. > >> > >> On the other hand, we should aim to close down CCAMP and move the > >> work to > >> other or new working groups. > >> > > > > coupling both sentences does it mean in your view > > > > 1) define several short-term milestones for refining topics that > > have been > > already initiated and requires further well scoped attention > > > > 2) (brand) new topics should target new working groups - as already > > initiated with L1VPN for inst. - or existing working groups > > > > nevertheless, i do not see why this automatically means CCAMP can not > > remain the focal point concerning *base* GMPLS signaling and routing > > protocols > Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 14:51:24 +0000 Content-Type: Multipart/Mixed; Boundary="NextPart" Mime-Version: 1.0 To: i-d-announce@ietf.org Cc: ccamp@ops.ietf.org From: Internet-Drafts@ietf.org Subject: I-D ACTION:draft-ietf-ccamp-rsvp-restart-ext-03.txt Message-Id: <E1DutQ1-0003cc-UO@newodin.ietf.org> Date: Tue, 19 Jul 2005 10:50:01 -0400 --NextPart A New Internet-Draft is available from the on-line Internet-Drafts directories. This draft is a work item of the Common Control and Measurement Plane Working Group of the IETF. Title : Extensions to GMPLS RSVP Graceful Restart Author(s) : A. Satyanarayana Filename : draft-ietf-ccamp-rsvp-restart-ext-03.txt Pages : 24 Date : 2005-7-19 This document describes extensions to the RSVP Graceful Restart mechanisms defined in RFC 3473. The extensions enable the recovery of RSVP signaling state based on the Path message last sent by the node being restarted. Previously defined Graceful Restart mechanisms, also called recovery from nodal faults, permit recovery of signaling state from adjacent nodes when the data plane has retained the associated forwarding state across a restart. These mechanisms do not fully support signaling state recovery on ingress nodes or recovery of all RSVP objects. The presented extensions use the RSVP Hello extensions defined in RFC 3209, and extensions for state recovery on nodal faults defined in RFC 3473. With the presented extensions the restarting node can recover all previously transmitted Path state including the Explicit Route Object and the downstream (outgoing) interface identifiers. The extensions can also be used to recover signaling state after the restart of an ingress node. The extensions optionally support the use of Summary Refresh, defined in RFC 2961, to reduce the number of messages exchanged during the Recovery Phase when the restarting node has recovered signaling state locally for one or more LSPs. A URL for this Internet-Draft is: http://www.ietf.org/internet-drafts/draft-ietf-ccamp-rsvp-restart-ext-03.txt To remove yourself from the I-D Announcement list, send a message to i-d-announce-request@ietf.org with the word unsubscribe in the body of the message. You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce to change your subscription settings. Internet-Drafts are also available by anonymous FTP. Login with the username "anonymous" and a password of your e-mail address. After logging in, type "cd internet-drafts" and then "get draft-ietf-ccamp-rsvp-restart-ext-03.txt". A list of Internet-Drafts directories can be found in http://www.ietf.org/shadow.html or ftp://ftp.ietf.org/ietf/1shadow-sites.txt Internet-Drafts can also be obtained by e-mail. Send a message to: mailserv@ietf.org. In the body type: "FILE /internet-drafts/draft-ietf-ccamp-rsvp-restart-ext-03.txt". NOTE: The mail server at ietf.org can return the document in MIME-encoded form by using the "mpack" utility. To use this feature, insert the command "ENCODING mime" before the "FILE" command. To decode the response(s), you will need "munpack" or a MIME-compliant mail reader. Different MIME-compliant mail readers exhibit different behavior, especially when dealing with "multipart" MIME messages (i.e. documents which have been split up into multiple messages), so check your local documentation on how to manipulate these messages. Below is the data which will enable a MIME compliant mail reader implementation to automatically retrieve the ASCII version of the Internet-Draft. --NextPart Content-Type: Multipart/Alternative; Boundary="OtherAccess" --OtherAccess Content-Type: Message/External-body; access-type="mail-server"; server="mailserv@ietf.org" Content-Type: text/plain Content-ID: <2005-7-19101112.I-D@ietf.org> ENCODING mime FILE /internet-drafts/draft-ietf-ccamp-rsvp-restart-ext-03.txt --OtherAccess Content-Type: Message/External-body; name="draft-ietf-ccamp-rsvp-restart-ext-03.txt"; site="ftp.ietf.org"; access-type="anon-ftp"; directory="internet-drafts" Content-Type: text/plain Content-ID: <2005-7-19101112.I-D@ietf.org> --OtherAccess-- --NextPart-- Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 14:24:57 +0000 Message-ID: <00b201c58c6d$96e0dc90$7a1810ac@movaz.com> From: "Igor Bryskin" <ibryskin@movaz.com> To: "Arthi Ayyangar" <arthi@juniper.net> Cc: <ccamp@ops.ietf.org> Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Date: Tue, 19 Jul 2005 10:24:35 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Arthi, One more comment. I think the draft should give a clear and unambiguous answer to how e2e LSP and a stitching LSP (S-LSP) relate from the switching capability point of view. Specifically, should they belong to the same network layer, should they share the same switching type or maybe there are no limitations at all, and you can stitch your e2e LSP pretty much with anything you got, provided that this "anything" has enough bandwidth. Let me give you a specific example. Is it possible to stitch a PSC1 e2e LSP with a PSC2 S-LSP? You probably remember our discussion with Kireeti in Minneapolis (and you were also part of it), and he said (and I 100% agree with him) that although it is doable (you can do anything you want with the label stack), it is not right and this should be logically prohibited. What you should do in this case is to nest the e2e LSP within the PSC2 LSP (that is, use it as H-LSP) even in case it consumes the entire PSC2 LSP bandwidth. It is more clear in non-PSC cases. Although it is possible to "stitch" an e2e VC12 LSP with a VC4 S-LSP, I could not imagine that a network operator would ever agree with such solution. He is likely to say " What ? If the service is supposed to stay for 12 month, than for entire year I pretty much burry my 62 VC12 channels - I cannot use them even conceptually in case when, say, on the next day I'll have to provide another 62 VC12 services that could've used these wasted resources" On the other hand I know plenty GMPLS people who would say: "Sure, no problem. Of course we can stitch a PSC e2e LSP with lambda S-LSP: If somebody wants to do a trick stupid as it is, why should we prevent him from doing so?" The bottom line is that a discussion like this in the draft and consensus on how we see the stitching IMO would be extremely valuable for the GMPLS multi-layer story, and I don't see a reason why this could not be achieved in the frame of this work. Igor ----- Original Message ----- From: "Arthi Ayyangar" <arthi@juniper.net> To: "Igor Bryskin" <ibryskin@movaz.com> Cc: <i-d-announce@ietf.org>; <Internet-Drafts@ietf.org>; <ccamp@ops.ietf.org> Sent: Monday, July 18, 2005 2:51 PM Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > Hi Igor, > > Please see inline. > > > 1. Why are saying that LSP Stitching is a private case of LSP Hierarchy? > > > there is more differences than similarities: > > The differences are: > > 1) In case of H-LSP there is a data plane adjacency, while in case of S-LSP > > there is none (as you correctly pointed out); > > 2) In case of H-LSP there is an adaptation in data plane (label push/pop for > > PSC), while in case of S-LSP there is none - just simple cross-connecting > > ( label swap) as in case of two "native " e2e LSP adjacent segments ; > > 3) H-LSP could be used by many e2e LSPs, while S-LSP could be used by > > exactly one e2e LSP > > 4) Signaling is different - there is no label negotiation in stitching > > 5) H-LSP is used as a "true" data link, specifically there is a resource > > allocation on the H-LSP edges, while in case of S-LSP there is none > > 6) From MLN point of view, H-LSP is created in a server (lower) layer, while > > the S-LSP is created in the client (same as e2e LSP) layer. > > > > There are two similarities that I can think of: > > 1) There is a signaling and possibly routing (see below) adjacencies between > > the ends; > > 2) Both H-LSP and S-LSP could be advertised as separate TE links or as TE > > bundles > > > > I would recommend to dedicate a paragraph and enlist there similarities and > > differencies > -------> Based on the discussions on the list over the last revision, the > message that the authors received was that we do not want to go about such > a lengthy discussion of similarities and differences (NOTE that the draft > already does clearly highlight them where applicable). Instead we want > this ID to simply explain how LSP stitching functions. In other words the > idea was to make this ID a complete document by itself. But since it does > borrow concepts from the LSP hierarchy ID, just state the concepts that > are applicable or inapplicable. > > > 2. Why are you saying that a TE Link based on S-LSP can be used for exactly > > one e2e LSP? Parallel S-LSPs could be advertised as a single TE link > > (bundle, see above) and hence can accomadate several e2e LSPs. > ----> Igor, the draft does talk about this Bundling case. > > > 3. You are saying that S-LSP does not have a routing peering. Actually, in > > this respect it is no different from H-LSP: if it is advertised as a TE link > > into the same TE domain that was used for S-LSP creation (unlikely IMO > > scenario) than it does not require the routing adjacency (in other words, it > > is an FA according to LSP-HIER definition), otherwise, it IS NOT and FA and > > does require the direct routing peering in the domain it is advertised to > > make it useful as a TE link in this domain > -----> I am sorry, but I am missing your point here completely. Are you > saying that the statement "the end points of an LSP segment do not have a > routing adjacency", is incorrect or are you saying that "this is > obvious" ? It is unclear to me what exactly your argument is. Please > clarify. > > thanks, > -arthi > > > > > Cheers, > > Igor > > > > > > > > ----- Original Message ----- > > From: <Internet-Drafts@ietf.org> > > To: <i-d-announce@ietf.org> > > Cc: <ccamp@ops.ietf.org> > > Sent: Friday, July 15, 2005 3:50 PM > > Subject: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > > > > > > > A New Internet-Draft is available from the on-line Internet-Drafts > > directories. > > > This draft is a work item of the Common Control and Measurement Plane > > Working Group of the IETF. > > > > > > Title : Label Switched Path Stitching with Generalized MPLS Traffic > > Engineering > > > Author(s) : A. Ayyangar, J. Vasseur > > > Filename : draft-ietf-ccamp-lsp-stitching-01.txt > > > Pages : 19 > > > Date : 2005-7-15 > > > > > > In certain scenarios, there may be a need to combine together two > > > different Generalized Multi-Protocol Label Switching (GMPLS) Label > > > Switched Paths (LSPs) such that in the data plane, a single end-to- > > > end (e2e) LSP is achieved and all traffic from one LSP is switched > > > onto the other LSP. We will refer to this as "LSP stitching". This > > > document covers cases where: a) the node performing the stitching > > > does not require configuration of every LSP pair to be stitched > > > together b) the node performing the stitching is not the egress of > > > any of the LSPs c) LSP stitching not only results in an end-to-end > > > LSP in the data plane, but there is also a corresponding end-to-end > > > LSP (RSVP session) in the control plane. It might be possible to > > > configure a GMPLS node to switch the traffic from an LSP for which it > > > is the egress, to another LSP for which it is the ingress, without > > > requiring any signaling or routing extensions whatsoever, completely > > > transparent to other nodes. This will also result in LSP stitching > > > in the data plane. However, this document does not cover this > > > scenario of LSP stitching. > > > > > > A URL for this Internet-Draft is: > > > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt > > > > > > To remove yourself from the I-D Announcement list, send a message to > > > i-d-announce-request@ietf.org with the word unsubscribe in the body of the > > message. > > > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > > > to change your subscription settings. > > > > > > > > > Internet-Drafts are also available by anonymous FTP. Login with the > > username > > > "anonymous" and a password of your e-mail address. After logging in, > > > type "cd internet-drafts" and then > > > "get draft-ietf-ccamp-lsp-stitching-01.txt". > > > > > > A list of Internet-Drafts directories can be found in > > > http://www.ietf.org/shadow.html > > > or ftp://ftp.ietf.org/ietf/1shadow-sites.txt > > > > > > > > > Internet-Drafts can also be obtained by e-mail. > > > > > > Send a message to: > > > mailserv@ietf.org. > > > In the body type: > > > "FILE /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt". > > > > > > NOTE: The mail server at ietf.org can return the document in > > > MIME-encoded form by using the "mpack" utility. To use this > > > feature, insert the command "ENCODING mime" before the "FILE" > > > command. To decode the response(s), you will need "munpack" or > > > a MIME-compliant mail reader. Different MIME-compliant mail readers > > > exhibit different behavior, especially when dealing with > > > "multipart" MIME messages (i.e. documents which have been split > > > up into multiple messages), so check your local documentation on > > > how to manipulate these messages. > > > > > > > > > Below is the data which will enable a MIME compliant mail reader > > > implementation to automatically retrieve the ASCII version of the > > > Internet-Draft. > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 14:14:01 +0000 Message-ID: <0e6c01c58c6c$66d02200$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <Dimitri.Papadimitriou@alcatel.be> Cc: <ccamp@ops.ietf.org>, "Thomas D. Nadeau" <tnadeau@cisco.com>, <zinin@psg.com>, "Bill Fenner" <fenner@research.att.com> Subject: Re: no more milestones? Date: Tue, 19 Jul 2005 15:15:50 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi Dimitri, > much thanks for your reply - just to clarify, in the below when you state: > >> "If, on the other hand, we are not going to add this stuff to the charter, >> it would be very important for us to be clear about this so that the people >> working in these areas know that they must take their work to another WG or >> BOF." > > when you state "... to another WG" do you mean an existing WG, and in such > a case do you have any specific WG in mind ? any clarification much > appreciated I have nothing specific in mind. Just to note that that which is not CCAMP is something else. A Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 14:03:37 +0000 Message-ID: <0e5301c58c6a$f1d1d5d0$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: "George Newsome" <gnewsome@ieee.org> Cc: <ccamp@ops.ietf.org> Subject: Re: Moving LSP ownership between control plane and management plane Date: Tue, 19 Jul 2005 15:02:24 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Thanks George, Useful input. > You asked for comments on this proposal, which was presented to ITU-T > Q14 experts in January of this year. Can you confirm that there were plenty of providers represented at the meeting. > There are so many issues involved in moving a connection from the > management plane to the control plane that involve intensive management > action, that it is not evident that signaling need be involved at all. Interesting. It is certainly valuable to have a list of these issues out in the open. I have made some comments in the appended meeting report. If the control plane is to take over the connection then we assume that signaling state must be built so that the control plane could then modify or tear down the connection. how would you propose that this be done without involving signaling? > As a personal opinion, it makes little sense to consider moving the > connection. It makes more sense to replace the connection with one set > up via the control plane, and to then delete the original management set > up connection. The question arrises of how this would be done in a network with limited resources. That is, if it is possible to set up a parallel conneciton then the method you describe would be fine, but if there are not enough parallel resources then would you propose that a lambda (for example) could be used for the management-established connection and for the control-plane-owned connection at the same time? This might present some difficulties as well. Do you have an opinion on my main question which is whether this tyoe of change-over would ever be done in service? Thanks, Adrian > Excerpt of the meeting report of Q14 Experts, January 2005 > > WD06 (Marconi) "G.7713 Modification in order to support Circuit > Migration" contains a modified version of draft G.7713 Revision as was > presented in the Berlin meeting (WD 11, i.e., TD50/3 Dec.2004). This > contribution addresses SP24 (PC ? SPC) of G.7713. It proposes to extend > the concept of Call/Connection setup to setup/adopt and release to > release/de-adopt. In Connection Adoption both the SNC are not actually > created and the LC are not actually established due to the fact that the > underlying physical resources are already in place. In the > Call/Connection dis-adoption SNC-LC-SNC is not actually de-allocated, > only the Control Plane information associated with then is removed and > the ownership is moved to Management Plane. Attributes for indicating > Adoption/Dis-adoption are proposed for the INNI Connection messages. > > The discussion was mainly on the Management plane and Control plane > actions required before signalling is involved. The group noted that the > following issues need to be addressed: > > - Naming of transport resources to the control plane. Before a call can > be placed under signalling control, links that are involved in the > connection need to be given control plane names. Without these, no > explicit route can be formed to match the resources of the connection. This is pretty obvious. But actually, this is an issue with moving the resources from one plane to another (or having them present in both planes). It is certainly a prerequisite for moving the connection, but it is not functionally an element of moving the connection. The draft, as I read it, assumes that these operations have been completed and then asks "how do I move the connection from one plane to the other?" > - Resource state while under dual CP and MP views. After resources are > given control plane names, the resources are still viewed by the > management plane. It may be necessary to create a new state for the > resource to indicate that the management plane cannot perform some > actions on the connection points as the control plane is about to take over. Yes. There is certainly scope for a race condition here. I believe that the I-D proposes to (try to) make the transfer an atomic action from the point of view of the change-over. Note, however, that it is wrong to assume that a resource belongs only to the control or to the management plane at any moment in time. This may be true with respect to the ability to provision/control connections, but resources that are available for use by the control plane can still be managed through the management plane. Thus, you take a laser out of service through the management plane, not through the control plane. > - Role of discovery functions (esp. CELA). After control plane names > are allocated, distributed control plane functions need to be associated > and communication adjacencies formed. This too is a precursor to any > signalling procedures in migrating from the MP to CP. Note to CCAMP: CELA is Control Entity Logical Adjacency. This too is "obvious". But this discussion seems to be wide of the mark. Before deploying a control plane, it is clear that this type of operation needs to take place. Since (within the context of CCAMP) a control plane *is* going to be deployed, we can assume this operation and move on to the discussion in the draft. > - Similarity of migration to synchronization after CP failure and > subsequent recovery. Connection control procedures might be the same in > migrating a call from the MP to the CP as the situation when the CP has > failed and is recovering. Here, the network connections are already in > place, but connection and call state need to be created to match it. > Knowledge of the call and what the connection should be is obtained from > the MP for migration, and from some reliable database in the recovery case. This is a valid point. Control plane failure, however, is unlikely to be total. That is, it would be unusual for more than one control plane link or node to fail. Thus, in the control plane, the "reliable database" is distributed and recovery can follow the procedures outlined in draft-ietf-ccamp-rsvp-restart-ext-03.txt It would certainly be worth considering whether the procedures used for control plane recovery could be applied to transfer of ownership from the management to control plane. at the moment, I suspect that they could not because the restart extensions assume a specific period during which signaling is associated with recovery. We would also need to look at the reverse process (transfer from CP to MP) that is not possible to cover by CP restart procedures. > - Call awareness of migration vs. connection being unaware of migration. > When connection state is being created to match an existing > connection, the connection controllers (CCs) do not require awareness of > why this is happening as the context could be migration or recovery. A > mechanism to create control state without affecting transport plane > state is needed in the CCs. Call controllers though do need migration > awareness as they need to obtain/derive call identification from the MP. I disagree. Connection controllers must be able to distinguish between resources that are already in use and cannot be assigned to a new connection, and those that are already in use and can be assigned to a new connection because the connection is replacing an existing connection. If this were happening entirely within the control plane we would use resource sharing, but when the operation spans the control and management planes resource ownership/sharing is more complex and the conneciotn controller needs to be able to issue the approriate instrucitons for the assignment of the resources. On the other hand, call controllers do not need awareness of the details of the realisation of the connections. > In summary, the group decided that it is premature to consider > signalling procedure before the above issues, amongst other, have been > studied. However the contribution does enable the group to expose in a > larger context the interaction between CP and MP for connection > migration. The above issues were included in SP24 of the G.7713 Living > List (see WD22). Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 13:50:02 +0000 To: "Adrian Farrel" <adrian@olddog.co.uk> Cc: <Dimitri.Papadimitriou@alcatel.be>, <ccamp@ops.ietf.org>, "Thomas D. Nadeau" <tnadeau@cisco.com>, <zinin@psg.com>, "Bill Fenner" <fenner@research.att.com> From: Dimitri.Papadimitriou@alcatel.be Subject: Re: no more milestones? MIME-Version: 1.0 Date: Tue, 19 Jul 2005 15:49:05 +0200 Message-ID: <OF7F19D94A.47A82D28-ONC1257043.004BE752-C1257043.004BE7CF@netfr.alcatel.fr> Content-type: text/plain; charset=us-ascii hi adrian, much thanks for your reply - just to clarify, in the below when you state: "If, on the other hand, we are not going to add this stuff to the charter, it would be very important for us to be clear about this so that the people working in these areas know that they must take their work to another WG or BOF." when you state "... to another WG" do you mean an existing WG, and in such a case do you have any specific WG in mind ? any clarification much appreciated thanks, - dimitri. --- Hi Dimitri, > > I would like to see: > > - a refinement of the charter to focus the next work items > > - a year's worth of milestones to ensure we stay focused. > > > > On the other hand, we should aim to close down CCAMP and move the work to > > other or new working groups. > > coupling both sentences does it mean in your view > > 1) define several short-term milestones for refining topics that have been > already initiated and requires further well scoped attention > > 2) (brand) new topics should target new working groups - as already > initiated with L1VPN for inst. - or existing working groups > > nevertheless, i do not see why this automatically means CCAMP can not > remain the focal point concerning *base* GMPLS signaling and routing > protocols To answer your question: yes, I would like to see both. We need some short-term milestones to guide and refine our work on existing topics. For example, to complete the inter-domain work, possibly including issues of protection and restoration. But there are also some topics that are broadly covered by the CCAMP scope, where we would benefit greatly by having a specific description on the charter, and detailed milestones. I am thinking about L2SC, MRN, MPLS/GMPLS interworking as discussed on the mailing list last November. While we could sort of take on these things under the vague scope of the current charter, I would prefer for us to be open, clear and focused. And this covers your last point. I, too, don't see why CCAMP should not remain the focal point for GMPLS signaling and routing protocols. However, to do so in an efficient and proper manner, we should add the appropriate milestones to the charter. If, on the other hand, we are not going to add this stuff to the charter, it would be very important for us to be clear about this so that the people working in these areas know that they must take their work to another WG or BOF. Thanks, Adrian Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 13:37:26 +0000 MIME-Version: 1.0 To: ccamp@ops.ietf.org Cc: dpapadimitriou@psg.com From: Dimitri.Papadimitriou@alcatel.be Subject: Feedback on draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt Date: Tue, 19 Jul 2005 15:36:19 +0200 Message-ID: <OFC9875A24.B2B17922-ONC1257043.004ABC1B-C1257043.004ABC89@netfr.alcatel.fr> Content-type: text/plain; charset=us-ascii ccamp'ers, all, as already stated on this list we have put together our thoughts and produced the above mentioned document; its purpose is to initiate DISCUSSION on the following points: - are the proposed GMPLS for Ethernet deployment scenarios representative enough of the user community's expectations ? - if yes, is the interest in GMPLS for Ethernet in-/out-side the CCAMP WG community strong enough to consider the corr. GMPLS signaling and routing requirements (second part of the document) as part of the CCAMP WG effort ? - if not, would you let us know why ? much thanks for your feedback - dimitri. Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 12:35:51 +0000 Date: Tue, 19 Jul 2005 21:35:28 +0900 (JST) From: Kenichi Nagami <nagami@wide.ad.jp> Message-Id: <200507191235.j6JCZS4p013108@sh.wide.ad.jp> To: ccamp@ops.ietf.org, otani@kddilabs.jp Subject: Re: GMPLS OSPF-TE MIB Hi tomo, This work is interesting and important from management point of view. I have questions and comments. - Does this MIB support for only GMPLS OSPF-TE or both GMPLS OSPF-TE and OSPF-TE (RFC3630) ? - I think INDEX of the ospfTeLsdbEntry need to add Area ID. Because OSPF opaque LSA type 10 is "area local scope". - There are a log of error for MIB definition. Could you fix them? For example, 1) "--" is needed "in xxx" current DESCRIPTION "Initial version. Published as RFC xxxx." -- RFC-editor pls fill in xxx" changed DESCRIPTION "Initial version. Published as RFC xxxx." -- RFC-editor pls fill -- in xxx" 2) need to "," in ospfTeMainGroup OBJECT-GROUP ospfTeSwitchingType ospfTeEncoding are changed to ospfTeSwitchingType, ospfTeEncoding, Regards, Ken Nagami Intec Netcore > >At 08:52 05/07/15, Tomohiro Otani wrote: >Hi everyone, >We posted a new draft about GMPLS OSPF-TE MIB as follows. >http://www.ietf.org/internet-drafts/draft-otani-ccamp-gmpls-ospf-mib-00.txt >We welcome your comments and feedbacks, pls. >And please make a comment whether this is within ccamp WG or not (OSPF-WG?). >We believe that this is a very significant definition of OSPF-TE MIB, >being complementary to other GMPLS related MIBs (TE/LSR/LMP mibs) in order >to deploy GMPLS control plane in the actual environment with simple but >reliable network management. >With best regards, >tomo Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 12:26:43 +0000 Message-ID: <0df301c58c5d$5e1a5b80$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <ccamp@ops.ietf.org> Cc: "'Kireeti Kompella'" <kireeti@juniper.net>, <zinin@psg.com>, "Bill Fenner" <fenner@research.att.com>, "Scott Bradner" <sob@harvard.edu> Subject: Proposed response to ITU-T SG15 Questions 12 and 14 on ASON Lexicography Date: Tue, 19 Jul 2005 12:59:12 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi, Please comment on this draft response liaison which I intend to send on Monday 25th July. You can see the original incoming liaison at http://www.olddog.co.uk/incoming.htm Thanks, Adrian ===== To: ITU-T SG15 Questions 12 and 14 From: IETF CCAMP Working Group Subject: Response to your liaison on GMPLS/ASON Lexicography For: Action Deadline: August 31st 2005 The CCAMP Working Group would like to thank SG15 and especially Questions 12 and 14 for the time and effort they put into providing input and feedback on the GMPLS/ASON lexicography Internet-Draft that is being developed by CCAMP. Much of the material supplied as direct comments or as mark-ups to the Internet-Draft has been incorporated into the latest revision which is available at http://www.ietf.org/internet-drafts/draft-ietf-ccamp-gmpls-ason-lexicography-03.txt In your liaison you state some assumptions which we would like to comment on as follows. - The purpose of the document is to enable those familiar with ASON to understand GMPLS terminology in an ASON context Correct. - GMPLS terms are described informally in the document in a way that is generally consistent with GMPLS (existing RFCs and work in progress) While the definition of GMPLS terms would be helpful for understanding GMPLS terms in a GMPLS context, that is not the purpose of this document. - Based on these descriptions, interpretations of these GMPLS terms in ASON terminology are provided It is intended that the interpretation of the GMPLS terms is not based on these descriptions, but based on the full meaning of the GMPLS terms. The descriptions of the GMPLS terms are provided in this Internet-Draft for context and a brief summary. - CCAMP has primarily responsibility for the descriptions of GMPLS terms Correct. - Q12/15 has responsibility to provide input to the GMPLS-ASON interpretations (based on the descriptions in this draft) We welcome this assumption of responsibility by Q12/15 to provide input to the Internet-Draft. While the Internet-Draft remains under the overall editorial control of the CCAMP working group, we hope to be able to use the text supplied by Q12/15 with only editorial changes. Additionally, we welcome the pledge to continue work on this Internet-Draft through correspondence on the Q12/15 mailing list, and thank Q12/15 for assigning Ben Mack-Crane as the SG15 representative on this matter. Ben has proved very helpful in clarifying and developing comments received from SG15. We have had several email exchanges with Ben about a few of the specific comments made in your review of the 02 revision of the Internet-Draft. Although he has taken these comments to the Q12/15 mailing list no response has been received, perhaps due to pressure of time. We are taking the opportunity to restate several questions here for your convenience and would appreciate your answers as part of the response to this liaison. 1. Definition of a Resource (section 3.2) You have supplied us with the simple text that says... In ASON applications: A resource refers only to a link connection (e.g. VC-11, VC-4 etc.). We would like to clarify this because we don't understand how the original text is wrong. It used to say... ITU-T [should be ASON] terms for resource: - Connection point (cp) in the context of link discovery and resource management (allocation, binding into cross-connects, etc.); - Link connection or trail termination in the context of routing, path computation and signaling. Recall that we are talking about the ASON terms for what GMPLS calls a resource. We think the following... In ASON a link connection is an association of output cp on one side of the link and input cp on the other side of the link. In GMPLS by resource we mean in most of the cases the local end of the resource (which is cp in ASON terminology). This is true in the context of link discovery. More importantly, this is also true in the context of a signaling application. Specifically, when a GMPLS signaling sub-system requests allocation of a resource, the GMPLS Resource Manager allocates only the local end of the resource. This happens on both sides of the link, that is, we allocate input and output resources. This contrasts with ASON, where there is a much more versatile LRM, and the connection manager (signaling application) allocates link connections only on the output side, leaving LRM to coordinate the allocation with the neighbor. However, in GMPLS TE advertisements we account only for outgoing resources, that is, we don't advertise incoming resources on the assumption that they match the outgoing resources for the discovered TE links. Bottom line: we believe that we should retain the previous definitions, and have not made this change in the new revision. We would like to discuss this point further with Q12/15. 2. Link ends and Data links (Section 3.5 - was section 3.4) 2a.Although we understand that ASON does not require the concept of a link end, GMPLS does. Therefore we should provide some form of language to help people do the mapping. The proposed new text from Q12/15 removes all reference to link ends from the ASON section, and we feel this is a mistake and propose the following text: In ASON, a GMPLS unidirectional data link end is a collection of connection points from the same client layer that are supported by a single trail termination (access point). This text depends on the fact that a link end is described as a set of resources that transfer traffic to a neighbor. 2b.The new text from Q12/15 gives us... A GMPLS data link is an ASON link supported by a single server trail. We are not sure that we understand this correctly. It appears to say that a unidirectional data link could be supported by multiple trail terminations (access points). This seems to suggest that a client would have to make a choice about where to send data, which seems to imply it really has to choose between data links! We are confused! We think some of this problem may come from the *need* in GMPLS to identify data link ends. There may be some value in stressing that a GMPLS link end is supported by exactly one trail termination. When several trail terminations are involved, we can still talk about single TE link (as a bundle), but each component link will be a distinct data link. We think that in ASON the SNPP (TE link in GMPLS language) could be supported by multiple terminations. 3. Link interfaces (section 3.6 - was 3.5) The new text from Q12/15 gives us... A GMPLS interface is "all the stuff between a physical interface and a matrix in an ASON network element." More precisely, it is a trail termination function and adaptation function for which adapted client layer connection points are bound to a matrix. We think "physical interface" may be ambiguous in the multi- layer context. We think the "stuff" should be between a link and a matrix, since we are talking about links in distinct layers. In addition to your feedback on the specific points above, we would welcome your continued comments and review of the Internet-Draft. Regards, Adrian Farrel and Kireeti Kompella IETF CCAMP Working Group Co-chairs Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 11:47:14 +0000 Message-ID: <42DCE79D.10405@ieee.org> Date: Tue, 19 Jul 2005 07:44:29 -0400 From: George Newsome <gnewsome@ieee.org> User-Agent: Mozilla/5.0 (Windows; U; Win98; en-US; rv:1.4) Gecko/20030624 Netscape/7.1 (ax) MIME-Version: 1.0 To: Adrian Farrel <adrian@olddog.co.uk> Cc: ccamp@ops.ietf.org Subject: Re: Moving LSP ownership between control plane and management plane Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 8bit Adrian, You asked for comments on this proposal, which was presented to ITU-T Q14 experts in January of this year. There are so many issues involved in moving a connection from the management plane to the control plane that involve intensive management action, that it is not evident that signaling need be involved at all. As a personal opinion, it makes little sense to consider moving the connection. It makes more sense to replace the connection with one set up via the control plane, and to then delete the original management set up connection. Appended is an excerpt of the meeting report detailing some of the issues that need to be resolved before signaling need be considered. Regards George Excerpt of the meeting report of Q14 Experts, January 2005 WD06 (Marconi) "G.7713 Modification in order to support Circuit Migration" contains a modified version of draft G.7713 Revision as was presented in the Berlin meeting (WD 11, i.e., TD50/3 Dec.2004). This contribution addresses SP24 (PC ? SPC) of G.7713. It proposes to extend the concept of Call/Connection setup to setup/adopt and release to release/de-adopt. In Connection Adoption both the SNC are not actually created and the LC are not actually established due to the fact that the underlying physical resources are already in place. In the Call/Connection dis-adoption SNC-LC-SNC is not actually de-allocated, only the Control Plane information associated with then is removed and the ownership is moved to Management Plane. Attributes for indicating Adoption/Dis-adoption are proposed for the INNI Connection messages. The discussion was mainly on the Management plane and Control plane actions required before signalling is involved. The group noted that the following issues need to be addressed: - Naming of transport resources to the control plane. Before a call can be placed under signalling control, links that are involved in the connection need to be given control plane names. Without these, no explicit route can be formed to match the resources of the connection. - Resource state while under dual CP and MP views. After resources are given control plane names, the resources are still viewed by the management plane. It may be necessary to create a new state for the resource to indicate that the management plane cannot perform some actions on the connection points as the control plane is about to take over. - Role of discovery functions (esp. CELA). After control plane names are allocated, distributed control plane functions need to be associated and communication adjacencies formed. This too is a precursor to any signalling procedures in migrating from the MP to CP. - Similarity of migration to synchronization after CP failure and subsequent recovery. Connection control procedures might be the same in migrating a call from the MP to the CP as the situation when the CP has failed and is recovering. Here, the network connections are already in place, but connection and call state need to be created to match it. Knowledge of the call and what the connection should be is obtained from the MP for migration, and from some reliable database in the recovery case. - Call awareness of migration vs. connection being unaware of migration. When connection state is being created to match an existing connection, the connection controllers (CCs) do not require awareness of why this is happening as the context could be migration or recovery. A mechanism to create control state without affecting transport plane state is needed in the CCs. Call controllers though do need migration awareness as they need to obtain/derive call identification from the MP. In summary, the group decided that it is premature to consider signalling procedure before the above issues, amongst other, have been studied. However the contribution does enable the group to expose in a larger context the interaction between CP and MP for connection migration. The above issues were included in SP24 of the G.7713 Living List (see WD22). Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 08:39:11 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01C58C3D.1688A89B" Subject: TR : I-D ACTION:draft-leroux-ccamp-ctrl-saturation-01.txt Date: Tue, 19 Jul 2005 10:37:23 +0200 Message-ID: <D109C8C97C15294495117745780657AE02E04114@ftrdmel1.rd.francetelecom.fr> Thread-Topic: I-D ACTION:draft-leroux-ccamp-ctrl-saturation-01.txt Thread-Index: AcWMOqCIJ1enawf0Ry2TRvWFIUEtPAAAEfbg From: "LE ROUX Jean-Louis RD-CORE-LAN" <jeanlouis.leroux@francetelecom.com> To: <ccamp@ops.ietf.org> This is a multi-part message in MIME format. ------_=_NextPart_001_01C58C3D.1688A89B Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Hi all, Here is a new version of the control plane saturation ID, which takes into account comments received on=20 the list and also some offline comments. Here are the major changes: -A new co-author -Motivations and elements of procedure have been clarified -Added new flags in the IGP Saturation TLV so as to indicate the reasons for the saturation -Added Inter-AS considerations -Some rewording for the sake of clarity Your comments on this new revision would be really appreciated. Adrian, Kireeti, we would like to see this item as part of the potential rechartering discussion... Best Regards, JL et al. > A New Internet-Draft is available from the on-line Internet-Drafts > directories. >=20 >=20 > Title : Procedure to handle (G)MPLS-TE control plane > saturation > Author(s) : J. Le Roux, et al. > Filename : draft-leroux-ccamp-ctrl-saturation-01.txt > Pages : 10 > Date : 2005-7-18 > =09 > This document defines extensions to RSVP-TE (Resource Reservation=20 > Protocol-Traffic Engineering) and IGP (IS-IS and OSPF), in order to >=20 > notify about control plane resources saturation, when an LSR runs > out=20 > of control plane resources available to support any additional LSP. >=20 > Such notification may serve as trigger for the impacted Head-end > LSR=20 > to take appropriate actions. >=20 > A URL for this Internet-Draft is: > http://www.ietf.org/internet-drafts/draft-leroux-ccamp-ctrl-saturation > -01.txt >=20 > To remove yourself from the I-D Announcement list, send a message to=20 > i-d-announce-request at ietf.org with the word unsubscribe in the body > of the message. =20 > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce >=20 > to change your subscription settings. >=20 >=20 > Internet-Drafts are also available by anonymous FTP. Login with the > username > "anonymous" and a password of your e-mail address. After logging in, > type "cd internet-drafts" and then > "get draft-leroux-ccamp-ctrl-saturation-01.txt". >=20 > A list of Internet-Drafts directories can be found in > http://www.ietf.org/shadow.html=20 > or ftp://ftp.ietf.org/ietf/1shadow-sites.txt >=20 >=20 > Internet-Drafts can also be obtained by e-mail. >=20 > Send a message to: > mailserv at ietf.org. > In the body type: > "FILE > /internet-drafts/draft-leroux-ccamp-ctrl-saturation-01.txt". > =09 > NOTE: The mail server at ietf.org can return the document in > MIME-encoded form by using the "mpack" utility. To use this > feature, insert the command "ENCODING mime" before the "FILE" > command. To decode the response(s), you will need "munpack" or > a MIME-compliant mail reader. Different MIME-compliant mail > readers > exhibit different behavior, especially when dealing with > "multipart" MIME messages (i.e. documents which have been split > up into multiple messages), so check your local documentation on > how to manipulate these messages. > =09 > =09 > Below is the data which will enable a MIME compliant mail reader > implementation to automatically retrieve the ASCII version of the > Internet-Draft. > <ftp://ftp.ietf.org/internet-drafts/draft-leroux-ccamp-ctrl-saturation > -01.txt> > <ftp://ftp.ietf.org/internet-drafts/draft-leroux-ccamp-ctrl-saturation > -01.txt> =20 > _______________________________________________ > I-D-Announce mailing list > I-D-Announce at ietf.org > https://www1.ietf.org/mailman/listinfo/i-d-announce >=20 ------_=_NextPart_001_01C58C3D.1688A89B Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN"> <HTML> <HEAD> <META HTTP-EQUIV=3D"Content-Type" CONTENT=3D"text/html; = charset=3Dus-ascii"> <META NAME=3D"Generator" CONTENT=3D"MS Exchange Server version = 6.5.7226.0"> <TITLE>TR : I-D ACTION:draft-leroux-ccamp-ctrl-saturation-01.txt = </TITLE> </HEAD> <BODY> <!-- Converted from text/rtf format --> <BR> <UL> <P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">Hi all,</FONT> </P> <P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">Here is a new version = of the control plane saturation ID, which takes into account comments = received on </FONT> <BR><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">the list and also = some offline comments.</FONT> <BR><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">Here are the major = changes:</FONT> <BR><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">-A new = co-author</FONT> <BR><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">-Motivations and = elements of procedure have been clarified</FONT> <BR><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">-Added new flags in = the IGP Saturation TLV so as to indicate the reasons for the = saturation</FONT> <BR><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">-Added Inter-AS = considerations</FONT> <BR><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">-Some rewording for = the sake of clarity</FONT> </P> <P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">Your comments on this = new revision would be really appreciated.</FONT> </P> <P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">Adrian, Kireeti, we = would like to see this item as part of the potential rechartering = discussion...</FONT> </P> <P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">Best Regards,</FONT> </P> <P><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Arial">JL et al.</FONT> </P> <BR> <BR> <P><FONT SIZE=3D2 FACE=3D"Courier New">A New Internet-Draft is available = from the on-line Internet-Drafts directories.</FONT> </P> <BR> <P> <FONT SIZE=3D2 = FACE=3D"Courier New">Title = : Procedure to handle = (G)MPLS-TE control plane</FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier = New"> &n= bsp; &nb= sp; saturation</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">Author(s) : J. = Le Roux, et al.</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">Filename = : draft-leroux-ccamp-ctrl-saturation-01.txt</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">Pages = : 10</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">Date = : 2005-7-18</FONT> <BR> =20 <BR><FONT SIZE=3D2 FACE=3D"Courier New">This document defines extensions = to RSVP-TE (Resource Reservation </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New"> Protocol-Traffic = Engineering) and IGP (IS-IS and OSPF), in order to </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New"> notify about = control plane resources saturation, when an LSR runs out </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New"> of control plane = resources available to support any additional LSP. </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New"> Such notification = may serve as trigger for the impacted Head-end LSR </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New"> to take appropriate = actions.</FONT> </P> <P><FONT SIZE=3D2 FACE=3D"Courier New">A URL for this Internet-Draft = is:</FONT> <BR><A = HREF=3D"http://www.ietf.org/internet-drafts/draft-leroux-ccamp-ctrl-satur= ation-01.txt"><U><FONT COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Courier = New">http://www.ietf.org/internet-drafts/draft-leroux-ccamp-ctrl-saturati= on-01.txt</FONT></U></A> </P> <P><FONT SIZE=3D2 FACE=3D"Courier New">To remove yourself from the I-D = Announcement list, send a message to </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">i-d-announce-request at ietf.org = with the word unsubscribe in the body of the message. </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">You can also visit </FONT><A = HREF=3D"https://www1.ietf.org/mailman/listinfo/I-D-announce"><U><FONT = COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Courier = New">https://www1.ietf.org/mailman/listinfo/I-D-announce</FONT></U></A><F= ONT SIZE=3D2 FACE=3D"Courier New"> </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">to change your subscription = settings.</FONT> </P> <BR> <P><FONT SIZE=3D2 FACE=3D"Courier New">Internet-Drafts are also = available by anonymous FTP. Login with the username</FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">"anonymous" and a = password of your e-mail address. After logging in,</FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">type "cd = internet-drafts" and then</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">"get = draft-leroux-ccamp-ctrl-saturation-01.txt".</FONT> </P> <P><FONT SIZE=3D2 FACE=3D"Courier New">A list of Internet-Drafts = directories can be found in</FONT> <BR><A HREF=3D"http://www.ietf.org/shadow.html"><U><FONT = COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Courier = New">http://www.ietf.org/shadow.html</FONT></U></A><FONT SIZE=3D2 = FACE=3D"Courier New"> </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">or </FONT><A = HREF=3D"ftp://ftp.ietf.org/ietf/1shadow-sites.txt"><U><FONT = COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Courier = New">ftp://ftp.ietf.org/ietf/1shadow-sites.txt</FONT></U></A> </P> <BR> <P><FONT SIZE=3D2 FACE=3D"Courier New">Internet-Drafts can also be = obtained by e-mail.</FONT> </P> <P><FONT SIZE=3D2 FACE=3D"Courier New">Send a message to:</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">mailserv at ietf.org.</FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">In the body type:</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">"FILE = /internet-drafts/draft-leroux-ccamp-ctrl-saturation-01.txt".</FONT> <BR> =20 <BR><FONT SIZE=3D2 FACE=3D"Courier New">NOTE: The mail = server at ietf.org can return the document in</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">MIME-encoded form by using the "mpack" = utility. To use this</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">feature, insert the command "ENCODING = mime" before the "FILE"</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">command. To decode the response(s), you will = need "munpack" or</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">a MIME-compliant mail reader. Different = MIME-compliant mail readers</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">exhibit different behavior, especially when dealing = with</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">"multipart" MIME messages (i.e. documents = which have been split</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">up into multiple messages), so check your local = documentation on</FONT> <BR> <FONT SIZE=3D2 = FACE=3D"Courier New">how to manipulate these messages.</FONT> <BR> = =20 <BR> = =20 <BR><FONT SIZE=3D2 FACE=3D"Courier New">Below is the data which will = enable a MIME compliant mail reader</FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">implementation to automatically = retrieve the ASCII version of the</FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">Internet-Draft.</FONT> <BR><A = HREF=3D"ftp://ftp.ietf.org/internet-drafts/draft-leroux-ccamp-ctrl-satura= tion-01.txt"><U><FONT COLOR=3D"#0000FF" FACE=3D"Times New = Roman"><ftp://ftp.ietf.org/internet-drafts/draft-leroux-ccamp-ctrl-sat= uration-01.txt></FONT></U></A><FONT FACE=3D"Times New Roman"> </FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier = New">_______________________________________________</FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">I-D-Announce mailing list</FONT> <BR><FONT SIZE=3D2 FACE=3D"Courier New">I-D-Announce at ietf.org</FONT> <BR><A = HREF=3D"https://www1.ietf.org/mailman/listinfo/i-d-announce"><U><FONT = COLOR=3D"#0000FF" SIZE=3D2 FACE=3D"Courier = New">https://www1.ietf.org/mailman/listinfo/i-d-announce</FONT></U></A> </P> </UL> </BODY> </HTML> ------_=_NextPart_001_01C58C3D.1688A89B-- Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 08:39:06 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: RE: no more milestones? Date: Tue, 19 Jul 2005 09:36:18 +0100 Message-ID: <0536FC9B908BEC4597EE721BE6A353890A9F1DED@i2km07-ukbr.domain1.systemhost.net> Thread-Topic: no more milestones? Thread-Index: AcWL2gpHx44PAcioRP6bjKbT0RZ4CwAUdAwQ From: <neil.2.harrison@bt.com> To: <adrian@olddog.co.uk>, <tnadeau@cisco.com>, <Dimitri.Papadimitriou@alcatel.be> Cc: <ccamp@ops.ietf.org>, <zinin@psg.com>, <fenner@research.att.com> Thanks Adrian, >=20 > Hi Neil and Tom, >=20 > >> There are for instance, ops/management-related items > >> for "base" CCAMP items that need to be worked on. The=20 > question for=20 > >> me is why can't we continue to work on these? I thought=20 > these things=20 > >> were automatic chartered items, and thus should be ongoing work=20 > >> items. > > > >You are right Tom to raise the point about the importance of network=20 > >management here. When we are dealing with large BW=20 > granularity co-cs=20 > >mode technologies then virtually all the really important issues are=20 > >concerned with the efficacy of the management-plane solution....the=20 > >control-plane plays a minor support role (for a whole raft of=20 > >technical/commercial/operational reasons). This is of course=20 > quite the=20 > >opposite to a cl-ps layer network like IP. We look to TMF for=20 > >expertise here in developing the required management=20 > information models=20 > >which are based on functional architecture. >=20 > I think you are right, Neil, that we would largely look=20 > outside the IETF for management plane solutions. NH=3D> I know some folks must think 'why does he keep banging about = stuff like functional architecture, 3 network modes, client/server relationships, etc'....well, its for sound practical reasons, and one of them being the importance of the management information models we operators need if we are going to get the right equipment specifications and maximally effective downstream NMS/OSS. These facets can only be as good as the base networking specification of the data and control planes, ie if these violate architectural rules of the mode concerned or have functional deficiencies then no amount of effort/money subsequently thrown at the management piece can really recover the situation. And speaking as an operator I am very concerned about whole-life costs and this is why the above stuff is important to me......esp noting there is a significant time-constant between cause and effect, so its often very easy to make mistakes early on that don't become apparent wrt implications until several years later.....and then its invariably too late to fix easily. So if there is one key message I'd like folks to get it's this: We must strive to get the networking aspects of the data-plane and control-plane right/complete for the mode we are considering from t=3D0....if we = don't, the NMS/OSS folks who pick this up later cannot recover the situation. Folks involved in any aspect of networking (esp those creating standards) should therefore try and make the effort to better understand functional architecture stuff.=20 >=20 > What Tom is referring to, of course, is the management=20 > aspects necessary as part of the control plane solutions.=20 > Aspects such as MIB modules, PIBs, OAM etc. NH=3D> Understood and agreed Adrian.....and I'm glad to note here you mention OAM as a 'for instance' here. This is a good example of what I was referring to above, ie this needs to be an integral part of the trail's characteristic information specification at t=3D0. Some folks = may remember that ~3-4 years ago I spent a futile year or so trying to convince the MPLS crowd that they must consider this aspect AND that it must not be proxied by a control-plane solution (this becomes very obviously true when the data-plane and control-plane protocols run on logically distinct layer networks within the SAME networking system, as indeed they are forced to in the co-cs mode). However, I should also mention that if one breaks the connectivity rules of the co mode (ie merging as we find with LDP and PHP in some spins of MPLS) then there are no good/simple solutions to OAM/fault management (or indeed traffic or performance management).....this is just a consequence you now have to accept and live with. >=20 > I'm sure that I don't have to remind you that GMPLS is a=20 > control plane technology not a data plane technology. The=20 > applicability of GMPLS is to any co data plane technology and=20 > therefore it embraces both co-cs and co-ps. NH=3D> No you don't have to remind me of this.....and you are mentioning = a really important thing here that, if I may, I'd like to amplify a little: - all technologies map to one of the 3 networking modes.....the exact behaviour of a given technology is then dependent on well/completely it is functionally specified for its target environment, and how well it respects modal networking constraints. In terms of what this means practically, then (i) it makes huge amounts of sense to seek convergence within each mode but it is not sensible to think exactly the same approach can be applied across all 3 modes, and (ii) each mode has very important behavioural differences that we should seek to exploit as these are the key to addressing what are otherwise very hard networking problems. {Aside - There is actually a unifying model based on Shannon's Information Theory which brings the G.805 (co) and G.809 (cl) func arch perspectives together which one of my colleagues is working on (paper to be issued later this year), but in terms of what we see manifested we simply need to note that its these differences that are the key to tackling some otherwise very hard networking problems in the areas of QoS/performance-inheritance/pricing/complexity/etc management....I gave some hints on this in the thread 'End-to-end L2-LSP' 15 July 05.} - the major difference between GMPLS and MPLS is, as you correctly observe Adrian, that no data-plane structures have been specified for the former in IETF whereas they have for the latter (and it is this that is the source of the problems for MPLS as-is...we could actually fix these if there was the will to let go of a few misplaced beliefs.....but I am not sure that there is that will given how far some of this stuff has now gone). - unlike the co-ps mode the co-cs mode is rather 'self-protecting' of abuse, ie its very difficult to do bad things to it even if you try....and the reasons why this is so is hidden-away in the underlying Information Theory constraints of what 'labelling' means in this mode (see also my rather detailed response to Igor Bryskin of recent, 'Network modes and layer network relationships' 13 July 05, which explains why one can't have a single control-plane running 'IP to Optics'). So in the co-cs mode: OOB control/management is forced, there are no 'QoS classes', we can always have proper client/server functional decoupling (subject to my observations wrt the nonsense of 'topology on the fly', etc) one cannot merge co-cs trails and of course trail must go from source to sink (ie can't prematurely stop at the wrong place). There are also some interesting consequential issues wrt performance inheritance in nested client/server layer network relationships if one thinks any mode can arbitrarily support any other mode (eg PWE3 stuff). - it ought to be somewhat intuitively obvious that its a good idea to try and re-use the same base control-plane components/design for the co-ps mode as the co-cs mode. And this is one reason (there are actually lots of reasons) why I am supportive of the folks who also advocate this view.....and I think you share this view Adrian. regards, Neil >=20 > In answer to Tom's question, we are not prohibited from=20 > working on these things, but we are rather adrift without any=20 > milestones. i would certainly say that the right place to=20 > bring drafts for discussion is the CCAMP list. >=20 > However, before committing the WG to any specific drafts the=20 > chairs need to be guided by an up-to-date charter with real=20 > milestones. I do hope that the discussions in Paris will lead=20 > to these so that we can get on with the work that everyone is=20 > so keen to do. >=20 > Cheers, > Adrian >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 19 Jul 2005 00:48:54 +0000 Message-ID: <42DC4DE0.3000902@kddilabs.jp> Date: Tue, 19 Jul 2005 09:48:32 +0900 From: Tomohiro Otani <otani@kddilabs.jp> User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; ja-JP; rv:1.7.2) Gecko/20040803 MIME-Version: 1.0 To: Dimitri.Papadimitriou@alcatel.be Cc: ccamp@ops.ietf.org Subject: Re: http://www.ietf.org/internet-drafts/draft-otani-ccamp-gmpls-cspf-constraints-00 Content-Type: text/plain; charset=ISO-2022-JP Content-Transfer-Encoding: 7bit Hi Dimitri, Per our initial discussion, we would like to formulate the unified view on CSPF consideration, and we simply thought it is BCP. Considering more deeply, our draft does not define the protocol itself but the protocol related specification In that sense, you are right, it is prescriptive more than descriptive for guiding operations. Moreover, I came up my mind that it may be more standard track than informational. But I am not sure. What's your thought on this ? With best regards, tomo PS: We posted 01 version. Dimitri.Papadimitriou@alcatel.be wrote: >authors, is the purpose of the above mentioned draft: > > 1. prescriptive (even if informational) > > 2. descriptive for guiding operations (BCP) > > 3. informative (survey-like) > >from what i read, it is more 1 like - is my understanding correct ? > > > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 18 Jul 2005 21:47:50 +0000 Message-ID: <008101c58be2$5082b420$7a1810ac@movaz.com> From: "Igor Bryskin" <ibryskin@movaz.com> To: "Arthi Ayyangar" <arthi@juniper.net> Cc: <ccamp@ops.ietf.org> Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Date: Mon, 18 Jul 2005 17:47:37 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Arthi, ----- Original Message ----- From: "Arthi Ayyangar" <arthi@juniper.net> To: "Igor Bryskin" <ibryskin@movaz.com> Cc: <i-d-announce@ietf.org>; <Internet-Drafts@ietf.org>; <ccamp@ops.ietf.org> Sent: Monday, July 18, 2005 2:51 PM Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > Hi Igor, > > Please see inline. > > > 1. Why are saying that LSP Stitching is a private case of LSP Hierarchy? > > > there is more differences than similarities: > > The differences are: > > 1) In case of H-LSP there is a data plane adjacency, while in case of S-LSP > > there is none (as you correctly pointed out); > > 2) In case of H-LSP there is an adaptation in data plane (label push/pop for > > PSC), while in case of S-LSP there is none - just simple cross-connecting > > ( label swap) as in case of two "native " e2e LSP adjacent segments ; > > 3) H-LSP could be used by many e2e LSPs, while S-LSP could be used by > > exactly one e2e LSP > > 4) Signaling is different - there is no label negotiation in stitching > > 5) H-LSP is used as a "true" data link, specifically there is a resource > > allocation on the H-LSP edges, while in case of S-LSP there is none > > 6) From MLN point of view, H-LSP is created in a server (lower) layer, while > > the S-LSP is created in the client (same as e2e LSP) layer. > > > > There are two similarities that I can think of: > > 1) There is a signaling and possibly routing (see below) adjacencies between > > the ends; > > 2) Both H-LSP and S-LSP could be advertised as separate TE links or as TE > > bundles > > > > I would recommend to dedicate a paragraph and enlist there similarities and > > differencies > -------> Based on the discussions on the list over the last revision, the > message that the authors received was that we do not want to go about such > a lengthy discussion of similarities and differences (NOTE that the draft > already does clearly highlight them where applicable). Instead we want > this ID to simply explain how LSP stitching functions. In other words the > idea was to make this ID a complete document by itself. But since it does > borrow concepts from the LSP hierarchy ID, just state the concepts that > are applicable or inapplicable. IB>> I see a problem here. Since we do not have another draft providing the tight definition of LSP stitching, it is not good enough to describe it in some light narrative manner. In section 1 you are saying: 1. Introduction LSP stitching is a special case of LSP hierarchy. In case of LSP stitching, instead of an FA LSP, an "LSP segment" is created between two GMPLS nodes .... I don't think that "LSP stitching is a special case of LSP hierarchy". The two mechanism have some remote IMO resemblance in control plane and have nothing in common in the data plane. The biggest difference is that the stitching happens within the same network layer and hence in sharp contrast to LSP hierarchy does not require reconfigurations in lower (fat pipe) layer(s). Just because of that the stitching could be much more attractive mechanism for control plane. At least there will be no questions like " Who is going to pay for unused bandwidth?" Anyway, all I asked was for a small paragraph summarizing similarities and differences of both planes. That IMO would help. One more point. You use terms FAs and FA-LSPs everywhere in the draft. This is confusing because what you really mean hear is Hierarchical LSPs (H-LSPs) and TE links enabled by H-LSPs. As I mentioned earlier, there is no place for these terms in your draft because it mostly concerns about TE inter-domain scenarios, while [LSP-HIER] has a focus on a single domain and strictly mandates the LSP to be advertised into the *same* domain to be considered as FA-LSP > > > 2. Why are you saying that a TE Link based on S-LSP can be used for exactly > > one e2e LSP? Parallel S-LSPs could be advertised as a single TE link > > (bundle, see above) and hence can accomadate several e2e LSPs. > ----> Igor, the draft does talk about this Bundling case. IB>> My problem here is with the statement from section 2: "An LSP segment TE link SHOULD NOT admit more than one e2e LSP into it. " You do describe the bundling, which contradicts the statement above. It may accept as many e2e LSPs as it has component links > > > 3. You are saying that S-LSP does not have a routing peering. Actually, in > > this respect it is no different from H-LSP: if it is advertised as a TE link > > into the same TE domain that was used for S-LSP creation (unlikely IMO > > scenario) than it does not require the routing adjacency (in other words, it > > is an FA according to LSP-HIER definition), otherwise, it IS NOT and FA and > > does require the direct routing peering in the domain it is advertised to > > make it useful as a TE link in this domain > -----> I am sorry, but I am missing your point here completely. Are you > saying that the statement "the end points of an LSP segment do not have a > routing adjacency", is incorrect or are you saying that "this is > obvious" ? It is unclear to me what exactly your argument is. Please > clarify. IB>> To be more clear I should say the following. If an LSP segment is advertised into a different TE domain it SHOULD have a routing peering between the ends in the domain it is advertised, otherwise, it could not be used as a TE link in this domain. This is because generally speaking the data plane connectivity does not guarantee control plane connectivity in the domain the link is advertised, and in GMPLS a connection could not be used as a TE link without control channel connecting nodes handling the link. The same applies to hierarchical LSPs as well. The exception is LSP (stitching or hierarchy) advertised into the same domain, in this case it is FA-LSP and does not need direct routing peering to be used as a TE link. Cheers, Igor > > thanks, > -arthi > > > > > Cheers, > > Igor > > > > > > > > ----- Original Message ----- > > From: <Internet-Drafts@ietf.org> > > To: <i-d-announce@ietf.org> > > Cc: <ccamp@ops.ietf.org> > > Sent: Friday, July 15, 2005 3:50 PM > > Subject: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > > > > > > > A New Internet-Draft is available from the on-line Internet-Drafts > > directories. > > > This draft is a work item of the Common Control and Measurement Plane > > Working Group of the IETF. > > > > > > Title : Label Switched Path Stitching with Generalized MPLS Traffic > > Engineering > > > Author(s) : A. Ayyangar, J. Vasseur > > > Filename : draft-ietf-ccamp-lsp-stitching-01.txt > > > Pages : 19 > > > Date : 2005-7-15 > > > > > > In certain scenarios, there may be a need to combine together two > > > different Generalized Multi-Protocol Label Switching (GMPLS) Label > > > Switched Paths (LSPs) such that in the data plane, a single end-to- > > > end (e2e) LSP is achieved and all traffic from one LSP is switched > > > onto the other LSP. We will refer to this as "LSP stitching". This > > > document covers cases where: a) the node performing the stitching > > > does not require configuration of every LSP pair to be stitched > > > together b) the node performing the stitching is not the egress of > > > any of the LSPs c) LSP stitching not only results in an end-to-end > > > LSP in the data plane, but there is also a corresponding end-to-end > > > LSP (RSVP session) in the control plane. It might be possible to > > > configure a GMPLS node to switch the traffic from an LSP for which it > > > is the egress, to another LSP for which it is the ingress, without > > > requiring any signaling or routing extensions whatsoever, completely > > > transparent to other nodes. This will also result in LSP stitching > > > in the data plane. However, this document does not cover this > > > scenario of LSP stitching. > > > > > > A URL for this Internet-Draft is: > > > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt > > > > > > To remove yourself from the I-D Announcement list, send a message to > > > i-d-announce-request@ietf.org with the word unsubscribe in the body of the > > message. > > > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > > > to change your subscription settings. > > > > > > > > > Internet-Drafts are also available by anonymous FTP. Login with the > > username > > > "anonymous" and a password of your e-mail address. After logging in, > > > type "cd internet-drafts" and then > > > "get draft-ietf-ccamp-lsp-stitching-01.txt". > > > > > > A list of Internet-Drafts directories can be found in > > > http://www.ietf.org/shadow.html > > > or ftp://ftp.ietf.org/ietf/1shadow-sites.txt > > > > > > > > > Internet-Drafts can also be obtained by e-mail. > > > > > > Send a message to: > > > mailserv@ietf.org. > > > In the body type: > > > "FILE /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt". > > > > > > NOTE: The mail server at ietf.org can return the document in > > > MIME-encoded form by using the "mpack" utility. To use this > > > feature, insert the command "ENCODING mime" before the "FILE" > > > command. To decode the response(s), you will need "munpack" or > > > a MIME-compliant mail reader. Different MIME-compliant mail readers > > > exhibit different behavior, especially when dealing with > > > "multipart" MIME messages (i.e. documents which have been split > > > up into multiple messages), so check your local documentation on > > > how to manipulate these messages. > > > > > > > > > Below is the data which will enable a MIME compliant mail reader > > > implementation to automatically retrieve the ASCII version of the > > > Internet-Draft. > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 18 Jul 2005 21:40:36 +0000 Message-ID: <0d5e01c58be1$966ebf70$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <ccamp@ops.ietf.org> Subject: Moving LSP ownership between control plane and management plane Date: Mon, 18 Jul 2005 22:40:11 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi, You may not have seen this I-D because it lacks the "ccamp" element in its name. It discusses the issues for switching ownership of an LSP from the management plane to the control plane and back again. Can we have some comments from the CCAMP community (and especially from the providers) on whether this function is necessary. Would you consider deploying a control plane into an active network, or would you shut the network (or at least the LSPs) for maintenance, and bring it (or them) back up using the control plane? The I-D also proposes a relatively simple solution and I would welcome comments on this, too. Thanks, Adrian ----- Original Message ----- From: <Internet-Drafts@ietf.org> To: <i-d-announce@ietf.org> Sent: Thursday, June 30, 2005 11:50 PM Subject: I-D ACTION:draft-caviglia-mp2cpcp2mp-02.txt > A New Internet-Draft is available from the on-line Internet-Drafts directories. > > > Title : GMPLS Signaling Extensions for the Transfer of > Ownership of Label Switched Paths Between > the Management and Control Planes > Author(s) : D. Caviglia, et al. > Filename : draft-caviglia-mp2cpcp2mp-02.txt > Pages : 13 > Date : 2005-6-30 > > During migration scenarios it may be desirable to transfer the > ownership of a Label Switched Path (LSP) from the Management Plane > (MP) to the Control Plane (CP), or vice versa. If the LSP is carrying > traffic this change needs to be made "in service," that is, without > affecting traffic. > This memo provides minor extensions to the Generalized Multi Protocol > Label Switching (GMPLS) signaling protocol, GRSVP_TE (Generalized > Resource Reservation Protocol with Traffic Engineering Extensions), > to enable such transfer of ownership and describes the proposed > procedures. > > A URL for this Internet-Draft is: > http://www.ietf.org/internet-drafts/draft-caviglia-mp2cpcp2mp-02.txt > > To remove yourself from the I-D Announcement list, send a message to > i-d-announce-request@ietf.org with the word unsubscribe in the body of the message. > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > to change your subscription settings. > > > Internet-Drafts are also available by anonymous FTP. Login with the username > "anonymous" and a password of your e-mail address. After logging in, > type "cd internet-drafts" and then > "get draft-caviglia-mp2cpcp2mp-02.txt". > > A list of Internet-Drafts directories can be found in > http://www.ietf.org/shadow.html > or ftp://ftp.ietf.org/ietf/1shadow-sites.txt > > > Internet-Drafts can also be obtained by e-mail. > > Send a message to: > mailserv@ietf.org. > In the body type: > "FILE /internet-drafts/draft-caviglia-mp2cpcp2mp-02.txt". > > NOTE: The mail server at ietf.org can return the document in > MIME-encoded form by using the "mpack" utility. To use this > feature, insert the command "ENCODING mime" before the "FILE" > command. To decode the response(s), you will need "munpack" or > a MIME-compliant mail reader. Different MIME-compliant mail readers > exhibit different behavior, especially when dealing with > "multipart" MIME messages (i.e. documents which have been split > up into multiple messages), so check your local documentation on > how to manipulate these messages. > > > Below is the data which will enable a MIME compliant mail reader > implementation to automatically retrieve the ASCII version of the > Internet-Draft. > -------------------------------------------------------------------------- ------ > _______________________________________________ > I-D-Announce mailing list > I-D-Announce@ietf.org > https://www1.ietf.org/mailman/listinfo/i-d-announce > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 18 Jul 2005 21:40:29 +0000 Message-ID: <0d5c01c58be1$8ca4cc00$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: "Arthi Ayyangar" <arthi@juniper.net>, <Dimitri.Papadimitriou@alcatel.be> Cc: "Igor Bryskin" <ibryskin@movaz.com>, <ccamp@ops.ietf.org> Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Date: Mon, 18 Jul 2005 22:23:17 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi, [i-d-announce@ietf.org and Internet-Drafts@ietf.org omitted from the cc list] [SNIP] > > [dp] i would simply recommend to state what an LSP segment is (as stated, > > this comparison has been used to show differences in terms of control > > plane processing and keeping it at that level is sensible) and not embark > > this document into terms and complex comparisons that are at the end of > > no real help > <AA>----> I tend to agree with this. Yes. If there is a need to compare the two techniques or to do applicbality then that is another I-D. I am not sure that there is a need. I would like to see this I-D restricted, as Arthi said, to a description of what Stitching is and how it works. > > 2. Why are you saying that a TE Link based on S-LSP can be used for > > exactly > > one e2e LSP? Parallel S-LSPs could be advertised as a single TE link > > (bundle, see above) and hence can accomadate several e2e LSPs. > > > > [dp] i think the purpose is to say that a "triggered" LSP segment can be > > used by a single end-to-end LSP compared to the situation occurring with > > FA, where the triggered FA-LSP can then carry multiple nested LSPs > <AA>------> As I mentioned in the email to Igor, the sentence immediately > following this does talk about the bundling case. What is missing is > probably an additional sentence which explicitly states that when such > bundling is use, more that one e2e LSP can be admitted over the LSP > segment. That can be clarified. Say rather, "can be admitted over a TE link supported by LSP segments". That is, a segment can only admit one LSP, but a bundle of segments advertised as a TE link, can admit multiple LSPs. A Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 18 Jul 2005 20:49:54 +0000 Message-ID: <0cf601c58bda$60c2fcd0$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <neil.2.harrison@bt.com>, "Thomas D. Nadeau" <tnadeau@cisco.com>, "Dimitri Papadimitriou" <Dimitri.Papadimitriou@alcatel.be> Cc: <ccamp@ops.ietf.org>, <zinin@psg.com>, "Bill Fenner" <fenner@research.att.com> Subject: Re: no more milestones? Date: Mon, 18 Jul 2005 21:40:59 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi Neil and Tom, >> There are for instance, ops/management-related items >> for "base" CCAMP items that need to be worked on. The question >> for me is why can't we continue to work on these? I thought >> these things were automatic chartered items, and thus should >> be ongoing work items. > >You are right Tom to raise the point about the importance of >network management here. When we are dealing with large BW >granularity co-cs mode technologies then virtually all the >really important issues are concerned with the efficacy of the >management-plane solution....the control-plane plays a minor >support role (for a whole raft of technical/commercial/operational >reasons). This is of course quite the opposite to a cl-ps layer >network like IP. We look to TMF for expertise here in developing >the required management information models which are based on >functional architecture. I think you are right, Neil, that we would largely look outside the IETF for management plane solutions. What Tom is referring to, of course, is the management aspects necessary as part of the control plane solutions. Aspects such as MIB modules, PIBs, OAM etc. I'm sure that I don't have to remind you that GMPLS is a control plane technology not a data plane technology. The applicability of GMPLS is to any co data plane technology and therefore it embraces both co-cs and co-ps. In answer to Tom's question, we are not prohibited from working on these things, but we are rather adrift without any milestones. i would certainly say that the right place to bring drafts for discussion is the CCAMP list. However, before committing the WG to any specific drafts the chairs need to be guided by an up-to-date charter with real milestones. I do hope that the discussions in Paris will lead to these so that we can get on with the work that everyone is so keen to do. Cheers, Adrian Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 18 Jul 2005 20:49:46 +0000 Message-ID: <0cf701c58bda$6c49bb70$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <Dimitri.Papadimitriou@alcatel.be> Cc: <ccamp@ops.ietf.org>, "Thomas D. Nadeau" <tnadeau@cisco.com>, <zinin@psg.com>, "Bill Fenner" <fenner@research.att.com> Subject: Re: no more milestones? Date: Mon, 18 Jul 2005 21:47:09 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi Dimitri, > > I would like to see: > > - a refinement of the charter to focus the next work items > > - a year's worth of milestones to ensure we stay focused. > > > > On the other hand, we should aim to close down CCAMP and move the work to > > other or new working groups. > > coupling both sentences does it mean in your view > > 1) define several short-term milestones for refining topics that have been > already initiated and requires further well scoped attention > > 2) (brand) new topics should target new working groups - as already > initiated with L1VPN for inst. - or existing working groups > > nevertheless, i do not see why this automatically means CCAMP can not > remain the focal point concerning *base* GMPLS signaling and routing > protocols To answer your question: yes, I would like to see both. We need some short-term milestones to guide and refine our work on existing topics. For example, to complete the inter-domain work, possibly including issues of protection and restoration. But there are also some topics that are broadly covered by the CCAMP scope, where we would benefit greatly by having a specific description on the charter, and detailed milestones. I am thinking about L2SC, MRN, MPLS/GMPLS interworking as discussed on the mailing list last November. While we could sort of take on these things under the vague scope of the current charter, I would prefer for us to be open, clear and focused. And this covers your last point. I, too, don't see why CCAMP should not remain the focal point for GMPLS signaling and routing protocols. However, to do so in an efficient and proper manner, we should add the appropriate milestones to the charter. If, on the other hand, we are not going to add this stuff to the charter, it would be very important for us to be clear about this so that the people working in these areas know that they must take their work to another WG or BOF. Thanks, Adrian Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 18 Jul 2005 19:23:00 +0000 MIME-Version: 1.0 To: "Thomas D. Nadeau" <tnadeau@cisco.com> Cc: Dimitri.Papadimitriou@alcatel.be, "Adrian Farrel" <adrian@olddog.co.uk>, <ccamp@ops.ietf.org>, <zinin@psg.com>, "Bill Fenner" <fenner@research.att.com> From: Dimitri.Papadimitriou@alcatel.be Subject: Re: no more milestones? Date: Mon, 18 Jul 2005 21:20:39 +0200 Message-ID: <OF38AF2CA2.71FA7800-ONC1257042.00565140-C1257042.006A42CD@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPnRvbSAtIDwvRk9OVD48L1A+PFA+PEZP TlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPml0IHNlZW1zIHRvIG1lIHRvIGJlIHRoZSByaWdo dCB0aW1lIHRvIHdyYXAgdXAgYW5kIGxvb2sgYWZ0ZXIgdGhlIGJhc2Ugb3BlcmF0aW9uIG1lY2hh bmlzbXMgdG8gYmUgcHV0IGluIHBsYWNlIGZvciB0aGUgZmlyc3QgcGhhc2Ugb2YgYmFzZSBHTVBM UyBkZXBsb3ltZW50IGFzIHdlbGwgLSBhbmQgbGVhdmUgdGhlIG1vcmUgYWR2YW5jZWQgbWVjaGFu aXNtcyB0byAmbmJzcDsmcXVvdDtDQ0FNUC1JTkcmcXVvdDsgbGlrZS1XRyhzKSBhcyBpbmRpY2F0 ZWQgaW4gYSBwcmV2aW91cyBlLW1haWwgLTwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3Nw YWNlLENvdXJpZXIiPml0IGlzIGluZGVlZCBkaWZmaWN1bHQgdG8gdW5kZXJzdGFuZCB3aHkgdGhp cyBpcyBub3QgY291cGxlZCB0byB0aGUgYmFzZSBwcm90b2NvbCBkZXZlbG9wbWVudCBhcyBNSUJz IGFyZSBvdGhlcndpc2UgR01QTFMgd2lsbCBhcHBlYXIgbGlrZSBhIHdvbmRlcmZ1bCBhbmQgcG93 ZXJmdWwgdG9vbGJveCBkZXZlbG9wZWQgYnkgdGhlIElFVEYgYnV0IHdpdGhvdXQgdGhlIGFwcHJv cHJpYXRlIHVzZXIgbWFudWFsIDwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENv dXJpZXIiPmkgd291bGQgbmV2ZXJ0aGVsZXNzIGFkZCB3ZSBoYXZlIGFsc28gYSBzdHJvbmcgbmVl ZCBmb3IgbW9yZSBvcGVyYXRpb25hbCBmZWVkYmFjayBvdGhlcndpc2UgdGhlIHVzZXIgbWFudWFs IHdpbGwgdHJlYXQgbW9yZSBhYm91dCAmcXVvdDtob3cgZGV2ZWxvcG1lbnQgY29tbXVuaXR5IHNl ZXMgbmV0d29yayBvcGVyYXRpb25zJnF1b3Q7IHJhdGhlciB0aGFuIHByYWN0aWNlICZxdW90O3do aWNoIG1lY2hhbmlzbXMgZGV2ZWxvcG1lbnQgY29tbXVuaXR5IGhhcyB0byBjb25zaWRlciB0byBt ZWV0IHVzZXIgY29tbXVuaXR5IG5lZWRzJnF1b3Q7IC4uLiBhIHdheSB0byBhY2hpZXZlIHRoaXMg aXMgd291bGQgYmUgdG8gJnF1b3Q7aW52aXRlJnF1b3Q7IHRoZSBicm9hZCB1c2VyIGNvbW11bml0 eSB0byBleHByZXNzIHRoZWlyIHZpZXdzIGFuZCBjb25jZXJucyBpbiB0ZXJtcyBvZiBHTVBMUyBk ZXBsb3ltZW50IHRoaXMgbWF5IHRha2UgdGhlIGZvcm0gb2YgYSBzdXJ2ZXkgPzwvRk9OVD48QlI+ PEJSPjxGT05UIFNJWkU9Mj48Qj4mcXVvdDtUaG9tYXMgRC4gTmFkZWF1JnF1b3Q7ICZsdDt0bmFk ZWF1QGNpc2NvLmNvbSZndDs8L0I+PC9GT05UPjxCUj48Rk9OVCBTSVpFPTI+U2VudCBieTogb3du ZXItY2NhbXBAb3BzLmlldGYub3JnPC9GT05UPjxCUj48Rk9OVCBTSVpFPTI+MDcvMTgvMjAwNSAw OTowNiBBU1Q8L0ZPTlQ+PEJSPjxCUj4gPEZPTlQgU0laRT0yPlRvOjwvRk9OVD4gPEZPTlQgU0la RT0yPkRpbWl0cmkgUEFQQURJTUlUUklPVS9CRS9BTENBVEVMQEFMQ0FURUw8L0ZPTlQ+PEJSPiA8 Rk9OVCBTSVpFPTI+Y2M6PC9GT05UPiA8Rk9OVCBTSVpFPTI+JnF1b3Q7QWRyaWFuIEZhcnJlbCZx dW90OyAmbHQ7YWRyaWFuQG9sZGRvZy5jby51ayZndDssICZsdDtjY2FtcEBvcHMuaWV0Zi5vcmcm Z3Q7LCAmbHQ7emluaW5AcHNnLmNvbSZndDssICZxdW90O0JpbGwgRmVubmVyJnF1b3Q7ICZsdDtm ZW5uZXJAcmVzZWFyY2guYXR0LmNvbSZndDs8L0ZPTlQ+PEJSPiA8Rk9OVCBTSVpFPTI+YmNjOjwv Rk9OVD4gPEJSPjxGT05UIFNJWkU9Mj5TdWJqZWN0OjwvRk9OVD4gPEZPTlQgU0laRT0yPlJlOiBu byBtb3JlIG1pbGVzdG9uZXM/PC9GT05UPjxCUj4gPEJSPjxCUj48L1A+PFA+PEZPTlQgRkFDRT0i TW9ub3NwYWNlLENvdXJpZXIiPk9uIEp1bCAxOCwgMjAwNSwgYXQgOTowMCBBTSwgRGltaXRyaS5Q YXBhZGltaXRyaW91QGFsY2F0ZWwuYmUgd3JvdGU6PEJSPjwvRk9OVD48QlI+PEZPTlQgRkFDRT0i TW9ub3NwYWNlLENvdXJpZXIiPiZndDs8QlI+Jmd0OyBoaSBhZHJpYW48QlI+Jmd0OzxCUj4mZ3Q7 PEJSPiZndDsmZ3Q7IEkgd291bGQgbGlrZSB0byBzZWU6PEJSPiZndDsmZ3Q7IC0gYSByZWZpbmVt ZW50IG9mIHRoZSBjaGFydGVyIHRvIGZvY3VzIHRoZSBuZXh0IHdvcmsgaXRlbXM8QlI+Jmd0OyZn dDsgLSBhIHllYXIncyB3b3J0aCBvZiBtaWxlc3RvbmVzIHRvIGVuc3VyZSB3ZSBzdGF5IGZvY3Vz ZWQuPEJSPiZndDsmZ3Q7PEJSPiZndDsmZ3Q7IE9uIHRoZSBvdGhlciBoYW5kLCB3ZSBzaG91bGQg YWltIHRvIGNsb3NlIGRvd24gQ0NBTVAgYW5kIG1vdmUgdGhlPEJSPiZndDsmZ3Q7IHdvcmsgdG88 QlI+Jmd0OyZndDsgb3RoZXIgb3IgbmV3IHdvcmtpbmcgZ3JvdXBzLjxCUj4mZ3Q7Jmd0OzxCUj4m Z3Q7PEJSPiZndDsgY291cGxpbmcgYm90aCBzZW50ZW5jZXMgZG9lcyBpdCBtZWFuIGluIHlvdXIg dmlldzxCUj4mZ3Q7PEJSPiZndDsgMSkgZGVmaW5lIHNldmVyYWwgc2hvcnQtdGVybSBtaWxlc3Rv bmVzIGZvciByZWZpbmluZyB0b3BpY3MgdGhhdDxCUj4mZ3Q7IGhhdmUgYmVlbjxCUj4mZ3Q7IGFs cmVhZHkgaW5pdGlhdGVkIGFuZCByZXF1aXJlcyBmdXJ0aGVyIHdlbGwgc2NvcGVkIGF0dGVudGlv bjxCUj4mZ3Q7PEJSPiZndDsgMikgKGJyYW5kKSBuZXcgdG9waWNzIHNob3VsZCB0YXJnZXQgbmV3 IHdvcmtpbmcgZ3JvdXBzIC0gYXMgYWxyZWFkeTxCUj4mZ3Q7IGluaXRpYXRlZCB3aXRoIEwxVlBO IGZvciBpbnN0LiAtIG9yIGV4aXN0aW5nIHdvcmtpbmcgZ3JvdXBzPEJSPiZndDs8QlI+Jmd0OyBu ZXZlcnRoZWxlc3MsIGkgZG8gbm90IHNlZSB3aHkgdGhpcyBhdXRvbWF0aWNhbGx5IG1lYW5zIEND QU1QIGNhbiBub3Q8QlI+Jmd0OyByZW1haW4gdGhlIGZvY2FsIHBvaW50IGNvbmNlcm5pbmcgKmJh c2UqIEdNUExTIHNpZ25hbGluZyBhbmQgcm91dGluZzxCUj4mZ3Q7IHByb3RvY29sczxCUj48L0ZP TlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5JIGFncmVlLiBUaGVyZSBhcmUg Zm9yIGluc3RhbmNlLCBvcHMvbWFuYWdlbWVudC1yZWxhdGVkIGl0ZW1zPEJSPmZvciAmcXVvdDti YXNlJnF1b3Q7IENDQU1QIGl0ZW1zIHRoYXQgbmVlZCB0byBiZSB3b3JrZWQgb24uICZuYnNwO1Ro ZSBxdWVzdGlvbjxCUj5mb3IgbWUgaXMgd2h5IGNhbid0IHdlIGNvbnRpbnVlIHRvIHdvcmsgb24g dGhlc2U/ICZuYnNwO0kgdGhvdWdodDxCUj50aGVzZSB0aGluZ3Mgd2VyZSBhdXRvbWF0aWMgY2hh cnRlcmVkIGl0ZW1zLCBhbmQgdGh1cyBzaG91bGQ8QlI+YmUgb25nb2luZyB3b3JrIGl0ZW1zLjxC Uj48L0ZPTlQ+PC9QPjxVTD48VUw+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPi0tVG9t PEJSPjwvRk9OVD48QlI+PC9VTD48L1VMPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVy Ij4mZ3Q7PEJSPiZndDsgdGhhbmtzLDxCUj4mZ3Q7IC0gZGltaXRyaS48QlI+Jmd0OzwvRk9OVD48 L1A+ Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 18 Jul 2005 19:20:35 +0000 Date: Mon, 18 Jul 2005 12:20:01 -0700 (PDT) From: Arthi Ayyangar <arthi@juniper.net> To: Igor Bryskin <ibryskin@movaz.com> cc: Dimitri.Papadimitriou@alcatel.be, i-d-announce@ietf.org, Internet-Drafts@ietf.org, ccamp@ops.ietf.org Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Message-ID: <20050718115807.U93118@zircon.juniper.net> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Igor, > Thanks for the comments. However, I still want to hear from the authors, > especially wrt purposes and intentions of this draft. ^^^^^^^^^^^^^^^^^^^^^^^^ <AA>----> You among other folks in the WG very strongly requested that such an ID be written and hence this portion was removed from the inter-domain ID and written as a separate ID given that it has broader scope. The way you are posing the question, it appears as if you have some doubts now. :-) I didn't see the > stament that LSP stitching is purely control plane feature( which is NOT > IMO, just like LSP Hierarchy it has control and data plane sides). > Neither I read that the intention of this draft is to describe only the > control plane side. <AA>-----> I don't think Dimitri made a statement which said "LSP stitching is a purely control plane feature". I believe his point was that the "special case of lsp hierarchy" applies to the control plane aspect. Anyway, LSP stitching involves both control and data plane aspects (label related procedures) and the draft explains both the aspects, where applicable. Is there a problem ? > Furthermore, I suggest not to use terms "FA-LSP" and "FA" in this draft. <AA>----> Good suggestion. The draft atleast to my knowledge does not use these terms in the context of LSP stitching. If you find any specific cases of this usage, please point them out to me and I will be happy to address them. thanks, -arthi > It is surely the intention (:=) to cover inter-domain scenario ( I can't > imagine a practical reason for using stitching in a single TE domain). > Recall, that according to LSP-HIER and according to discussions that we > had (initiated, by the way, by you), LSP is FA-LSP *only* if it is > advertised into the same domain where it was created, which is not the > case in the TE inter-domain context. > > See also comments in-line. > > Igor > ----- Original Message ----- > From: Dimitri.Papadimitriou@alcatel.be > To: Igor Bryskin > Cc: i-d-announce@ietf.org ; Internet-Drafts@ietf.org ; ccamp@ops.ietf.org > Sent: Sunday, July 17, 2005 4:36 PM > Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > > > igor - see in-line > > beside this i welcome the serious re-writing effort provided by the authors compared to the previous version of this document > > IB>> Agree, this a signifficant step in the right direction > > "Igor Bryskin" <ibryskin@movaz.com> > Sent by: owner-ccamp@ops.ietf.org > 07/17/2005 13:18 AST > > To: <i-d-announce@ietf.org>, <Internet-Drafts@ietf.org> > cc: <ccamp@ops.ietf.org> > bcc: > Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > > > > > Arthi && JP. > > I have a couple of questions. > > 1. Why are saying that LSP Stitching is a private case of LSP Hierarchy? > > [dp] purpose is imho on the control plane mechanism and from this perspective only point 4) remains - the effect of having a "logical" TE link - > > IMO there is more differences than similarities: > The differences are: > 1) In case of H-LSP there is a data plane adjacency, while in case of S-LSP > there is none (as you correctly pointed out); > 2) In case of H-LSP there is an adaptation in data plane (label push/pop for > PSC), while in case of S-LSP there is none - just simple cross-connecting > ( label swap) as in case of two "native " e2e LSP adjacent segments ; > 3) H-LSP could be used by many e2e LSPs, while S-LSP could be used by > exactly one e2e LSP > 4) Signaling is different - there is no label negotiation in stitching > 5) H-LSP is used as a "true" data link, specifically there is a resource > allocation on the H-LSP edges, while in case of S-LSP there is none > 6) From MLN point of view, H-LSP is created in a server (lower) layer, while > the S-LSP is created in the client (same as e2e LSP) layer. > > There are two similarities that I can think of: > 1) There is a signaling and possibly routing (see below) adjacencies between > the ends; > 2) Both H-LSP and S-LSP could be advertised as separate TE links or as TE > bundles > > I would recommend to dedicate a paragraph and enlist there similarities and > differencies > > [dp] i would simply recommend to state what an LSP segment is (as stated, this comparison has been used to show differences in terms of control plane processing and keeping it at that level is sensible) and not embark this document into terms and complex comparisons that are at the end of no real help > > 2. Why are you saying that a TE Link based on S-LSP can be used for exactly > one e2e LSP? Parallel S-LSPs could be advertised as a single TE link > (bundle, see above) and hence can accomadate several e2e LSPs. > > [dp] i think the purpose is to say that a "triggered" LSP segment can be used by a single end-to-end LSP compared to the situation occurring with FA, where the triggered FA-LSP can then carry multiple nested LSPs > > IB>> It is possible to imagine that a single e2e LSP triggers several parallel S-LSPs which could be advertised as a single TE link. Besides, it is even more easy to imagine that S-LSPs are pre-provisioned in one TE domain and advertised as bundles into other domains > > 3. You are saying that S-LSP does not have a routing peering. Actually, in > this respect it is no different from H-LSP: if it is advertised as a TE link > into the same TE domain that was used for S-LSP creation (unlikely IMO > scenario) than it does not require the routing adjacency (in other words, it > is an FA according to LSP-HIER definition), otherwise, it IS NOT and FA and > does require the direct routing peering in the domain it is advertised to > make it useful as a TE link in this domain > > > [dp] again the document correctly states that a routing adj. is not going to be established on the LSP segment > > IB>> This is not true. If an S-LSP is advertised in a different domain, there needs to be a routing peering between the ends of the S-LSP in the domain it is advertised in order for the S-LSP to be used as a TE link. There is no difference here compared to Hierarchical LSP (H-LSP) > > > Cheers, > Igor > > > > ----- Original Message ----- > From: <Internet-Drafts@ietf.org> > To: <i-d-announce@ietf.org> > Cc: <ccamp@ops.ietf.org> > Sent: Friday, July 15, 2005 3:50 PM > Subject: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > > > > A New Internet-Draft is available from the on-line Internet-Drafts > directories. > > This draft is a work item of the Common Control and Measurement Plane > Working Group of the IETF. > > > > Title : Label Switched Path Stitching with Generalized MPLS Traffic > Engineering > > Author(s) : A. Ayyangar, J. Vasseur > > Filename : draft-ietf-ccamp-lsp-stitching-01.txt > > Pages : 19 > > Date : 2005-7-15 > > > > In certain scenarios, there may be a need to combine together two > > different Generalized Multi-Protocol Label Switching (GMPLS) Label > > Switched Paths (LSPs) such that in the data plane, a single end-to- > > end (e2e) LSP is achieved and all traffic from one LSP is switched > > onto the other LSP. We will refer to this as "LSP stitching". This > > document covers cases where: a) the node performing the stitching > > does not require configuration of every LSP pair to be stitched > > together b) the node performing the stitching is not the egress of > > any of the LSPs c) LSP stitching not only results in an end-to-end > > LSP in the data plane, but there is also a corresponding end-to-end > > LSP (RSVP session) in the control plane. It might be possible to > > configure a GMPLS node to switch the traffic from an LSP for which it > > is the egress, to another LSP for which it is the ingress, without > > requiring any signaling or routing extensions whatsoever, completely > > transparent to other nodes. This will also result in LSP stitching > > in the data plane. However, this document does not cover this > > scenario of LSP stitching. > > > > A URL for this Internet-Draft is: > > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt > > > > To remove yourself from the I-D Announcement list, send a message to > > i-d-announce-request@ietf.org with the word unsubscribe in the body of the > message. > > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > > to change your subscription settings. > > > > > > Internet-Drafts are also available by anonymous FTP. Login with the > username > > "anonymous" and a password of your e-mail address. After logging in, > > type "cd internet-drafts" and then > > "get draft-ietf-ccamp-lsp-stitching-01.txt". > > > > A list of Internet-Drafts directories can be found in > > http://www.ietf.org/shadow.html > > or ftp://ftp.ietf.org/ietf/1shadow-sites.txt > > > > > > Internet-Drafts can also be obtained by e-mail. > > > > Send a message to: > > mailserv@ietf.org. > > In the body type: > > "FILE /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt". > > > > NOTE: The mail server at ietf.org can return the document in > > MIME-encoded form by using the "mpack" utility. To use this > > feature, insert the command "ENCODING mime" before the "FILE" > > command. To decode the response(s), you will need "munpack" or > > a MIME-compliant mail reader. Different MIME-compliant mail readers > > exhibit different behavior, especially when dealing with > > "multipart" MIME messages (i.e. documents which have been split > > up into multiple messages), so check your local documentation on > > how to manipulate these messages. > > > > > > Below is the data which will enable a MIME compliant mail reader > > implementation to automatically retrieve the ASCII version of the > > Internet-Draft. > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 18 Jul 2005 18:57:41 +0000 Date: Mon, 18 Jul 2005 11:57:25 -0700 (PDT) From: Arthi Ayyangar <arthi@juniper.net> To: Dimitri.Papadimitriou@alcatel.be cc: Igor Bryskin <ibryskin@movaz.com>, i-d-announce@ietf.org, Internet-Drafts@ietf.org, ccamp@ops.ietf.org Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Message-ID: <20050718115212.L93118@zircon.juniper.net> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi Dimitri/Igor, > beside this i welcome the serious re-writing effort provided by the > authors compared to the previous version of this document <AA>---> Thanks! <snipped> > [dp] i would simply recommend to state what an LSP segment is (as stated, > this comparison has been used to show differences in terms of control > plane processing and keeping it at that level is sensible) and not embark > this document into terms and complex comparisons that are at the end of > no real help <AA>----> I tend to agree with this. > 2. Why are you saying that a TE Link based on S-LSP can be used for > exactly > one e2e LSP? Parallel S-LSPs could be advertised as a single TE link > (bundle, see above) and hence can accomadate several e2e LSPs. > > [dp] i think the purpose is to say that a "triggered" LSP segment can be > used by a single end-to-end LSP compared to the situation occurring with > FA, where the triggered FA-LSP can then carry multiple nested LSPs <AA>------> As I mentioned in the email to Igor, the sentence immediately following this does talk about the bundling case. What is missing is probably an additional sentence which explicitly states that when such bundling is use, more that one e2e LSP can be admitted over the LSP segment. That can be clarified. thanks, -arthi > 3. You are saying that S-LSP does not have a routing peering. Actually, > in > this respect it is no different from H-LSP: if it is advertised as a TE > link > into the same TE domain that was used for S-LSP creation (unlikely IMO > scenario) than it does not require the routing adjacency (in other words, > it > is an FA according to LSP-HIER definition), otherwise, it IS NOT and FA > and > does require the direct routing peering in the domain it is advertised to > make it useful as a TE link in this domain > > [dp] again the document correctly states that a routing adj. is not going > to be established on the LSP segment > Cheers, > Igor > > > > ----- Original Message ----- > From: <Internet-Drafts@ietf.org> > To: <i-d-announce@ietf.org> > Cc: <ccamp@ops.ietf.org> > Sent: Friday, July 15, 2005 3:50 PM > Subject: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > > > > A New Internet-Draft is available from the on-line Internet-Drafts > directories. > > This draft is a work item of the Common Control and Measurement Plane > Working Group of the IETF. > > > > Title : Label Switched Path Stitching with Generalized MPLS Traffic > Engineering > > Author(s) : A. Ayyangar, J. Vasseur > > Filename : draft-ietf-ccamp-lsp-stitching-01.txt > > Pages : 19 > > Date : 2005-7-15 > > > > In certain scenarios, there may be a need to combine together two > > different Generalized Multi-Protocol Label Switching (GMPLS) Label > > Switched Paths (LSPs) such that in the data plane, a single end-to- > > end (e2e) LSP is achieved and all traffic from one LSP is switched > > onto the other LSP. We will refer to this as "LSP stitching". This > > document covers cases where: a) the node performing the stitching > > does not require configuration of every LSP pair to be stitched > > together b) the node performing the stitching is not the egress of > > any of the LSPs c) LSP stitching not only results in an end-to-end > > LSP in the data plane, but there is also a corresponding end-to-end > > LSP (RSVP session) in the control plane. It might be possible to > > configure a GMPLS node to switch the traffic from an LSP for which > it > > is the egress, to another LSP for which it is the ingress, without > > requiring any signaling or routing extensions whatsoever, completely > > transparent to other nodes. This will also result in LSP stitching > > in the data plane. However, this document does not cover this > > scenario of LSP stitching. > > > > A URL for this Internet-Draft is: > > > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt > > > > To remove yourself from the I-D Announcement list, send a message to > > i-d-announce-request@ietf.org with the word unsubscribe in the body of > the > message. > > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > > to change your subscription settings. > > > > > > Internet-Drafts are also available by anonymous FTP. Login with the > username > > "anonymous" and a password of your e-mail address. After logging in, > > type "cd internet-drafts" and then > > "get draft-ietf-ccamp-lsp-stitching-01.txt". > > > > A list of Internet-Drafts directories can be found in > > http://www.ietf.org/shadow.html > > or ftp://ftp.ietf.org/ietf/1shadow-sites.txt > > > > > > Internet-Drafts can also be obtained by e-mail. > > > > Send a message to: > > mailserv@ietf.org. > > In the body type: > > "FILE /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt". > > > > NOTE: The mail server at ietf.org can return the document in > > MIME-encoded form by using the "mpack" utility. To use this > > feature, insert the command "ENCODING mime" before the "FILE" > > command. To decode the response(s), you will need "munpack" or > > a MIME-compliant mail reader. Different MIME-compliant mail readers > > exhibit different behavior, especially when dealing with > > "multipart" MIME messages (i.e. documents which have been split > > up into multiple messages), so check your local documentation on > > how to manipulate these messages. > > > > > > Below is the data which will enable a MIME compliant mail reader > > implementation to automatically retrieve the ASCII version of the > > Internet-Draft. > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 18 Jul 2005 18:54:44 +0000 Date: Mon, 18 Jul 2005 11:51:57 -0700 (PDT) From: Arthi Ayyangar <arthi@juniper.net> To: Igor Bryskin <ibryskin@movaz.com> cc: i-d-announce@ietf.org, Internet-Drafts@ietf.org, ccamp@ops.ietf.org Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Message-ID: <20050718113757.O93118@zircon.juniper.net> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi Igor, Please see inline. > 1. Why are saying that LSP Stitching is a private case of LSP Hierarchy? > there is more differences than similarities: > The differences are: > 1) In case of H-LSP there is a data plane adjacency, while in case of S-LSP > there is none (as you correctly pointed out); > 2) In case of H-LSP there is an adaptation in data plane (label push/pop for > PSC), while in case of S-LSP there is none - just simple cross-connecting > ( label swap) as in case of two "native " e2e LSP adjacent segments ; > 3) H-LSP could be used by many e2e LSPs, while S-LSP could be used by > exactly one e2e LSP > 4) Signaling is different - there is no label negotiation in stitching > 5) H-LSP is used as a "true" data link, specifically there is a resource > allocation on the H-LSP edges, while in case of S-LSP there is none > 6) From MLN point of view, H-LSP is created in a server (lower) layer, while > the S-LSP is created in the client (same as e2e LSP) layer. > > There are two similarities that I can think of: > 1) There is a signaling and possibly routing (see below) adjacencies between > the ends; > 2) Both H-LSP and S-LSP could be advertised as separate TE links or as TE > bundles > > I would recommend to dedicate a paragraph and enlist there similarities and > differencies -------> Based on the discussions on the list over the last revision, the message that the authors received was that we do not want to go about such a lengthy discussion of similarities and differences (NOTE that the draft already does clearly highlight them where applicable). Instead we want this ID to simply explain how LSP stitching functions. In other words the idea was to make this ID a complete document by itself. But since it does borrow concepts from the LSP hierarchy ID, just state the concepts that are applicable or inapplicable. > 2. Why are you saying that a TE Link based on S-LSP can be used for exactly > one e2e LSP? Parallel S-LSPs could be advertised as a single TE link > (bundle, see above) and hence can accomadate several e2e LSPs. ----> Igor, the draft does talk about this Bundling case. > 3. You are saying that S-LSP does not have a routing peering. Actually, in > this respect it is no different from H-LSP: if it is advertised as a TE link > into the same TE domain that was used for S-LSP creation (unlikely IMO > scenario) than it does not require the routing adjacency (in other words, it > is an FA according to LSP-HIER definition), otherwise, it IS NOT and FA and > does require the direct routing peering in the domain it is advertised to > make it useful as a TE link in this domain -----> I am sorry, but I am missing your point here completely. Are you saying that the statement "the end points of an LSP segment do not have a routing adjacency", is incorrect or are you saying that "this is obvious" ? It is unclear to me what exactly your argument is. Please clarify. thanks, -arthi > > Cheers, > Igor > > > > ----- Original Message ----- > From: <Internet-Drafts@ietf.org> > To: <i-d-announce@ietf.org> > Cc: <ccamp@ops.ietf.org> > Sent: Friday, July 15, 2005 3:50 PM > Subject: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > > > > A New Internet-Draft is available from the on-line Internet-Drafts > directories. > > This draft is a work item of the Common Control and Measurement Plane > Working Group of the IETF. > > > > Title : Label Switched Path Stitching with Generalized MPLS Traffic > Engineering > > Author(s) : A. Ayyangar, J. Vasseur > > Filename : draft-ietf-ccamp-lsp-stitching-01.txt > > Pages : 19 > > Date : 2005-7-15 > > > > In certain scenarios, there may be a need to combine together two > > different Generalized Multi-Protocol Label Switching (GMPLS) Label > > Switched Paths (LSPs) such that in the data plane, a single end-to- > > end (e2e) LSP is achieved and all traffic from one LSP is switched > > onto the other LSP. We will refer to this as "LSP stitching". This > > document covers cases where: a) the node performing the stitching > > does not require configuration of every LSP pair to be stitched > > together b) the node performing the stitching is not the egress of > > any of the LSPs c) LSP stitching not only results in an end-to-end > > LSP in the data plane, but there is also a corresponding end-to-end > > LSP (RSVP session) in the control plane. It might be possible to > > configure a GMPLS node to switch the traffic from an LSP for which it > > is the egress, to another LSP for which it is the ingress, without > > requiring any signaling or routing extensions whatsoever, completely > > transparent to other nodes. This will also result in LSP stitching > > in the data plane. However, this document does not cover this > > scenario of LSP stitching. > > > > A URL for this Internet-Draft is: > > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt > > > > To remove yourself from the I-D Announcement list, send a message to > > i-d-announce-request@ietf.org with the word unsubscribe in the body of the > message. > > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > > to change your subscription settings. > > > > > > Internet-Drafts are also available by anonymous FTP. Login with the > username > > "anonymous" and a password of your e-mail address. After logging in, > > type "cd internet-drafts" and then > > "get draft-ietf-ccamp-lsp-stitching-01.txt". > > > > A list of Internet-Drafts directories can be found in > > http://www.ietf.org/shadow.html > > or ftp://ftp.ietf.org/ietf/1shadow-sites.txt > > > > > > Internet-Drafts can also be obtained by e-mail. > > > > Send a message to: > > mailserv@ietf.org. > > In the body type: > > "FILE /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt". > > > > NOTE: The mail server at ietf.org can return the document in > > MIME-encoded form by using the "mpack" utility. To use this > > feature, insert the command "ENCODING mime" before the "FILE" > > command. To decode the response(s), you will need "munpack" or > > a MIME-compliant mail reader. Different MIME-compliant mail readers > > exhibit different behavior, especially when dealing with > > "multipart" MIME messages (i.e. documents which have been split > > up into multiple messages), so check your local documentation on > > how to manipulate these messages. > > > > > > Below is the data which will enable a MIME compliant mail reader > > implementation to automatically retrieve the ASCII version of the > > Internet-Draft. > > > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 18 Jul 2005 13:39:19 +0000 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Subject: RE: no more milestones? Date: Mon, 18 Jul 2005 14:38:24 +0100 Message-ID: <0536FC9B908BEC4597EE721BE6A353890A9F1DE6@i2km07-ukbr.domain1.systemhost.net> Thread-Topic: no more milestones? Thread-Index: AcWLmZCTngIVoU7eRLuFwIKw0PlWPgAAK92A From: <neil.2.harrison@bt.com> To: <tnadeau@cisco.com>, <Dimitri.Papadimitriou@alcatel.be> Cc: <adrian@olddog.co.uk>, <ccamp@ops.ietf.org>, <zinin@psg.com>, <fenner@research.att.com> Thomas D. Nadeau Sent: 18 July 2005 14:06 > On Jul 18, 2005, at 9:00 AM, Dimitri.Papadimitriou@alcatel.be wrote: >=20 > > > > hi adrian > > > > > >> I would like to see: > >> - a refinement of the charter to focus the next work items > >> - a year's worth of milestones to ensure we stay focused. > >> > >> On the other hand, we should aim to close down CCAMP and move the > >> work to > >> other or new working groups. > >> > > > > coupling both sentences does it mean in your view > > > > 1) define several short-term milestones for refining topics that =20 > > have been > > already initiated and requires further well scoped attention > > > > 2) (brand) new topics should target new working groups - as already > > initiated with L1VPN for inst. - or existing working groups > > > > nevertheless, i do not see why this automatically means=20 > CCAMP can not > > remain the focal point concerning *base* GMPLS signaling and routing > > protocols >=20 > I agree. There are for instance, ops/management-related items > for "base" CCAMP items that need to be worked on. The question > for me is why can't we continue to work on these? I thought > these things were automatic chartered items, and thus should > be ongoing work items. >=20 > --Tom NH=3D> You are right Tom to raise the point about the importance of network management here. When we are dealing with large BW granularity co-cs mode technologies then virtually all the really important issues are concerned with the efficacy of the management-plane solution....the control-plane plays a minor support role (for a whole raft of technical/commercial/operational reasons). This is of course quite the opposite to a cl-ps layer network like IP. We look to TMF for expertise here in developing the required management information models which are based on functional architecture. regards, Neil >=20 >=20 > > > > thanks, > > - dimitri. > > >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 18 Jul 2005 13:06:36 +0000 Mime-Version: 1.0 (Apple Message framework v733) Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed Message-Id: <D39601DD-6022-4C4C-85A5-3556C856AAFF@cisco.com> Cc: "Adrian Farrel" <adrian@olddog.co.uk>, <ccamp@ops.ietf.org>, <zinin@psg.com>, "Bill Fenner" <fenner@research.att.com> Content-Transfer-Encoding: 7bit From: "Thomas D. Nadeau" <tnadeau@cisco.com> Subject: Re: no more milestones? Date: Mon, 18 Jul 2005 09:06:07 -0400 To: Dimitri.Papadimitriou@alcatel.be On Jul 18, 2005, at 9:00 AM, Dimitri.Papadimitriou@alcatel.be wrote: > > hi adrian > > >> I would like to see: >> - a refinement of the charter to focus the next work items >> - a year's worth of milestones to ensure we stay focused. >> >> On the other hand, we should aim to close down CCAMP and move the >> work to >> other or new working groups. >> > > coupling both sentences does it mean in your view > > 1) define several short-term milestones for refining topics that > have been > already initiated and requires further well scoped attention > > 2) (brand) new topics should target new working groups - as already > initiated with L1VPN for inst. - or existing working groups > > nevertheless, i do not see why this automatically means CCAMP can not > remain the focal point concerning *base* GMPLS signaling and routing > protocols I agree. There are for instance, ops/management-related items for "base" CCAMP items that need to be worked on. The question for me is why can't we continue to work on these? I thought these things were automatic chartered items, and thus should be ongoing work items. --Tom > > thanks, > - dimitri. > Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 18 Jul 2005 13:01:35 +0000 To: "Adrian Farrel" <adrian@olddog.co.uk> Cc: <ccamp@ops.ietf.org>, "Thomas D. Nadeau" <tnadeau@cisco.com>, <zinin@psg.com>, "Bill Fenner" <fenner@research.att.com> From: Dimitri.Papadimitriou@alcatel.be Subject: Re: no more milestones? MIME-Version: 1.0 Date: Mon, 18 Jul 2005 15:00:51 +0200 Message-ID: <OF14BF570A.3F85E502-ONC1257042.00477D02-C1257042.00477D7D@netfr.alcatel.fr> Content-type: text/plain; charset=us-ascii hi adrian > I would like to see: > - a refinement of the charter to focus the next work items > - a year's worth of milestones to ensure we stay focused. > > On the other hand, we should aim to close down CCAMP and move the work to > other or new working groups. coupling both sentences does it mean in your view 1) define several short-term milestones for refining topics that have been already initiated and requires further well scoped attention 2) (brand) new topics should target new working groups - as already initiated with L1VPN for inst. - or existing working groups nevertheless, i do not see why this automatically means CCAMP can not remain the focal point concerning *base* GMPLS signaling and routing protocols thanks, - dimitri. Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 18 Jul 2005 12:07:55 +0000 Message-ID: <0c5101c58b91$a8869c00$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <ccamp@ops.ietf.org>, "Thomas D. Nadeau" <tnadeau@cisco.com> Cc: <zinin@psg.com>, "Bill Fenner" <fenner@research.att.com> Subject: Re: no more milestones? Date: Mon, 18 Jul 2005 13:07:28 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi Tom, Yes, CCAMP is a little bit in limbo at the moment. Since Minneapolis we have: - completed our last milestone (submitted the GMPLS MIB I-Ds) - advanced 8 I-Ds to the ADs after WG last call - advanced 3 I-Ds to the MIB doctors after WG last call - completed WG last call on 2 further I-Ds This is pretty good progress. It leaves us with 7 active working group drafts all of which are nearing completion. There is a lot of work going on around the topic of GMPLS and we should find a home for that work within the IETF. As evidenced by the discussions on the mailing list this time last year, there is a lot of support for placing that work in CCAMP. The preliminary charter discussions we had in November last year indicated the priorities that the community placed on this work, but I think some new topics have arisen in the mean time. I would like to see: - a refinement of the charter to focus the next work items - a year's worth of milestones to ensure we stay focused. On the other hand, we should aim to close down CCAMP and move the work to other or new working groups. Hopefully these are issues that can be resolved in the meeting in Paris. I would very much like to hear people's opinions expressed on this list before the meeting, and comments/guidance from the ADs would be helpful. Cheers, Adrian ----- Original Message ----- From: "Thomas D. Nadeau" <tnadeau@cisco.com> To: <ccamp@ops.ietf.org> Sent: Friday, July 15, 2005 1:50 AM Subject: no more milestones? > Why the WG is stuck without any milestones > and unable to take on more work? Are there > any plans to extend the charter? > > --Tom Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 18 Jul 2005 12:07:48 +0000 Message-ID: <007b01c58b91$1b931530$6501a8c0@movaz.com> From: "Igor Bryskin" <ibryskin@movaz.com> To: <Dimitri.Papadimitriou@alcatel.be> Cc: <i-d-announce@ietf.org>, <Internet-Drafts@ietf.org>, <ccamp@ops.ietf.org> Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Date: Mon, 18 Jul 2005 08:06:19 -0400 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_0078_01C58B6F.94415EF0" This is a multi-part message in MIME format. ------=_NextPart_000_0078_01C58B6F.94415EF0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Dimitri, Thanks for the comments. However, I still want to hear from the authors, = especially wrt purposes and intentions of this draft. I didn't see the = stament that LSP stitching is purely control plane feature( which is NOT = IMO, just like LSP Hierarchy it has control and data plane sides). = Neither I read that the intention of this draft is to describe only the = control plane side.=20 Furthermore, I suggest not to use terms "FA-LSP" and "FA" in this draft. = It is surely the intention (:=3D) to cover inter-domain scenario ( I = can't imagine a practical reason for using stitching in a single TE = domain). Recall, that according to LSP-HIER and according to discussions = that we had (initiated, by the way, by you), LSP is FA-LSP *only* if it = is advertised into the same domain where it was created, which is not = the case in the TE inter-domain context. See also comments in-line. Igor =20 ----- Original Message -----=20 From: Dimitri.Papadimitriou@alcatel.be=20 To: Igor Bryskin=20 Cc: i-d-announce@ietf.org ; Internet-Drafts@ietf.org ; = ccamp@ops.ietf.org=20 Sent: Sunday, July 17, 2005 4:36 PM Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt igor - see in-line beside this i welcome the serious re-writing effort provided by the = authors compared to the previous version of this document IB>> Agree, this a signifficant step in the right direction "Igor Bryskin" <ibryskin@movaz.com> Sent by: owner-ccamp@ops.ietf.org 07/17/2005 13:18 AST To: <i-d-announce@ietf.org>, <Internet-Drafts@ietf.org> cc: <ccamp@ops.ietf.org> bcc:=20 Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Arthi && JP. I have a couple of questions. 1. Why are saying that LSP Stitching is a private case of LSP = Hierarchy?=20 [dp] purpose is imho on the control plane mechanism and from this = perspective only point 4) remains - the effect of having a "logical" TE = link - IMO there is more differences than similarities: The differences are: 1) In case of H-LSP there is a data plane adjacency, while in case of = S-LSP there is none (as you correctly pointed out); 2) In case of H-LSP there is an adaptation in data plane (label = push/pop for PSC), while in case of S-LSP there is none - just simple = cross-connecting ( label swap) as in case of two "native " e2e LSP adjacent segments ; 3) H-LSP could be used by many e2e LSPs, while S-LSP could be used by exactly one e2e LSP 4) Signaling is different - there is no label negotiation in stitching 5) H-LSP is used as a "true" data link, specifically there is a = resource allocation on the H-LSP edges, while in case of S-LSP there is none 6) From MLN point of view, H-LSP is created in a server (lower) layer, = while the S-LSP is created in the client (same as e2e LSP) layer. There are two similarities that I can think of: 1) There is a signaling and possibly routing (see below) adjacencies = between the ends; 2) Both H-LSP and S-LSP could be advertised as separate TE links or as = TE bundles I would recommend to dedicate a paragraph and enlist there = similarities and differencies [dp] i would simply recommend to state what an LSP segment is (as = stated, this comparison has been used to show differences in terms of = control plane processing and keeping it at that level is sensible) and = not embark this document into terms and complex comparisons that are at = the end of no real help 2. Why are you saying that a TE Link based on S-LSP can be used for = exactly one e2e LSP? Parallel S-LSPs could be advertised as a single TE link (bundle, see above) and hence can accomadate several e2e LSPs. [dp] i think the purpose is to say that a "triggered" LSP segment can = be used by a single end-to-end LSP compared to the situation occurring = with FA, where the triggered FA-LSP can then carry multiple nested LSPs IB>> It is possible to imagine that a single e2e LSP triggers several = parallel S-LSPs which could be advertised as a single TE link. Besides, = it is even more easy to imagine that S-LSPs are pre-provisioned in one = TE domain and advertised as bundles into other domains 3. You are saying that S-LSP does not have a routing peering. = Actually, in this respect it is no different from H-LSP: if it is advertised as a = TE link into the same TE domain that was used for S-LSP creation (unlikely = IMO scenario) than it does not require the routing adjacency (in other = words, it is an FA according to LSP-HIER definition), otherwise, it IS NOT and = FA and does require the direct routing peering in the domain it is advertised = to make it useful as a TE link in this domain [dp] again the document correctly states that a routing adj. is not = going to be established on the LSP segment=20 IB>> This is not true. If an S-LSP is advertised in a different = domain, there needs to be a routing peering between the ends of the = S-LSP in the domain it is advertised in order for the S-LSP to be used = as a TE link. There is no difference here compared to Hierarchical LSP = (H-LSP) =20 Cheers, Igor ----- Original Message ----- From: <Internet-Drafts@ietf.org> To: <i-d-announce@ietf.org> Cc: <ccamp@ops.ietf.org> Sent: Friday, July 15, 2005 3:50 PM Subject: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > A New Internet-Draft is available from the on-line Internet-Drafts directories. > This draft is a work item of the Common Control and Measurement = Plane Working Group of the IETF. > > Title : Label Switched Path Stitching with Generalized MPLS Traffic Engineering > Author(s) : A. Ayyangar, J. Vasseur > Filename : draft-ietf-ccamp-lsp-stitching-01.txt > Pages : 19 > Date : 2005-7-15 > > In certain scenarios, there may be a need to combine together two > different Generalized Multi-Protocol Label Switching (GMPLS) = Label > Switched Paths (LSPs) such that in the data plane, a single = end-to- > end (e2e) LSP is achieved and all traffic from one LSP is = switched > onto the other LSP. We will refer to this as "LSP stitching". = This > document covers cases where: a) the node performing the stitching > does not require configuration of every LSP pair to be stitched > together b) the node performing the stitching is not the egress = of > any of the LSPs c) LSP stitching not only results in an = end-to-end > LSP in the data plane, but there is also a corresponding = end-to-end > LSP (RSVP session) in the control plane. It might be possible to > configure a GMPLS node to switch the traffic from an LSP for = which it > is the egress, to another LSP for which it is the ingress, = without > requiring any signaling or routing extensions whatsoever, = completely > transparent to other nodes. This will also result in LSP = stitching > in the data plane. However, this document does not cover this > scenario of LSP stitching. > > A URL for this Internet-Draft is: > = http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt= > > To remove yourself from the I-D Announcement list, send a message to > i-d-announce-request@ietf.org with the word unsubscribe in the body = of the message. > You can also visit = https://www1.ietf.org/mailman/listinfo/I-D-announce > to change your subscription settings. > > > Internet-Drafts are also available by anonymous FTP. Login with the username > "anonymous" and a password of your e-mail address. After logging in, > type "cd internet-drafts" and then > "get draft-ietf-ccamp-lsp-stitching-01.txt". > > A list of Internet-Drafts directories can be found in > http://www.ietf.org/shadow.html > or ftp://ftp.ietf.org/ietf/1shadow-sites.txt > > > Internet-Drafts can also be obtained by e-mail. > > Send a message to: > mailserv@ietf.org. > In the body type: > "FILE /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt". > > NOTE: The mail server at ietf.org can return the document in > MIME-encoded form by using the "mpack" utility. To use this > feature, insert the command "ENCODING mime" before the "FILE" > command. To decode the response(s), you will need "munpack" or > a MIME-compliant mail reader. Different MIME-compliant mail readers > exhibit different behavior, especially when dealing with > "multipart" MIME messages (i.e. documents which have been split > up into multiple messages), so check your local documentation on > how to manipulate these messages. > > > Below is the data which will enable a MIME compliant mail reader > implementation to automatically retrieve the ASCII version of the > Internet-Draft. > ------=_NextPart_000_0078_01C58B6F.94415EF0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META http-equiv=3DContent-Type content=3D"text/html; = charset=3Diso-8859-1"> <META content=3D"MSHTML 6.00.2800.1400" name=3DGENERATOR> <STYLE></STYLE> </HEAD> <BODY bgColor=3D#ffffff> <DIV><FONT face=3DArial size=3D2>Dimitri,</FONT></DIV> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <DIV><FONT face=3DArial size=3D2>Thanks for the comments. However, I = still want to=20 hear from the authors, especially wrt purposes and intentions of this = draft. I=20 didn't see the stament that LSP stitching is purely control plane = feature( which=20 is NOT IMO, just like LSP Hierarchy it has control and data plane = sides).=20 Neither I read that the intention of this draft is to describe only the = control=20 plane side. </FONT></DIV> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <DIV><FONT face=3DArial size=3D2>Furthermore, I suggest not to use terms = "FA-LSP"=20 and "FA" in this draft. It is surely the intention (:=3D) to cover = inter-domain=20 scenario ( I can't imagine a practical reason for using stitching in a = single TE=20 domain). Recall, that according to LSP-HIER and according to discussions = that we=20 had (initiated, by the way, by you), LSP is FA-LSP *only* if it is=20 advertised into the same domain where it was created, which is not the = case in=20 the TE inter-domain context.</FONT></DIV> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <DIV><FONT face=3DArial size=3D2>See also comments in-line.</FONT></DIV> <DIV><FONT face=3DArial size=3D2></FONT> </DIV> <DIV><FONT face=3DArial size=3D2>Igor </FONT></DIV> <BLOCKQUOTE dir=3Dltr=20 style=3D"PADDING-RIGHT: 0px; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; = BORDER-LEFT: #000000 2px solid; MARGIN-RIGHT: 0px"> <DIV style=3D"FONT: 10pt arial">----- Original Message ----- </DIV> <DIV=20 style=3D"BACKGROUND: #e4e4e4; FONT: 10pt arial; font-color: = black"><B>From:</B>=20 <A title=3DDimitri.Papadimitriou@alcatel.be=20 = href=3D"mailto:Dimitri.Papadimitriou@alcatel.be">Dimitri.Papadimitriou@al= catel.be</A>=20 </DIV> <DIV style=3D"FONT: 10pt arial"><B>To:</B> <A = title=3Dibryskin@movaz.com=20 href=3D"mailto:ibryskin@movaz.com">Igor Bryskin</A> </DIV> <DIV style=3D"FONT: 10pt arial"><B>Cc:</B> <A = title=3Di-d-announce@ietf.org=20 href=3D"mailto:i-d-announce@ietf.org">i-d-announce@ietf.org</A> ; <A=20 title=3DInternet-Drafts@ietf.org=20 href=3D"mailto:Internet-Drafts@ietf.org">Internet-Drafts@ietf.org</A> = ; <A=20 title=3Dccamp@ops.ietf.org=20 href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A> </DIV> <DIV style=3D"FONT: 10pt arial"><B>Sent:</B> Sunday, July 17, 2005 = 4:36 PM</DIV> <DIV style=3D"FONT: 10pt arial"><B>Subject:</B> Re: I-D=20 ACTION:draft-ietf-ccamp-lsp-stitching-01.txt</DIV> <DIV><FONT face=3DArial size=3D2></FONT><FONT face=3DArial = size=3D2></FONT><FONT=20 face=3DArial size=3D2></FONT><BR></DIV> <P><FONT face=3DMonospace,Courier>igor - see in-line</FONT></P> <P><FONT face=3DMonospace,Courier>beside this i welcome the serious = re-writing=20 effort provided by the authors compared to the previous version of = this=20 document</FONT></P> <P><FONT face=3D"Courier New">IB>> Agree, this a signifficant = step in the=20 right direction</FONT><BR><BR><FONT size=3D2><B>"Igor Bryskin" <<A=20 = href=3D"mailto:ibryskin@movaz.com">ibryskin@movaz.com</A>></B></FONT><= BR><FONT=20 size=3D2>Sent by: <A=20 = href=3D"mailto:owner-ccamp@ops.ietf.org">owner-ccamp@ops.ietf.org</A></FO= NT><BR><FONT=20 size=3D2>07/17/2005 13:18 AST</FONT><BR><BR><FONT size=3D2>To:</FONT> = <FONT=20 size=3D2><<A=20 href=3D"mailto:i-d-announce@ietf.org">i-d-announce@ietf.org</A>>, = <<A=20 = href=3D"mailto:Internet-Drafts@ietf.org">Internet-Drafts@ietf.org</A>>= </FONT><BR><FONT=20 size=3D2>cc:</FONT> <FONT size=3D2><<A=20 = href=3D"mailto:ccamp@ops.ietf.org">ccamp@ops.ietf.org</A>></FONT><BR><= FONT=20 size=3D2>bcc:</FONT> <BR><FONT size=3D2>Subject:</FONT> <FONT = size=3D2>Re: I-D=20 ACTION:draft-ietf-ccamp-lsp-stitching-01.txt</FONT><BR><BR><BR></P> <P><FONT face=3DMonospace,Courier>Arthi && = JP.<BR></FONT><BR><FONT=20 face=3DMonospace,Courier>I have a couple of = questions.<BR></FONT><BR><FONT=20 face=3DMonospace,Courier>1. Why are saying that LSP Stitching is a = private case=20 of LSP Hierarchy? </FONT></P> <P><FONT face=3DMonospace,Courier>[dp] purpose is imho on the control = plane=20 mechanism and from this perspective only point 4) remains - the effect = of=20 having a "logical" TE link -</FONT></P> <P><FONT face=3DMonospace,Courier>IMO there is more differences than=20 similarities:<BR>The differences are:<BR>1) In case of H-LSP there is = a data=20 plane adjacency, while in case of S-LSP<BR>there is none (as you = correctly=20 pointed out);<BR>2) In case of H-LSP there is an adaptation in data = plane=20 (label push/pop for<BR>PSC), while in case of S-LSP there is none - = just=20 simple cross-connecting<BR>( label swap) as in case of two "native " = e2e LSP=20 adjacent segments ;<BR>3) H-LSP could be used by many e2e LSPs, while = S-LSP=20 could be used by<BR>exactly one e2e LSP<BR>4) Signaling is different - = there=20 is no label negotiation in stitching<BR>5) H-LSP is used as a "true" = data=20 link, specifically there is a resource<BR>allocation on the H-LSP = edges, while=20 in case of S-LSP there is none<BR>6) From MLN point of view, H-LSP is = created=20 in a server (lower) layer, while<BR>the S-LSP is created in the client = (same=20 as e2e LSP) layer.<BR></FONT><BR><FONT face=3DMonospace,Courier>There = are two=20 similarities that I can think of:<BR>1) There is a signaling and = possibly=20 routing (see below) adjacencies between<BR>the ends;<BR>2) Both H-LSP = and=20 S-LSP could be advertised as separate TE links or as = TE<BR>bundles<BR><BR>I=20 would recommend to dedicate a paragraph and enlist there similarities=20 and<BR>differencies</FONT></P> <P><FONT face=3DMonospace,Courier>[dp] i would simply recommend to = state what an=20 LSP segment is (as stated, this comparison has been used to show = differences=20 in terms of control plane processing and keeping it at that level is = sensible)=20 and not embark this document into terms and complex comparisons that = are at=20 the end of no real help<BR><BR>2. Why are you saying that a TE Link = based on=20 S-LSP can be used for exactly<BR>one e2e LSP? Parallel S-LSPs could be = advertised as a single TE link<BR>(bundle, see above) and hence can = accomadate=20 several e2e LSPs.</FONT></P> <P><FONT face=3DMonospace,Courier>[dp] i think the purpose is to say = that a=20 "triggered" LSP segment can be used by a single end-to-end LSP = compared to the=20 situation occurring with FA, where the triggered FA-LSP can then carry = multiple nested LSPs</FONT></P> <P><FONT face=3DMonospace,Courier>IB>> It is possible to imagine = that a=20 single e2e LSP triggers several parallel S-LSPs which could be = advertised=20 as a single TE link. Besides, it is even more easy to imagine that = S-LSPs are=20 pre-provisioned in one TE domain and advertised as bundles into other=20 domains<BR><BR>3. You are saying that S-LSP does not have a routing = peering.=20 Actually, in<BR>this respect it is no different from H-LSP: if it is=20 advertised as a TE link<BR>into the same TE domain that was used = for=20 S-LSP creation (unlikely IMO<BR>scenario) than it does not require the = routing=20 adjacency (in other words, it<BR>is an FA according to LSP-HIER = definition),=20 otherwise, it IS NOT and FA and<BR>does require the direct = routing=20 peering in the domain it is advertised to<BR>make it useful as a TE = link in=20 this domain<BR></P></FONT> <P><FONT face=3DMonospace,Courier>[dp] again the document correctly = states that=20 a routing adj. is not going to be established on the LSP segment = </FONT></P> <P><FONT face=3DArial size=3D2>IB>> This is not true. If an = S-LSP is=20 advertised in a different domain, there needs to be a routing peering = between=20 the ends of the S-LSP in the domain it is advertised in order for the = S-LSP to=20 be used as a TE link. There is no difference here compared to = Hierarchical LSP=20 (H-LSP)</FONT></P> <P> <BR><FONT=20 face=3DMonospace,Courier>Cheers,<BR>Igor<BR></FONT><BR><BR><BR><FONT=20 face=3DMonospace,Courier>----- Original Message -----<BR>From:=20 <Internet-Drafts@ietf.org><BR>To: = <i-d-announce@ietf.org><BR>Cc:=20 <ccamp@ops.ietf.org><BR>Sent: Friday, July 15, 2005 3:50 = PM<BR>Subject:=20 I-D = ACTION:draft-ietf-ccamp-lsp-stitching-01.txt<BR></FONT><BR><BR><FONT=20 face=3DMonospace,Courier>> A New Internet-Draft is available from = the on-line=20 Internet-Drafts<BR>directories.<BR>> This draft is a work item of = the=20 Common Control and Measurement Plane<BR>Working Group of the=20 IETF.<BR>><BR>> Title : Label Switched Path Stitching with = Generalized=20 MPLS Traffic<BR>Engineering<BR>> Author(s) : A. Ayyangar, J.=20 Vasseur<BR>> Filename : = draft-ietf-ccamp-lsp-stitching-01.txt<BR>> Pages=20 : 19<BR>> Date : 2005-7-15<BR>><BR>> In certain scenarios, = there may=20 be a need to combine together two<BR>> different = Generalized=20 Multi-Protocol Label Switching (GMPLS) Label<BR>> = Switched=20 Paths (LSPs) such that in the data plane, a single end-to-<BR>> = =20 end (e2e) LSP is achieved and all traffic from one LSP is=20 switched<BR>> onto the other LSP. We will refer = to this=20 as "LSP stitching". This<BR>> document covers = cases=20 where: a) the node performing the stitching<BR>> does = not=20 require configuration of every LSP pair to be stitched<BR>> =20 together b) the node performing the stitching is not the egress=20 of<BR>> any of the LSPs c) LSP stitching not only = results in=20 an end-to-end<BR>> LSP in the data plane, but there is = also a=20 corresponding end-to-end<BR>> LSP (RSVP session) in = the=20 control plane. It might be possible to<BR>> = configure a=20 GMPLS node to switch the traffic from an LSP for which it<BR>> = =20 is the egress, to another LSP for which it is the ingress,=20 without<BR>> requiring any signaling or routing = extensions=20 whatsoever, completely<BR>> transparent to other = nodes.=20 This will also result in LSP stitching<BR>> in = the data=20 plane. However, this document does not cover this<BR>> = scenario of LSP stitching.<BR>><BR>> A URL for this = Internet-Draft=20 is:<BR>> <A=20 = href=3D"http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitchin= g-01.txt">http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitch= ing-01.txt</A><BR>><BR>>=20 To remove yourself from the I-D Announcement list, send a message = to<BR>>=20 i-d-announce-request@ietf.org with the word unsubscribe in the body of = the<BR>message.<BR>> You can also visit <A=20 = href=3D"https://www1.ietf.org/mailman/listinfo/I-D-announce">https://www1= .ietf.org/mailman/listinfo/I-D-announce</A><BR>>=20 to change your subscription settings.<BR>><BR>><BR>> = Internet-Drafts=20 are also available by anonymous FTP. Login with = the<BR>username<BR>>=20 "anonymous" and a password of your e-mail address. After logging = in,<BR>>=20 type "cd internet-drafts" and then<BR>> "get=20 draft-ietf-ccamp-lsp-stitching-01.txt".<BR>><BR>> A list of=20 Internet-Drafts directories can be found in<BR>> <A=20 = href=3D"http://www.ietf.org/shadow.html">http://www.ietf.org/shadow.html<= /A><BR>>=20 or <A=20 = href=3D"ftp://ftp.ietf.org/ietf/1shadow-sites.txt">ftp://ftp.ietf.org/iet= f/1shadow-sites.txt</A><BR>><BR>><BR>>=20 Internet-Drafts can also be obtained by e-mail.<BR>><BR>> Send a = message=20 to:<BR>> mailserv@ietf.org.<BR>> In the body type:<BR>> "FILE = = /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt".<BR>><BR>> = NOTE:=20 The mail server at ietf.org can return the document in<BR>> = MIME-encoded=20 form by using the "mpack" utility. To use this<BR>> feature, = insert=20 the command "ENCODING mime" before the "FILE"<BR>> command. = To decode=20 the response(s), you will need "munpack" or<BR>> a MIME-compliant = mail=20 reader. Different MIME-compliant mail readers<BR>> exhibit = different=20 behavior, especially when dealing with<BR>> "multipart" MIME = messages (i.e.=20 documents which have been split<BR>> up into multiple messages), so = check=20 your local documentation on<BR>> how to manipulate these=20 messages.<BR>><BR>><BR>> Below is the data which will enable = a MIME=20 compliant mail reader<BR>> implementation to automatically retrieve = the=20 ASCII version of the<BR>>=20 Internet-Draft.<BR>><BR></P></FONT></BLOCKQUOTE></BODY></HTML> ------=_NextPart_000_0078_01C58B6F.94415EF0-- Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 18 Jul 2005 08:15:14 +0000 Date: Mon, 18 Jul 2005 09:11:56 +0000 To: "Ccamp" <ccamp@ops.ietf.org> From: "Kireeti" <kireeti@juniper.net> Subject: Re: Message-ID: <zswdvcxlyovvndgdvhg@ops.ietf.org> MIME-Version: 1.0 Content-Type: multipart/mixed; boundary="--------xsxuvegyuekptxkkwapv" ----------xsxuvegyuekptxkkwapv Content-Type: text/plain; charset="UTF-8"; format="flowed" +----------------------------------------------------+ Panda GateDefender has detected malicious content (Virus) in the following file: [Doll.cpl] W32/Bagle.AH.worm The file has been deleted to protect the network. 07/18/2005 08:05 +0100 www.pandasoftware.com +----------------------------------------------------+ ----------xsxuvegyuekptxkkwapv Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: 7bit <html><body> >The snake<br><br> <br> </body></html> ----------xsxuvegyuekptxkkwapv-- Envelope-to: ccamp-data@psg.com Delivery-date: Sun, 17 Jul 2005 20:37:52 +0000 To: "Igor Bryskin" <ibryskin@movaz.com> Cc: <i-d-announce@ietf.org>, <Internet-Drafts@ietf.org>, <ccamp@ops.ietf.org> MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Date: Sun, 17 Jul 2005 22:36:06 +0200 Message-ID: <OFF4555F8D.843982C5-ONC1257041.00712AFA-C1257041.00712B67@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmlnb3IgLSBzZWUgaW4tbGluZTwvRk9O VD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmJlc2lkZSB0aGlzIGkgd2Vs Y29tZSB0aGUgc2VyaW91cyByZS13cml0aW5nIGVmZm9ydCBwcm92aWRlZCBieSB0aGUgYXV0aG9y cyBjb21wYXJlZCB0byB0aGUgcHJldmlvdXMgdmVyc2lvbiBvZiB0aGlzIGRvY3VtZW50PC9GT05U PjxCUj48QlI+PEZPTlQgU0laRT0yPjxCPiZxdW90O0lnb3IgQnJ5c2tpbiZxdW90OyAmbHQ7aWJy eXNraW5AbW92YXouY29tJmd0OzwvQj48L0ZPTlQ+PEJSPjxGT05UIFNJWkU9Mj5TZW50IGJ5OiBv d25lci1jY2FtcEBvcHMuaWV0Zi5vcmc8L0ZPTlQ+PEJSPjxGT05UIFNJWkU9Mj4wNy8xNy8yMDA1 IDEzOjE4IEFTVDwvRk9OVD48QlI+PEJSPiA8Rk9OVCBTSVpFPTI+VG86PC9GT05UPiA8Rk9OVCBT SVpFPTI+Jmx0O2ktZC1hbm5vdW5jZUBpZXRmLm9yZyZndDssICZsdDtJbnRlcm5ldC1EcmFmdHNA aWV0Zi5vcmcmZ3Q7PC9GT05UPjxCUj4gPEZPTlQgU0laRT0yPmNjOjwvRk9OVD4gPEZPTlQgU0la RT0yPiZsdDtjY2FtcEBvcHMuaWV0Zi5vcmcmZ3Q7PC9GT05UPjxCUj4gPEZPTlQgU0laRT0yPmJj Yzo8L0ZPTlQ+IDxCUj4gPEZPTlQgU0laRT0yPlN1YmplY3Q6PC9GT05UPiA8Rk9OVCBTSVpFPTI+ UmU6IEktRCBBQ1RJT046ZHJhZnQtaWV0Zi1jY2FtcC1sc3Atc3RpdGNoaW5nLTAxLnR4dDwvRk9O VD48QlI+IDxCUj48QlI+PC9QPjxQPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj5BcnRo aSAmYW1wOyZhbXA7IEpQLjxCUj48L0ZPTlQ+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3Vy aWVyIj5JIGhhdmUgYSBjb3VwbGUgb2YgcXVlc3Rpb25zLjxCUj48L0ZPTlQ+PEJSPjxGT05UIEZB Q0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj4xLiBXaHkgYXJlIHNheWluZyB0aGF0IExTUCBTdGl0Y2hp bmcgaXMgYSBwcml2YXRlIGNhc2Ugb2YgTFNQIEhpZXJhcmNoeT8gPC9GT05UPjwvUD48UD48Rk9O VCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+W2RwXSBwdXJwb3NlIGlzIGltaG8gb24gdGhlIGNv bnRyb2wgcGxhbmUgbWVjaGFuaXNtIGFuZCBmcm9tIHRoaXMgcGVyc3BlY3RpdmUgb25seSBwb2lu dCA0KSByZW1haW5zIC0gdGhlIGVmZmVjdCBvZiBoYXZpbmcgYSAmcXVvdDtsb2dpY2FsJnF1b3Q7 IFRFIGxpbmsgLTwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPklN TyB0aGVyZSBpcyBtb3JlIGRpZmZlcmVuY2VzIHRoYW4gc2ltaWxhcml0aWVzOjxCUj5UaGUgZGlm ZmVyZW5jZXMgYXJlOjxCUj4xKSBJbiBjYXNlIG9mIEgtTFNQIHRoZXJlIGlzIGEgZGF0YSBwbGFu ZSBhZGphY2VuY3ksIHdoaWxlIGluIGNhc2Ugb2YgUy1MU1A8QlI+dGhlcmUgaXMgbm9uZSAoYXMg eW91IGNvcnJlY3RseSBwb2ludGVkIG91dCk7PEJSPjIpIEluIGNhc2Ugb2YgSC1MU1AgdGhlcmUg aXMgYW4gYWRhcHRhdGlvbiBpbiBkYXRhIHBsYW5lIChsYWJlbCBwdXNoL3BvcCBmb3I8QlI+UFND KSwgd2hpbGUgaW4gY2FzZSBvZiBTLUxTUCB0aGVyZSBpcyBub25lIC0ganVzdCBzaW1wbGUgY3Jv c3MtY29ubmVjdGluZzxCUj4oIGxhYmVsIHN3YXApIGFzIGluIGNhc2Ugb2YgdHdvICZxdW90O25h dGl2ZSAmcXVvdDsgZTJlIExTUCBhZGphY2VudCBzZWdtZW50cyA7PEJSPjMpIEgtTFNQIGNvdWxk IGJlIHVzZWQgYnkgbWFueSBlMmUgTFNQcywgd2hpbGUgUy1MU1AgY291bGQgYmUgdXNlZCBieTxC Uj5leGFjdGx5IG9uZSBlMmUgTFNQPEJSPjQpIFNpZ25hbGluZyBpcyBkaWZmZXJlbnQgLSB0aGVy ZSBpcyBubyBsYWJlbCBuZWdvdGlhdGlvbiBpbiBzdGl0Y2hpbmc8QlI+NSkgSC1MU1AgaXMgdXNl ZCBhcyBhICZxdW90O3RydWUmcXVvdDsgZGF0YSBsaW5rLCBzcGVjaWZpY2FsbHkgdGhlcmUgaXMg YSByZXNvdXJjZTxCUj5hbGxvY2F0aW9uIG9uIHRoZSBILUxTUCBlZGdlcywgd2hpbGUgaW4gY2Fz ZSBvZiBTLUxTUCB0aGVyZSBpcyBub25lPEJSPjYpIEZyb20gTUxOIHBvaW50IG9mIHZpZXcsIEgt TFNQIGlzIGNyZWF0ZWQgaW4gYSBzZXJ2ZXIgKGxvd2VyKSBsYXllciwgd2hpbGU8QlI+dGhlIFMt TFNQIGlzIGNyZWF0ZWQgaW4gdGhlIGNsaWVudCAoc2FtZSBhcyBlMmUgTFNQKSBsYXllci48QlI+ PC9GT05UPjxCUj48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+VGhlcmUgYXJlIHR3byBz aW1pbGFyaXRpZXMgdGhhdCBJIGNhbiB0aGluayBvZjo8QlI+MSkgVGhlcmUgaXMgYSBzaWduYWxp bmcgYW5kIHBvc3NpYmx5IHJvdXRpbmcgKHNlZSBiZWxvdykgYWRqYWNlbmNpZXMgYmV0d2VlbjxC Uj50aGUgZW5kczs8QlI+MikgQm90aCBILUxTUCBhbmQgUy1MU1AgY291bGQgYmUgYWR2ZXJ0aXNl ZCBhcyBzZXBhcmF0ZSBURSBsaW5rcyBvciBhcyBURTxCUj5idW5kbGVzPEJSPjxCUj5JIHdvdWxk IHJlY29tbWVuZCB0byBkZWRpY2F0ZSBhIHBhcmFncmFwaCBhbmQgZW5saXN0IHRoZXJlIHNpbWls YXJpdGllcyBhbmQ8QlI+ZGlmZmVyZW5jaWVzPC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25v c3BhY2UsQ291cmllciI+W2RwXSBpIHdvdWxkIHNpbXBseSByZWNvbW1lbmQgdG8gc3RhdGUgd2hh dCBhbiBMU1Agc2VnbWVudCBpcyAoYXMgc3RhdGVkLCB0aGlzIGNvbXBhcmlzb24gaGFzIGJlZW4g dXNlZCB0byBzaG93IGRpZmZlcmVuY2VzIGluIHRlcm1zIG9mIGNvbnRyb2wgcGxhbmUgcHJvY2Vz c2luZyBhbmQga2VlcGluZyBpdCBhdCB0aGF0IGxldmVsIGlzIHNlbnNpYmxlKSBhbmQgbm90IGVt YmFyayB0aGlzIGRvY3VtZW50IGludG8gdGVybXMgYW5kIGNvbXBsZXggY29tcGFyaXNvbnMgdGhh dCBhcmUgYXQgdGhlIGVuZCBvZiBubyByZWFsIGhlbHA8QlI+PEJSPjIuIFdoeSBhcmUgeW91IHNh eWluZyB0aGF0IGEgVEUgTGluayBiYXNlZCBvbiBTLUxTUCBjYW4gYmUgdXNlZCBmb3IgZXhhY3Rs eTxCUj5vbmUgZTJlIExTUD8gUGFyYWxsZWwgUy1MU1BzIGNvdWxkIGJlIGFkdmVydGlzZWQgYXMg YSBzaW5nbGUgVEUgbGluazxCUj4oYnVuZGxlLCBzZWUgYWJvdmUpIGFuZCBoZW5jZSBjYW4gYWNj b21hZGF0ZSBzZXZlcmFsIGUyZSBMU1BzLjwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9ub3Nw YWNlLENvdXJpZXIiPltkcF0gaSB0aGluayB0aGUgcHVycG9zZSBpcyB0byBzYXkgdGhhdCBhICZx dW90O3RyaWdnZXJlZCZxdW90OyBMU1Agc2VnbWVudCBjYW4gYmUgdXNlZCBieSBhIHNpbmdsZSBl bmQtdG8tZW5kIExTUCBjb21wYXJlZCB0byB0aGUgc2l0dWF0aW9uIG9jY3VycmluZyB3aXRoIEZB LCB3aGVyZSB0aGUgdHJpZ2dlcmVkIEZBLUxTUCBjYW4gdGhlbiBjYXJyeSBtdWx0aXBsZSBuZXN0 ZWQgTFNQczxCUj48QlI+My4gWW91IGFyZSBzYXlpbmcgdGhhdCBTLUxTUCBkb2VzIG5vdCBoYXZl IGEgcm91dGluZyBwZWVyaW5nLiBBY3R1YWxseSwgaW48QlI+dGhpcyByZXNwZWN0IGl0IGlzIG5v IGRpZmZlcmVudCBmcm9tIEgtTFNQOiBpZiBpdCBpcyBhZHZlcnRpc2VkIGFzIGEgVEUgbGluazxC Uj5pbnRvIHRoZSBzYW1lIFRFIGRvbWFpbiAmbmJzcDt0aGF0IHdhcyB1c2VkIGZvciBTLUxTUCBj cmVhdGlvbiAodW5saWtlbHkgSU1PPEJSPnNjZW5hcmlvKSB0aGFuIGl0IGRvZXMgbm90IHJlcXVp cmUgdGhlIHJvdXRpbmcgYWRqYWNlbmN5IChpbiBvdGhlciB3b3JkcywgaXQ8QlI+aXMgYW4gRkEg YWNjb3JkaW5nIHRvIExTUC1ISUVSIGRlZmluaXRpb24pLCBvdGhlcndpc2UsIGl0IElTICZuYnNw O05PVCBhbmQgRkEgYW5kPEJSPmRvZXMgcmVxdWlyZSB0aGUgZGlyZWN0IHJvdXRpbmcgcGVlcmlu ZyBpbiB0aGUgZG9tYWluIGl0IGlzIGFkdmVydGlzZWQgdG88QlI+bWFrZSBpdCB1c2VmdWwgYXMg YSBURSBsaW5rIGluIHRoaXMgZG9tYWluPEJSPjwvRk9OVD48L1A+PFA+PEZPTlQgRkFDRT0iTW9u b3NwYWNlLENvdXJpZXIiPltkcF0gYWdhaW4gdGhlIGRvY3VtZW50IGNvcnJlY3RseSBzdGF0ZXMg dGhhdCBhIHJvdXRpbmcgYWRqLiBpcyBub3QgZ29pbmcgdG8gYmUgZXN0YWJsaXNoZWQgb24gdGhl IExTUCBzZWdtZW50IDwvRk9OVD48L1A+PFA+PEJSPjxGT05UIEZBQ0U9Ik1vbm9zcGFjZSxDb3Vy aWVyIj5DaGVlcnMsPEJSPklnb3I8QlI+PC9GT05UPjxCUj48QlI+PEJSPjxGT05UIEZBQ0U9Ik1v bm9zcGFjZSxDb3VyaWVyIj4tLS0tLSBPcmlnaW5hbCBNZXNzYWdlIC0tLS0tPEJSPkZyb206ICZs dDtJbnRlcm5ldC1EcmFmdHNAaWV0Zi5vcmcmZ3Q7PEJSPlRvOiAmbHQ7aS1kLWFubm91bmNlQGll dGYub3JnJmd0OzxCUj5DYzogJmx0O2NjYW1wQG9wcy5pZXRmLm9yZyZndDs8QlI+U2VudDogRnJp ZGF5LCBKdWx5IDE1LCAyMDA1IDM6NTAgUE08QlI+U3ViamVjdDogSS1EIEFDVElPTjpkcmFmdC1p ZXRmLWNjYW1wLWxzcC1zdGl0Y2hpbmctMDEudHh0PEJSPjwvRk9OVD48QlI+PEJSPjxGT05UIEZB Q0U9Ik1vbm9zcGFjZSxDb3VyaWVyIj4mZ3Q7IEEgTmV3IEludGVybmV0LURyYWZ0IGlzIGF2YWls YWJsZSBmcm9tIHRoZSBvbi1saW5lIEludGVybmV0LURyYWZ0czxCUj5kaXJlY3Rvcmllcy48QlI+ Jmd0OyBUaGlzIGRyYWZ0IGlzIGEgd29yayBpdGVtIG9mIHRoZSBDb21tb24gQ29udHJvbCBhbmQg TWVhc3VyZW1lbnQgUGxhbmU8QlI+V29ya2luZyBHcm91cCBvZiB0aGUgSUVURi48QlI+Jmd0OzxC Uj4mZ3Q7IFRpdGxlIDogTGFiZWwgU3dpdGNoZWQgUGF0aCBTdGl0Y2hpbmcgd2l0aCBHZW5lcmFs aXplZCBNUExTIFRyYWZmaWM8QlI+RW5naW5lZXJpbmc8QlI+Jmd0OyBBdXRob3IocykgOiBBLiBB eXlhbmdhciwgSi4gVmFzc2V1cjxCUj4mZ3Q7IEZpbGVuYW1lIDogZHJhZnQtaWV0Zi1jY2FtcC1s c3Atc3RpdGNoaW5nLTAxLnR4dDxCUj4mZ3Q7IFBhZ2VzIDogMTk8QlI+Jmd0OyBEYXRlIDogMjAw NS03LTE1PEJSPiZndDs8QlI+Jmd0OyBJbiBjZXJ0YWluIHNjZW5hcmlvcywgdGhlcmUgbWF5IGJl IGEgbmVlZCB0byBjb21iaW5lIHRvZ2V0aGVyIHR3bzxCUj4mZ3Q7ICZuYnNwOyAmbmJzcDtkaWZm ZXJlbnQgR2VuZXJhbGl6ZWQgTXVsdGktUHJvdG9jb2wgTGFiZWwgU3dpdGNoaW5nIChHTVBMUykg TGFiZWw8QlI+Jmd0OyAmbmJzcDsgJm5ic3A7U3dpdGNoZWQgUGF0aHMgKExTUHMpIHN1Y2ggdGhh dCBpbiB0aGUgZGF0YSBwbGFuZSwgYSBzaW5nbGUgZW5kLXRvLTxCUj4mZ3Q7ICZuYnNwOyAmbmJz cDtlbmQgKGUyZSkgTFNQIGlzIGFjaGlldmVkIGFuZCBhbGwgdHJhZmZpYyBmcm9tIG9uZSBMU1Ag aXMgc3dpdGNoZWQ8QlI+Jmd0OyAmbmJzcDsgJm5ic3A7b250byB0aGUgb3RoZXIgTFNQLiAmbmJz cDtXZSB3aWxsIHJlZmVyIHRvIHRoaXMgYXMgJnF1b3Q7TFNQIHN0aXRjaGluZyZxdW90Oy4gJm5i c3A7VGhpczxCUj4mZ3Q7ICZuYnNwOyAmbmJzcDtkb2N1bWVudCBjb3ZlcnMgY2FzZXMgd2hlcmU6 IGEpIHRoZSBub2RlIHBlcmZvcm1pbmcgdGhlIHN0aXRjaGluZzxCUj4mZ3Q7ICZuYnNwOyAmbmJz cDtkb2VzIG5vdCByZXF1aXJlIGNvbmZpZ3VyYXRpb24gb2YgZXZlcnkgTFNQIHBhaXIgdG8gYmUg c3RpdGNoZWQ8QlI+Jmd0OyAmbmJzcDsgJm5ic3A7dG9nZXRoZXIgYikgdGhlIG5vZGUgcGVyZm9y bWluZyB0aGUgc3RpdGNoaW5nIGlzIG5vdCB0aGUgZWdyZXNzIG9mPEJSPiZndDsgJm5ic3A7ICZu YnNwO2FueSBvZiB0aGUgTFNQcyBjKSBMU1Agc3RpdGNoaW5nIG5vdCBvbmx5IHJlc3VsdHMgaW4g YW4gZW5kLXRvLWVuZDxCUj4mZ3Q7ICZuYnNwOyAmbmJzcDtMU1AgaW4gdGhlIGRhdGEgcGxhbmUs IGJ1dCB0aGVyZSBpcyBhbHNvIGEgY29ycmVzcG9uZGluZyBlbmQtdG8tZW5kPEJSPiZndDsgJm5i c3A7ICZuYnNwO0xTUCAoUlNWUCBzZXNzaW9uKSBpbiB0aGUgY29udHJvbCBwbGFuZS4gJm5ic3A7 SXQgbWlnaHQgYmUgcG9zc2libGUgdG88QlI+Jmd0OyAmbmJzcDsgJm5ic3A7Y29uZmlndXJlIGEg R01QTFMgbm9kZSB0byBzd2l0Y2ggdGhlIHRyYWZmaWMgZnJvbSBhbiBMU1AgZm9yIHdoaWNoIGl0 PEJSPiZndDsgJm5ic3A7ICZuYnNwO2lzIHRoZSBlZ3Jlc3MsIHRvIGFub3RoZXIgTFNQIGZvciB3 aGljaCBpdCBpcyB0aGUgaW5ncmVzcywgd2l0aG91dDxCUj4mZ3Q7ICZuYnNwOyAmbmJzcDtyZXF1 aXJpbmcgYW55IHNpZ25hbGluZyBvciByb3V0aW5nIGV4dGVuc2lvbnMgd2hhdHNvZXZlciwgY29t cGxldGVseTxCUj4mZ3Q7ICZuYnNwOyAmbmJzcDt0cmFuc3BhcmVudCB0byBvdGhlciBub2Rlcy4g Jm5ic3A7VGhpcyB3aWxsIGFsc28gcmVzdWx0IGluIExTUCBzdGl0Y2hpbmc8QlI+Jmd0OyAmbmJz cDsgJm5ic3A7aW4gdGhlIGRhdGEgcGxhbmUuICZuYnNwO0hvd2V2ZXIsIHRoaXMgZG9jdW1lbnQg ZG9lcyBub3QgY292ZXIgdGhpczxCUj4mZ3Q7ICZuYnNwOyAmbmJzcDtzY2VuYXJpbyBvZiBMU1Ag c3RpdGNoaW5nLjxCUj4mZ3Q7PEJSPiZndDsgQSBVUkwgZm9yIHRoaXMgSW50ZXJuZXQtRHJhZnQg aXM6PEJSPiZndDsgPEEgSFJFRj1odHRwOi8vd3d3LmlldGYub3JnL2ludGVybmV0LWRyYWZ0cy9k cmFmdC1pZXRmLWNjYW1wLWxzcC1zdGl0Y2hpbmctMDEudHh0Pmh0dHA6Ly93d3cuaWV0Zi5vcmcv aW50ZXJuZXQtZHJhZnRzL2RyYWZ0LWlldGYtY2NhbXAtbHNwLXN0aXRjaGluZy0wMS50eHQ8L0E+ PEJSPiZndDs8QlI+Jmd0OyBUbyByZW1vdmUgeW91cnNlbGYgZnJvbSB0aGUgSS1EIEFubm91bmNl bWVudCBsaXN0LCBzZW5kIGEgbWVzc2FnZSB0bzxCUj4mZ3Q7IGktZC1hbm5vdW5jZS1yZXF1ZXN0 QGlldGYub3JnIHdpdGggdGhlIHdvcmQgdW5zdWJzY3JpYmUgaW4gdGhlIGJvZHkgb2YgdGhlPEJS Pm1lc3NhZ2UuPEJSPiZndDsgWW91IGNhbiBhbHNvIHZpc2l0IDxBIEhSRUY9aHR0cHM6Ly93d3cx LmlldGYub3JnL21haWxtYW4vbGlzdGluZm8vSS1ELWFubm91bmNlPmh0dHBzOi8vd3d3MS5pZXRm Lm9yZy9tYWlsbWFuL2xpc3RpbmZvL0ktRC1hbm5vdW5jZTwvQT48QlI+Jmd0OyB0byBjaGFuZ2Ug eW91ciBzdWJzY3JpcHRpb24gc2V0dGluZ3MuPEJSPiZndDs8QlI+Jmd0OzxCUj4mZ3Q7IEludGVy bmV0LURyYWZ0cyBhcmUgYWxzbyBhdmFpbGFibGUgYnkgYW5vbnltb3VzIEZUUC4gTG9naW4gd2l0 aCB0aGU8QlI+dXNlcm5hbWU8QlI+Jmd0OyAmcXVvdDthbm9ueW1vdXMmcXVvdDsgYW5kIGEgcGFz c3dvcmQgb2YgeW91ciBlLW1haWwgYWRkcmVzcy4gQWZ0ZXIgbG9nZ2luZyBpbiw8QlI+Jmd0OyB0 eXBlICZxdW90O2NkIGludGVybmV0LWRyYWZ0cyZxdW90OyBhbmQgdGhlbjxCUj4mZ3Q7ICZxdW90 O2dldCBkcmFmdC1pZXRmLWNjYW1wLWxzcC1zdGl0Y2hpbmctMDEudHh0JnF1b3Q7LjxCUj4mZ3Q7 PEJSPiZndDsgQSBsaXN0IG9mIEludGVybmV0LURyYWZ0cyBkaXJlY3RvcmllcyBjYW4gYmUgZm91 bmQgaW48QlI+Jmd0OyA8QSBIUkVGPWh0dHA6Ly93d3cuaWV0Zi5vcmcvc2hhZG93Lmh0bWw+aHR0 cDovL3d3dy5pZXRmLm9yZy9zaGFkb3cuaHRtbDwvQT48QlI+Jmd0OyBvciA8QSBIUkVGPWZ0cDov L2Z0cC5pZXRmLm9yZy9pZXRmLzFzaGFkb3ctc2l0ZXMudHh0PmZ0cDovL2Z0cC5pZXRmLm9yZy9p ZXRmLzFzaGFkb3ctc2l0ZXMudHh0PC9BPjxCUj4mZ3Q7PEJSPiZndDs8QlI+Jmd0OyBJbnRlcm5l dC1EcmFmdHMgY2FuIGFsc28gYmUgb2J0YWluZWQgYnkgZS1tYWlsLjxCUj4mZ3Q7PEJSPiZndDsg U2VuZCBhIG1lc3NhZ2UgdG86PEJSPiZndDsgbWFpbHNlcnZAaWV0Zi5vcmcuPEJSPiZndDsgSW4g dGhlIGJvZHkgdHlwZTo8QlI+Jmd0OyAmcXVvdDtGSUxFIC9pbnRlcm5ldC1kcmFmdHMvZHJhZnQt aWV0Zi1jY2FtcC1sc3Atc3RpdGNoaW5nLTAxLnR4dCZxdW90Oy48QlI+Jmd0OzxCUj4mZ3Q7IE5P VEU6IFRoZSBtYWlsIHNlcnZlciBhdCBpZXRmLm9yZyBjYW4gcmV0dXJuIHRoZSBkb2N1bWVudCBp bjxCUj4mZ3Q7IE1JTUUtZW5jb2RlZCBmb3JtIGJ5IHVzaW5nIHRoZSAmcXVvdDttcGFjayZxdW90 OyB1dGlsaXR5LiAmbmJzcDtUbyB1c2UgdGhpczxCUj4mZ3Q7IGZlYXR1cmUsIGluc2VydCB0aGUg Y29tbWFuZCAmcXVvdDtFTkNPRElORyBtaW1lJnF1b3Q7IGJlZm9yZSB0aGUgJnF1b3Q7RklMRSZx dW90OzxCUj4mZ3Q7IGNvbW1hbmQuICZuYnNwO1RvIGRlY29kZSB0aGUgcmVzcG9uc2UocyksIHlv dSB3aWxsIG5lZWQgJnF1b3Q7bXVucGFjayZxdW90OyBvcjxCUj4mZ3Q7IGEgTUlNRS1jb21wbGlh bnQgbWFpbCByZWFkZXIuICZuYnNwO0RpZmZlcmVudCBNSU1FLWNvbXBsaWFudCBtYWlsIHJlYWRl cnM8QlI+Jmd0OyBleGhpYml0IGRpZmZlcmVudCBiZWhhdmlvciwgZXNwZWNpYWxseSB3aGVuIGRl YWxpbmcgd2l0aDxCUj4mZ3Q7ICZxdW90O211bHRpcGFydCZxdW90OyBNSU1FIG1lc3NhZ2VzIChp LmUuIGRvY3VtZW50cyB3aGljaCBoYXZlIGJlZW4gc3BsaXQ8QlI+Jmd0OyB1cCBpbnRvIG11bHRp cGxlIG1lc3NhZ2VzKSwgc28gY2hlY2sgeW91ciBsb2NhbCBkb2N1bWVudGF0aW9uIG9uPEJSPiZn dDsgaG93IHRvIG1hbmlwdWxhdGUgdGhlc2UgbWVzc2FnZXMuPEJSPiZndDs8QlI+Jmd0OzxCUj4m Z3Q7IEJlbG93IGlzIHRoZSBkYXRhIHdoaWNoIHdpbGwgZW5hYmxlIGEgTUlNRSBjb21wbGlhbnQg bWFpbCByZWFkZXI8QlI+Jmd0OyBpbXBsZW1lbnRhdGlvbiB0byBhdXRvbWF0aWNhbGx5IHJldHJp ZXZlIHRoZSBBU0NJSSB2ZXJzaW9uIG9mIHRoZTxCUj4mZ3Q7IEludGVybmV0LURyYWZ0LjxCUj4m Z3Q7PEJSPjwvRk9OVD48L1A+ Envelope-to: ccamp-data@psg.com Delivery-date: Sun, 17 Jul 2005 17:21:28 +0000 Message-ID: <001301c58af3$8a644400$6501a8c0@movaz.com> From: "Igor Bryskin" <ibryskin@movaz.com> To: <i-d-announce@ietf.org>, <Internet-Drafts@ietf.org> Cc: <ccamp@ops.ietf.org> Subject: Re: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Date: Sun, 17 Jul 2005 13:18:24 -0400 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Arthi && JP. I have a couple of questions. 1. Why are saying that LSP Stitching is a private case of LSP Hierarchy? IMO there is more differences than similarities: The differences are: 1) In case of H-LSP there is a data plane adjacency, while in case of S-LSP there is none (as you correctly pointed out); 2) In case of H-LSP there is an adaptation in data plane (label push/pop for PSC), while in case of S-LSP there is none - just simple cross-connecting ( label swap) as in case of two "native " e2e LSP adjacent segments ; 3) H-LSP could be used by many e2e LSPs, while S-LSP could be used by exactly one e2e LSP 4) Signaling is different - there is no label negotiation in stitching 5) H-LSP is used as a "true" data link, specifically there is a resource allocation on the H-LSP edges, while in case of S-LSP there is none 6) From MLN point of view, H-LSP is created in a server (lower) layer, while the S-LSP is created in the client (same as e2e LSP) layer. There are two similarities that I can think of: 1) There is a signaling and possibly routing (see below) adjacencies between the ends; 2) Both H-LSP and S-LSP could be advertised as separate TE links or as TE bundles I would recommend to dedicate a paragraph and enlist there similarities and differencies 2. Why are you saying that a TE Link based on S-LSP can be used for exactly one e2e LSP? Parallel S-LSPs could be advertised as a single TE link (bundle, see above) and hence can accomadate several e2e LSPs. 3. You are saying that S-LSP does not have a routing peering. Actually, in this respect it is no different from H-LSP: if it is advertised as a TE link into the same TE domain that was used for S-LSP creation (unlikely IMO scenario) than it does not require the routing adjacency (in other words, it is an FA according to LSP-HIER definition), otherwise, it IS NOT and FA and does require the direct routing peering in the domain it is advertised to make it useful as a TE link in this domain Cheers, Igor ----- Original Message ----- From: <Internet-Drafts@ietf.org> To: <i-d-announce@ietf.org> Cc: <ccamp@ops.ietf.org> Sent: Friday, July 15, 2005 3:50 PM Subject: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt > A New Internet-Draft is available from the on-line Internet-Drafts directories. > This draft is a work item of the Common Control and Measurement Plane Working Group of the IETF. > > Title : Label Switched Path Stitching with Generalized MPLS Traffic Engineering > Author(s) : A. Ayyangar, J. Vasseur > Filename : draft-ietf-ccamp-lsp-stitching-01.txt > Pages : 19 > Date : 2005-7-15 > > In certain scenarios, there may be a need to combine together two > different Generalized Multi-Protocol Label Switching (GMPLS) Label > Switched Paths (LSPs) such that in the data plane, a single end-to- > end (e2e) LSP is achieved and all traffic from one LSP is switched > onto the other LSP. We will refer to this as "LSP stitching". This > document covers cases where: a) the node performing the stitching > does not require configuration of every LSP pair to be stitched > together b) the node performing the stitching is not the egress of > any of the LSPs c) LSP stitching not only results in an end-to-end > LSP in the data plane, but there is also a corresponding end-to-end > LSP (RSVP session) in the control plane. It might be possible to > configure a GMPLS node to switch the traffic from an LSP for which it > is the egress, to another LSP for which it is the ingress, without > requiring any signaling or routing extensions whatsoever, completely > transparent to other nodes. This will also result in LSP stitching > in the data plane. However, this document does not cover this > scenario of LSP stitching. > > A URL for this Internet-Draft is: > http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt > > To remove yourself from the I-D Announcement list, send a message to > i-d-announce-request@ietf.org with the word unsubscribe in the body of the message. > You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce > to change your subscription settings. > > > Internet-Drafts are also available by anonymous FTP. Login with the username > "anonymous" and a password of your e-mail address. After logging in, > type "cd internet-drafts" and then > "get draft-ietf-ccamp-lsp-stitching-01.txt". > > A list of Internet-Drafts directories can be found in > http://www.ietf.org/shadow.html > or ftp://ftp.ietf.org/ietf/1shadow-sites.txt > > > Internet-Drafts can also be obtained by e-mail. > > Send a message to: > mailserv@ietf.org. > In the body type: > "FILE /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt". > > NOTE: The mail server at ietf.org can return the document in > MIME-encoded form by using the "mpack" utility. To use this > feature, insert the command "ENCODING mime" before the "FILE" > command. To decode the response(s), you will need "munpack" or > a MIME-compliant mail reader. Different MIME-compliant mail readers > exhibit different behavior, especially when dealing with > "multipart" MIME messages (i.e. documents which have been split > up into multiple messages), so check your local documentation on > how to manipulate these messages. > > > Below is the data which will enable a MIME compliant mail reader > implementation to automatically retrieve the ASCII version of the > Internet-Draft. > Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 15 Jul 2005 19:51:24 +0000 Content-Type: Multipart/Mixed; Boundary="NextPart" Mime-Version: 1.0 To: i-d-announce@ietf.org Cc: ccamp@ops.ietf.org From: Internet-Drafts@ietf.org Subject: I-D ACTION:draft-ietf-ccamp-lsp-stitching-01.txt Message-Id: <E1DtWCB-0001vx-7N@newodin.ietf.org> Date: Fri, 15 Jul 2005 15:50:03 -0400 --NextPart A New Internet-Draft is available from the on-line Internet-Drafts directories. This draft is a work item of the Common Control and Measurement Plane Working Group of the IETF. Title : Label Switched Path Stitching with Generalized MPLS Traffic Engineering Author(s) : A. Ayyangar, J. Vasseur Filename : draft-ietf-ccamp-lsp-stitching-01.txt Pages : 19 Date : 2005-7-15 In certain scenarios, there may be a need to combine together two different Generalized Multi-Protocol Label Switching (GMPLS) Label Switched Paths (LSPs) such that in the data plane, a single end-to- end (e2e) LSP is achieved and all traffic from one LSP is switched onto the other LSP. We will refer to this as "LSP stitching". This document covers cases where: a) the node performing the stitching does not require configuration of every LSP pair to be stitched together b) the node performing the stitching is not the egress of any of the LSPs c) LSP stitching not only results in an end-to-end LSP in the data plane, but there is also a corresponding end-to-end LSP (RSVP session) in the control plane. It might be possible to configure a GMPLS node to switch the traffic from an LSP for which it is the egress, to another LSP for which it is the ingress, without requiring any signaling or routing extensions whatsoever, completely transparent to other nodes. This will also result in LSP stitching in the data plane. However, this document does not cover this scenario of LSP stitching. A URL for this Internet-Draft is: http://www.ietf.org/internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt To remove yourself from the I-D Announcement list, send a message to i-d-announce-request@ietf.org with the word unsubscribe in the body of the message. You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce to change your subscription settings. Internet-Drafts are also available by anonymous FTP. Login with the username "anonymous" and a password of your e-mail address. After logging in, type "cd internet-drafts" and then "get draft-ietf-ccamp-lsp-stitching-01.txt". A list of Internet-Drafts directories can be found in http://www.ietf.org/shadow.html or ftp://ftp.ietf.org/ietf/1shadow-sites.txt Internet-Drafts can also be obtained by e-mail. Send a message to: mailserv@ietf.org. In the body type: "FILE /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt". NOTE: The mail server at ietf.org can return the document in MIME-encoded form by using the "mpack" utility. To use this feature, insert the command "ENCODING mime" before the "FILE" command. To decode the response(s), you will need "munpack" or a MIME-compliant mail reader. Different MIME-compliant mail readers exhibit different behavior, especially when dealing with "multipart" MIME messages (i.e. documents which have been split up into multiple messages), so check your local documentation on how to manipulate these messages. Below is the data which will enable a MIME compliant mail reader implementation to automatically retrieve the ASCII version of the Internet-Draft. --NextPart Content-Type: Multipart/Alternative; Boundary="OtherAccess" --OtherAccess Content-Type: Message/External-body; access-type="mail-server"; server="mailserv@ietf.org" Content-Type: text/plain Content-ID: <2005-7-15145250.I-D@ietf.org> ENCODING mime FILE /internet-drafts/draft-ietf-ccamp-lsp-stitching-01.txt --OtherAccess Content-Type: Message/External-body; name="draft-ietf-ccamp-lsp-stitching-01.txt"; site="ftp.ietf.org"; access-type="anon-ftp"; directory="internet-drafts" Content-Type: text/plain Content-ID: <2005-7-15145250.I-D@ietf.org> --OtherAccess-- --NextPart-- Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 15 Jul 2005 16:22:15 +0000 content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: End-to-end L2-LSP Date: Fri, 15 Jul 2005 17:20:56 +0100 Message-ID: <0536FC9B908BEC4597EE721BE6A35389101B81FE@i2km07-ukbr.domain1.systemhost.net> Thread-Topic: End-to-end L2-LSP Thread-Index: AcWII4lsKzFN0JPAQs66qP37pnL8MABM3i8w From: <neil.2.harrison@bt.com> To: <jaihyung@etri.re.kr> Cc: <ccamp@ops.ietf.org> Hi Jaihyung, Having had some private requests to 'say something' on the list here I = want to respond to a point you raise below, viz: "Ever increasing best effort bandwidth will eventually not satisfy=20 both providers and customers for many reasons. Service providers=20 will have difficulty in collecting expenses enough to invest on = increased=20 bandwidth. This will in turn gradually downgrade overall communication = quality,=20 and applications requiring real-time service will never satisfy = customers.=20 We already see such symptom in some low-price service providers.=20 I believe new service structure that can provide difference of service=20 quality will offer benefit to customers, providers as well as vendors in = the end." See if these observations help: Real world networks (any mode) run at low utilisations over most of = their working life for a whole raft of important reasons. In this = 'normal' operating region the key challenge for the network architecture = is that it must enable operators the ability to create consistently = meaningful/priced services over all load points.....else longer-term = investment will suffer. If the network architecture cannot address this = then we have a serious problem brewing for not only this industry but = all those dependent on it....and currently this would appear to be the = case. Many operators are already discovering that 'ever faster and = faster' access without some fair usage pricing is not satisfactory. And = please don't assume this is an operator only problem (ie a mindset of = 'here are some capabilities now you figure out how to use them' is not = good enough).....if the operator community cannot generate a fair return = on investment then both vendors and customers will ultimately all = suffer. There is no free lunch here....but we do need a fairly priced = lunch. When networks fail and enter the 'abnormal' operating region the nature = of the key networking challenge changes significantly. Here the key = issue for the network architecture is targeted traffic survivability. = So how does one decide which traffic units are more important than other = ones?...esp if I make the observation that ANY application can assume = ANY 'importance' but yet all applications of a common kind have a fixed = QoS requirement. And this should not be just considered in the context = of public services, we also need to include private network builder = services. I am of the strong opinion that trying to solve the above = service-oriented problems by having lots of QoS classes in a single = networking mode/technology won't work very well (understatement). And = there is a substantial body of historical networking evidence to support = this view (and some interesting analysis, eg by Andrew Odlyzko). Go = have look at the truly successful networks that have prrofs by long-term = existence and the one thing you will note that these all have in common = is they do not have multiple QoS classes. ATM tried multiple QoS = classes and it was a failure. Only AAL5 and AAL1 have any major use = today. So why is this? Well its actually rather easy to explain: Applications have only 3 broad classifications: messages, files and = streams. It's therefore intuitively obvious one does need a large set = of QoS/traffic classes to address this. Indeed, in IP networking we = find UDP, TCP and RTP map very nicely to the key major traffic = classes.....this is NOT to say the serving network can effectively (ie = includes cost) provide for these however. AAL5 caters largely for the = message/file class and AAL1 the stream class. Everything else boils = down to BW squeezing.....but successful networks run at low loads over = most of their lives and BW is no longer such a scarce commodity that we = need to try and conserve it everywhere. Indeed BW should be used as a = tool to drive down complexity....and complexity costs are indeed = increasing (ask any operator about opex). In fact a '1 mode/technology = mindset' incurs lots of complexity. Now although the above considers the public networking case it does not = really address the needs of the private networking case. Here, the = basic service building block is a client/server layer network = relationship where the client and server layer networks must be = functionally decoupled....if for no other reason (and there are many = other reasons) than the fact that the client and server layer networks = may be owned by different operating parties. In such a relationship how = can anyone make any meaningful judgement as to a given application's = importance?.....vis-=E0-vis the same application (but different user = instance) in the same private network service instance, or the same = application in some other customer's private network service instance. = And note that once we nest several layer networks deep then sight of = ultimate end-user applications gets lost...as indeed they must, else the = management of those lower layer networks would become intractably = complex. There is also a performance inheritance issue here, since the = performance of layer network N is a combination of (i) impairments = introduce by the nodes at layer N and (ii) the impairments introduced in = the link-connections of layer N which come from trails at layer = N-1.....this performance inheritance issue recurses to the duct. So how are we to address the above issues of: - consistently meaningful/priced services in the normal operating = regions? - correctly targeted traffic survivability in the abnormal operating = region? - and across all public and private services?.....noting the important = client/server relationship between layered networks mentioned above I actually think there is a good solution to these problems....but it = requires we let go of any technology religion and consider all 3 = networking modes working in concert. One key point to note is that in = the co modes the trail object allows us to decouple the QoS semantic of = child traffic units carried (that can change epoch by epoch) from the = importance semantic of the parent trail entity (which is temporally = invariant). For the private services case the issue is very simple. One must = provide a co-cs or co-ps CBR-like server layer. Indeed, its no = coincidence that as one nears the duct all transport networks look like = this. They are the most efficient way of handling bulk traffic of any = type as any major operator will tell you. For the public service case let us assume a customer could select the = mode he/she wanted to use and pay for this accordingly. So for message = and some file applications the customer may be happy with a simple cl-ps = mode service (1 QoS class). For stream and some file applications the = customer may be happy to pay for a simple CBR-like co-ps SVC-based = service (again 1 QoS class). Just have think about this...and I mean = really think....of how this could address some otherwise very hard = problems in the QoS/pricing area. And just consider how simple both the = data-plane (classifiers/marking/shaping/policing/CAC/whatever) and = control-plane (routing and signalling messages sets) components become. You can think of this as the marriage of the best of the Internet and = PSTN worlds if you like. I do believe this scenario is worthy of = serious consideration and should not be dismissed lightly. regards, Neil > -----Original Message----- > From: owner-ccamp@ops.ietf.org=20 > [mailto:owner-ccamp@ops.ietf.org] On Behalf Of CHO, JAI HYUNG > Sent: 14 July 2005 04:24 > To: Spencer,R,Richard,XGH5 R; ccamp@ops.ietf.org > Subject: RE: End-to-end L2-LSP >=20 >=20 > =20 > =20 > Dear Richard, >=20 > See in-line please, >=20 > >Thank you for taking the time to respond to my questions. I=20 > anticipated > >that your response would include "over-provisioning is the=20 > best answer"=20 > >to decrease the chance of congestion on non-GMPLS switches.=20 > >This is what is done in the LAN environment today in=20 > conjunction with CoS,=20 > >and this is why I believe there will be two opposing views on how=20 > >far L2-LSP connections should be extended down to the user. One view=20 > >is that over provisioning and CoS work fine in the=20 > enterprise network today,=20 > >and therefore the L2-LSP should only be extended down to the point=20 > >at which the enterprise network connects to the SP network,=20 > >e.g. Ethernet DSLAM. I think the other view will be that=20 > over provisioning=20 > >and CoS in the enterprise network don't work and therefore L2-LSPs=20 > >should be provided end-to-end using only GMPLS capable switches.=20 > >I wouldn't have thought there would be much call for mixed=20 > >GMPLS/non-GMPLS enterprise networks (except for during migration)=20 > >due to the opposing views, e.g. if over provisioning is good=20 > enough for=20 > >one switch, why isn't it good enough for the other? >=20 > Over-provisioned network is non-guaranteed network. It can't prevent=20 > occasional congestion at any rate. Further, network condition=20 > in private=20 > network is very diverse. You can't predict minimum service level=20 > inside every private network.=20 >=20 > Suppose that an ISP want to provide high-quality, video-phone service=20 > for enterprise users. Nothing can be guaranteed inside the=20 > over-provisioned=20 > customer network. If the service level is not satisfactory,=20 > customers will complain=20 > to service provider even if the main cause is in his network=20 > or destination network.=20 > We need some structure that coordinate service quality in=20 > both private sector,=20 > as well as in public sector consistently.=20 > An end-to-end signal is an effective means to communicate between=20 > user terminal, private switches and provider switches, and=20 > concert service=20 > quality at each different operations network. This is a goal=20 > that RSVP (RFC2205)=20 > pursued, however ineffective in local network because=20 > Ethernet industry=20 > didn't support RSVP over Ethernet (RFC2814, 2815, 2816).=20 > Now we have another opportunity to realize the goal using L2-LSP.=20 > Therefore, I believe end-to-end LSP is necessary even in=20 > enterprise network=20 > for the purpose of providing guaranteed commercial service. >=20 > >I think where most people will agree is that using connections > >in the SPs access network does make sense in order to=20 > provide efficient=20 > >link utilisation and per connection policing to ensure one customers=20 > >traffic can not affect another's. Regarding the ability to=20 > set up new=20 > >connections based on a request from the user (rather than=20 > constraining=20 > >RSVP signalling to inside the SP network), I think this will come=20 > >down to enterprise customer demand for improved service performance.=20 > >If a customer wasn't getting the performance they needed from their=20 > >SP and was offered a new service that allowed them to setup=20 > new L2-LSP=20 > >connections on demand that would offer better performance=20 > (e.g. for voice=20 > >calls or video conferences), then upgrading the user devices=20 > to support=20 > >the new service could well be an attractive proposition. If=20 > on the other=20 >=20 > Yes, this will be one of prospective service newly enabled=20 > using L2-LSP. >=20 > >hand a customer buys a best effort service and considers the=20 > >performance > >to be 'good enough', then there would be no reason to go to=20 > the expense=20 > >of upgrading the user devices to support the set up of L2-LSPs. >=20 > Ever increasing besteffort bandwidth will eventually not satisfy=20 > both providers and customers for many reasons. Service providers=20 > will have difficulty in collecting expenses enough to invest=20 > on increased=20 > bandwidth. This will in turn gradually downgrade overall=20 > communication quality,=20 > and applications requiring real-time service will never=20 > satisfy customers.=20 > We already see such symptom in some low-price service providers.=20 > I believe new service structure that can provide difference=20 > of service=20 > quality will offer benefit to customers, providers as well as=20 > vendors in the end. >=20 > I also look forward to see you at Paris and continue helpful=20 > discussion. >=20 > Sincerely >=20 > Jaihyung >=20 >=20 > =20 >=20 > Dr. Jaihyung Cho > ETRI, Korea > phone : 042) 860-5514 > oversea: +82-42-860-5514 > fax: +82-42-861-5550=20 >=20 >=20 >=20 >=20 > -----?? ???-----=20 > ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com>=20 > ?? ??: 2005-07-14 ?? 1:10:30=20 > ?? ??: "jaihyung@etri.re.kr" <jaihyung@etri.re.kr>,=20 > "ccamp@ops.ietf.org" <ccamp@ops.ietf.org>=20 > ??:=20 > ??: RE: End-to-end L2-LSP=20 >=20 >=20 >=20 > Jaihyung,=20 >=20 > Thank you for taking the time to respond to my questions. I=20 > anticipated that your response would include=20 > "over-provisioning is the best answer" to decrease the chance=20 > of congestion on non-GMPLS switches. This is what is done in=20 > the LAN environment today in conjunction with CoS, and this=20 > is why I believe there will be two opposing views on how far=20 > L2-LSP connections should be extended down to the user. One=20 > view is that over provisioning and CoS work fine in the=20 > enterprise network today, and therefore the L2-LSP should=20 > only be extended down to the point at which the enterprise=20 > network connects to the SP network, e.g. Ethernet DSLAM. I=20 > think the other view will be that over provisioning and CoS=20 > in the enterprise network don't work and therefore L2-LSPs=20 > should be provided end-to-end using only GMPLS capable=20 > switches. I wouldn't have thought there would be much call=20 > for mixed GMPLS/non-GMPLS enterprise networks (except for=20 > during migration) due to the opposing views, e.g. if over=20 > provisioning is good enough for one switch, why isn't it good=20 > enough for the other?=20 >=20 > I think where most people will agree is that using=20 > connections in the SPs access network does make sense in=20 > order to provide efficient link utilisation and per=20 > connection policing to ensure one customers traffic can not=20 > affect another's. Regarding the ability to set up new=20 > connections based on a request from the user (rather than=20 > constraining RSVP signalling to inside the SP network), I=20 > think this will come down to enterprise customer demand for=20 > improved service performance. If a customer wasn't getting=20 > the performance they needed from their SP and was offered a=20 > new service that allowed them to setup new L2-LSP connections=20 > on demand that would offer better performance (e.g. for voice=20 > calls or video conferences), then upgrading the user devices=20 > to support the new service could well be an attractive=20 > proposition. If on the other hand a customer buys a best=20 > effort service and considers the performance to be 'good=20 > enough', then there would be no reason to go to the expense=20 > of upgrading the user devices to support the set up of L2-LSPs.=20 >=20 > I look forward to the discussions in Paris.=20 >=20 > Regards,=20 > Richard=20 >=20 > > -----Original Message----- > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr]=20 > > Sent: 13 July 2005 15:14=20 > > To: Spencer,R,Richard,XDE73 R; ccamp@ops.ietf.org=20 > > Subject: RE: End-to-end L2-LSP=20 > >=20 > >=20 > > =20 > > =20 > > Dear Richard Spencer > > =20 > > see in-line > > =20 > > >Jaihyung, > > >=20 > > >Perhaps I'm becoming too much of a pessimist;-) > > =20 > > oh, not at all, > > you are reasonable enough to dig up unclear knowledge=20 > > please go ahead until you finally get satisfactory answer.=20 > > =20 > >=20 > > >Ignoring the commercial aspects, please can you help clarify > > a few technical > > >questions I have in order to help me gain a better > > understanding of the proposal? > > >=20 > > >1. You say that one of the key reasons for using GMPLS in > > the enterprise > > >network is to provide end-to-end QoS control. However, you > > also say that not > > >all the switches will need to support GMPLS, only those at > > bottle necks. > > >Please can you expand on this to explain how you can have=20 > end-to-end > > >QoS control when the L2-LSP is not end-to-end?=20 > > =20 > >=20 > > The contextual meaning of 'End-to-end LSP' is that GMPLS label is > > assigned from source terminal and removed at destination terminal,=20 > > because RSVP-TE message is exchanged from host to host.=20 > > It may not perfectly guarantee QoS when the GMPLS deployment=20 > > is not complete, however it will at least be good enough=20 > > for end-user applications to tell those GMPLS implemented=20 > L2 switches=20 > > what frame must be given priority service when congestion occur.=20 > > =20 > >=20 > > >As being discussed > > >on the PWE3 mailing list, the performance a service receives=20 > > is only as > > >good as the worst performing server layer. Therefore, if there is a > > >standard 802.1 workgroup switch between the user and the=20 > > L2-LSP that is > > >experiencing congestion (e.g. due to a large amount of > > peer-to-peer traffic), > > >then the QoS that the user receives will only be as good as > > that provided > > >by the congested workgroup switch. > > =20 > >=20 > > of course, ETH-GMPLS does not prevent congestion inevitably=20 > occurring > > in those non-GMPLS deployed part.=20 > > =20 > >=20 > > >The only way I can see a network using a mixture of GMPLS > > and non-GMPLS > > >switches guaranteeing end-to-end QoS would be to design the > > network in > > >such a way that the non-GMPLS switches could never become=20 > congested. > > =20 > >=20 > > Yes, that will be the way wise network designer improve=20 > their network. > > =20 > >=20 > > >To do this would require that RSVP requests be used for ALL > > traffic flows > > >to ensure that there was enough bandwidth available, along > > with strict > > >per L2-LSP policing on the user devices (e.g. PC). > > >However, the obvious flaw in this approach is that someone could=20 > > >just come along and connect=20 > > >an end device to a non-GMPLS switch and start transmitting traffic=20 > > >(without performing CAC/policing) causing the switch to=20 > > become congested. > > >Therefore, please can you explain how end-to-end QoS can be > > guaranteed > > >in a mixed non-GMPLS and GMPLS switching environment? > > =20 > >=20 > > Of course traffic came from non controlled ingress can not be > > prevented=20 > > in non-GMPLS node. There's no magic. Only the=20 > > over-provisioning is the=20 > > best answer. However, once the aggregated traffic pass 'ANY' one of=20 > > GMPLS enabled node, those uncontrolled traffic will suffer=20 > > queuing discipline.=20 > > =20 > >=20 > > >2. Section 5.1 currently states "When the customer initiates > > data transmission, > > >the access switch maps the user flow into the L2 LSP. > > Mapping procedure is > > >subject to implementation choices, and is out of the scope of this > > >document." In order to be able to map any traffic, first of=20 > > all the switch needs > > >to know what type of traffic (e.g. based on source/dest MAC, > > VLAN ID, 802.1p > > >bits, etc) should be mapped to the L2-LSP. In Figure 1, if > > an L2-LSP is set up > > >from the DSLAM to the edge router based on a request from > > the user (e.g. via > > >a PC or VoIP phone), then won't the user also need to signal > > what type of > > >traffic should to be mapped to the L2-LSP? > > =20 > >=20 > > Thank you for pointing out important thing. > > That's the most unfortunate part I reluctantly agreed because the=20 > > work condition in the framework DT was not to span L2-LSP=20 > > beyond provider network.=20 > > As you noticed, mapping user traffic at provider access is awkward.=20 > > A smooth method will be for user terminal aware of the L2 label=20 > > to use by receiving RSVP-TE message.=20 > > However, there was strong argument that RSVP-TE must not=20 > > go beyond provider network because the DT is not chartered=20 > > to work in customer network.=20 > > That's why I requested amendment of CCAMP charter=20 > > to include user terminal explicitly in CCAMP work scope, so that=20 > > avoid such silly argument in future.=20 > > =20 > >=20 > > >3. In section 5.1, the text states that the aggregation > > switch between the > > >DSLAM and edge router may be a 802.1 switch, i.e. it doesn't > > have to support > > >GMPLS. To set up a connection, first of all there needs to > > be a data plane > > >capable of carrying control plane traffic between GMPLS > > peers. Secondly, > > >the GMPLS peers need reachability information in order to > > forward control > > >messages onto the next peer (i.e. a static or IGP route). If > > the GMPLS peers > > >(e.g. the DSLAM and edge router in Figure 1) are not > > directly connected, > > >then the GMPLS control packets will need to be forwarded by > > the 802.1 switch. > > >Would the idea be to have say a "control traffic VLAN" > > configured between the > > >DSLAM and edge router via the transit 802.1 switch so that > > control packets > > >could be forwarded transparently by the 802.1 switch? > > =20 > >=20 > > Although it doesn't preclude use of VLAN, I think such > > special configuration=20 > > for RSVP-TE control message is not necessary in legacy switches.=20 > > Legacy switches only need to pass RSVP-TE messages as they=20 > > pass normal data frames. Only the GMPLS implemented switches=20 > > need to distinguish frames containing RSVP-TE messages and=20 > > snoop them.=20 > > Detail of implementation is, of course, for future discussion.=20 > > =20 > >=20 > > >4. If we assume that control messages can be transparently=20 > forwarded > > >by the intermediate 802.1 aggregation switch then how are=20 > > user frames > > >forwarded in the data plane by the 802.1 switch? > > =20 > >=20 > > Legacy switches do not distinguish control/data frames. > > =20 > >=20 > > >If the DSLAM sets up a L2-LSP > > >to the edge router using say VLAN ID 99, how does the switch=20 > > know that any > > >packets coming in with a VLAN ID of 99 should be forwarded > > out of the port > > >connecting to the edge router? > > =20 > >=20 > > oh, you misunderstood the L2-LSP. > > The question you are asking depends on how L2 label is encoded.=20 > > there was hot debate in the DT about the issue how L2 label=20 > > is encoded.=20 > > My proposal is to use part of MAC address for label and the=20 > > other people=20 > > argued to use VLAN ID for label. The debate was not concluded=20 > > because the goal of the L2SC framework document is not to select=20 > > implementation method but to draw necessary work scope.=20 > > I believe this issue will be dealt in depth after Paris meeting.=20 > > So, please reserve your question until then and ask to those=20 > > VLAN proponents.=20 > > =20 > >=20 > > >5. During connection setup, I take it that the GMPLS peers > > assume that > > >resources are available within the intermediate 802.1 switch > > to support > > >the connection (as its transparent). However, as there is no > > >CAC/policing being performed by the 802.1 switch=20 > > >(the L2-LSPs are transparent), and due to the any-to-any and=20 > > broadcast > > >based nature of Ethernet (i.e. its CL properties), then=20 > how can this > > >assumption be made? This comes back to my original question, how=20 > > >can end-to-end QoS be guaranteed in an environment where there is=20 > > >a mixture of 802.1 and GMPLS enabled switches?=20 > > =20 > >=20 > > yes, you came back to the question I answered in above ;-) > > so, please see above.=20 > > GMPLS enabled nodes must be positioned at points congestion=20 > > is likely to be expected.=20 > > =20 > > Thanks > > =20 > > Jaihyung > > =20 > >=20 > > = =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D > >=20 > > =20 > >=20 > > -----?? ???----- > > ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com>=20 > > ?? ??: 2005-07-13 ?? 8:53:13=20 > > ?? ??: "jaihyung@etri.re.kr" <jaihyung@etri.re.kr>,=20 > > "ccamp@ops.ietf.org" <ccamp@ops.ietf.org>=20 > > ??:=20 > > ??: RE: End-to-end L2-LSP=20 > >=20 > >=20 > >=20 > >=20 > > Jaihyung, > >=20 > > Perhaps I'm becoming too much of a pessimist;-) Ignoring the > > commercial aspects, please can you help clarify a few=20 > > technical questions I have in order to help me gain a better=20 > > understanding of the proposal?=20 > >=20 > > 1. You say that one of the key reasons for using GMPLS in the > > enterprise network is to provide end-to-end QoS control.=20 > > However, you also say that not all the switches will need to=20 > > support GMPLS, only those at bottle necks. Please can you=20 > > expand on this to explain how you can have end-to-end QoS=20 > > control when the L2-LSP is not end-to-end? As being discussed=20 > > on the PWE3 mailing list, the performance a service receives=20 > > is only as good as the worst performing server layer.=20 > > Therefore, if there is a standard 802.1 workgroup switch=20 > > between the user and the L2-LSP that is experiencing=20 > > congestion (e.g. due to a large amount of peer-to-peer=20 > > traffic), then the QoS that the user receives will only be as=20 > > good as that provided by the congested workgroup switch.=20 > >=20 > > The only way I can see a network using a mixture of GMPLS and > > non-GMPLS switches guaranteeing end-to-end QoS would be to=20 > > design the network in such a way that the non-GMPLS switches=20 > > could never become congested. To do this would require that=20 > > RSVP requests be used for ALL traffic flows to ensure that=20 > > there was enough bandwidth available, along with strict per=20 > > L2-LSP policing on the user devices (e.g. PC). However, the=20 > > obvious flaw in this approach is that someone could just come=20 > > along and connect an end device to a non-GMPLS switch and=20 > > start transmitting traffic (without performing CAC/policing)=20 > > causing the switch to become congested. Therefore, please can=20 > > you explain how end-to-end QoS can be guaranteed in a mixed=20 > > non-GMPLS and GMPLS switching environment?=20 > >=20 > > 2. Section 5.1 currently states "When the customer initiates > > data transmission, the access switch maps the user flow into=20 > > the L2 LSP. Mapping procedure is subject to implementation=20 > > choices, and is out of the scope of this document." In order=20 > > to be able to map any traffic, first of all the switch needs=20 > > to know what type of traffic (e.g. based on source/dest MAC,=20 > > VLAN ID, 802.1p bits, etc) should be mapped to the L2-LSP. In=20 > > Figure 1, if an L2-LSP is set up from the DSLAM to the edge=20 > > router based on a request from the user (e.g. via a PC or=20 > > VoIP phone), then won't the user also need to signal what=20 > > type of traffic should to be mapped to the L2-LSP?=20 > >=20 > > 3. In section 5.1, the text states that the aggregation > > switch between the DSLAM and edge router may be a 802.1=20 > > switch, i.e. it doesn't have to support GMPLS. To set up a=20 > > connection, first of all there needs to be a data plane=20 > > capable of carrying control plane traffic between GMPLS=20 > > peers. Secondly, the GMPLS peers need reachability=20 > > information in order to forward control messages onto the=20 > > next peer (i.e. a static or IGP route). If the GMPLS peers=20 > > (e.g. the DSLAM and edge router in Figure 1) are not directly=20 > > connected, then the GMPLS control packets will need to be=20 > > forwarded by the 802.1 switch. Would the idea be to have say=20 > > a "control traffic VLAN" configured between the DSLAM and=20 > > edge router via the transit 802.1 switch so that control=20 > > packets could be forwarded transparently by the 802.1 switch?=20 > >=20 > > 4. If we assume that control messages can be transparently > > forwarded by the intermediate 802.1 aggregation switch then=20 > > how are user frames forwarded in the data plane by the 802.1=20 > > switch? If the DSLAM sets up a L2-LSP to the edge router=20 > > using say VLAN ID 99, how does the switch know that any=20 > > packets coming in with a VLAN ID of 99 should be forwarded=20 > > out of the port connecting to the edge router?=20 > >=20 > > 5. During connection setup, I take it that the GMPLS peers > > assume that resources are available within the intermediate=20 > > 802.1 switch to support the connection (as its transparent).=20 > > However, as there is no CAC/policing being performed by the=20 > > 802.1 switch (the L2-LSPs are transparent), and due to the=20 > > any-to-any and broadcast based nature of Ethernet (i.e. its=20 > > CL properties), then how can this assumption be made? This=20 > > comes back to my original question, how can end-to-end QoS be=20 > > guaranteed in an environment where there is a mixture of=20 > > 802.1 and GMPLS enabled switches?=20 > >=20 > > Thanks, > > Richard=20 > >=20 > > > -----Original Message----- > > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr]=20 > > > Sent: 12 July 2005 09:21=20 > > > To: Spencer,R,Richard,XDE73 R; ccamp@ops.ietf.org=20 > > > Subject: RE: End-to-end L2-LSP=20 > > >=20 > > >=20 > > >=20 > > >=20 > > > Dear Richard, > > >=20 > > > Yes, I believe it is not only commercially profitable but it > > > may also grow total=20 > > > revenue of service operators by introducing new services.=20 > > >=20 > > > For example, think the example of IP phone you referred. > > > How do you think you can provide reliable VoIP service and=20 > > > safely authenticate phone users and charge fees as counted by=20 > > > packet usage=20 > > > using simple Ethernet based xDSL network ?=20 > > >=20 > > > There's hardly new service that providers may offer using > > > Ethernet based=20 > > > architecture, and collect sufficient profit to investment.=20 > > >=20 > > > A sophisticated ATM switch might do the work if it is not > > expensive. > > > However, my intension is modifiying existing design of > > > Ethernet hardware=20 > > > with a little bit additional cost, and provide the level of=20 > > > service that=20 > > > ATM switches might offer.=20 > > >=20 > > > We have initial design of the Ethernet-MPLS hardware. I=20 > don't think > > > GMPLS enabled Ethernet switch will be so complex. The=20 > market price=20 > > > of GMPLS enabled switch will not be so expensive than normal=20 > > > Ethernet switches.=20 > > >=20 > > > Further, it only requires upgrade of switches in some > > > bottleneck points of network.=20 > > > Read the framework documents of L2SC. It doesn't require=20 > > > entire replacement=20 > > > of switches to introduce new service.=20 > > >=20 > > > Every year, ISPs purchase new access switches and replaces > > > old machines.=20 > > > With a bit additional cost, ISPs may use the GMPLS enabled=20 > > switch for > > > normal Ethernet switching in initial deployment. > > >=20 > > > When the GMPLS function is enabled in some upgraded part of > > network, > > > ISPs may provide high quality, reliable VoIP service to > > xDSL customers > > > using RSVP-TE enabled VoIP phone. > > > The high quality video-phone service enabled by GMPLS will be=20 > > > competable to legacy voice only phone service.=20 > > >=20 > > >=20 > > > There are lots of things I can argue about commercial aspects > > > of this technique.=20 > > > However I don't think it is not appropriate to discuss this=20 > > > issue in this mailing list.=20 > > > All I can say is,=20 > > > yes, I believe that it may be commercially viable.=20 > > >=20 > > >=20 > > > Thanks > > >=20 > > > Jaihyung > > >=20 > > > = =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D > > >=20 > > > Do you really think the benefits of extended GMPLS down to > > > the end user could ever justify the costs involved in=20 > > > upgrading LAN switches and things like IP phones (simple low=20 > > > cost commodity items) to support GMPLS signalling/routing?=20 > > >=20 > > > Your proposal may be technically possible (though it would > > > require new hardware and significant software upgrades), but=20 > > > commercially I would consider this to be a non-starter.=20 > > >=20 > > > Regards, > > > Richard=20 > > >=20 > > > > -----Original Message----- > > > > From: owner-ccamp@ops.ietf.org=20 > > [mailto:owner-ccamp@ops.ietf.org]On > > > > Behalf Of CHO, JAI HYUNG > > > > Sent: 12 July 2005 01:10=20 > > > > To: ccamp@ops.ietf.org=20 > > > > Subject: End-to-end L2-LSP=20 > > > >=20 > > > >=20 > > > > =20 > > > > Dear all, > > > >=20 > > > > Greetings. > > > >=20 > > > > I am a co-editor of the Ethernet-GMPLS framework document. > > > > (draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt)=20 > > > >=20 > > > > I am writing this to explain the motivation behind the > > > > proposal of scenario-1, and to raise the issue of=20 > > > > end-to-end LSP to be included in CCAMP charter.=20 > > > >=20 > > > > An important goal intended in scenario-1 is > > > > QoS control over end-to-end packet delivery using LSP.=20 > > > > In other words, the purpose of scenario-1 is extending coverage=20 > > > > of GMPLS control including customer network and=20 > end-user device,=20 > > > > such as PC, VoIP phone, IP-TV, or even future 4G mobile phone.=20 > > > > End-user applications will have a means to request resources=20 > > > > to network using GMPLS signal.=20 > > > >=20 > > > > Unfortunately, this view is not included in the > > > > framework because there was strong opinion that=20 > > > > L2-LSP must not span beyond provider network because=20 > > > > CCAMP is not chartered to work on customer=20 > > > > network. I am wondering if this is right argument.=20 > > > > Perhaps the charter can be amended to include this work.=20 > > > >=20 > > > > I believe there is little technical reason to preclude > > > > LSP being used in customer network. It is a reasonable=20 > > > > assumption that private company may also use GMPLS=20 > > > > implemented Ethernet switch, once the L2SC work is successful.=20 > > > > Currently, CCAMP is the only place discussing the matter of=20 > > > > Ethernet and LSP. There is no other WG dealing similar issue.=20 > > > > If CCAMP wants to work on Ethernet,=20 > > > > I think the coverage of common GMPLS control must be=20 > > > > extended to include all area where Ethernet is deployed.=20 > > > > Currently, Ethernet is dominantly deployed in first-mile=20 > > > > network where access network as well as customer network=20 > > > > and user hosts are usually included. Therefore, user host must=20 > > > > be considered as initiation and termination point of L2-LSP.=20 > > > >=20 > > > > Another important reason to include user host in CCAMP > > > > work scope is to achieve consistent end-to-end control of QoS.=20 > > > > End-to-end QoS has long been desired goal since RSVP was=20 > > > > first designed. This goal can be best achieved using=20 > > > > end-to-end GMPLS signal and L2-LSP.=20 > > > > Currently, RSVP (RFC2205) is available in many user hosts.=20 > > > > There is only a little difference between RSVP and RSVP-TE.=20 > > > > If the RSVP-TE signal of provider network can reach user host=20 > > > > across Ethernet based private network, an LSP can be=20 > established=20 > > > > from application to application via the provider backbone.=20 > > > > Service providers will be able to control application flows=20 > > > > in L2-LSP level as well as in aggregated LSP pipe of similar=20 > > > > application type. A variety of new services will be=20 > enabled using=20 > > > > the enhanced capability of distinguishing and controlling each=20 > > > > individual application flow.=20 > > > >=20 > > > > I personally believe that end-to-end L2-LSP may offer > > > > innovative solutions for servicing IP-TV as well as fast=20 > > mobility. > > > > There have been some experimental approaches combining > > > > MPLS and IP mobility. However such potential of the new=20 > > > > application can only be explored when end-user host is=20 > > > > included in the scope of CCAMP work area.=20 > > > >=20 > > > > Therefore, I hereby sincerely request people's attention > > > > and support on this perspective of end-to-end LSP.=20 > > > > If there is enough number of people agree on this view,=20 > > > > the WG can request the necessary update of the charter.=20 > > > >=20 > > > > Thank you. > > > > =20 > > > > Jaihyung Cho > > > > =20 > > > > =20 > > > > =20 > > > > ETRI, Korea=20 > > > > phone : 042) 860-5514=20 > > > > oversea: +82-42-860-5514=20 > > > > fax: +82-42-861-5550=20 > > > >=20 > > > >=20 > > > >=20 > > > >=20 > > > >=20 > > > >=20 > > > >=20 > > >=20 > > >=20 > > >=20 > > >=20 > >=20 > >=20 > >=20 > >=20 >=20 >=20 >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 15 Jul 2005 14:00:47 +0000 MIME-Version: 1.0 To: ccamp@ops.ietf.org Cc: adrian@olddog.co.uk, kireeti@juniper.net From: Dimitri.Papadimitriou@alcatel.be Subject: Re: Draft agenda for Paris Date: Fri, 15 Jul 2005 15:58:41 +0200 Message-ID: <OF2FD03775.F7E74021-ONC125703F.004CC852-C125703F.004CC8AF@netfr.alcatel.fr> Content-type: text/plain; charset=us-ascii hi adrian, kireeti, thanks for posting this agenda, this said, two important issues are imho creating increasing concern in this WG: 1. CCAMP WG backlog -> if one considers the current size of the WG I-Ds still under discussion, it has reached more than a reasonable size now since roughly 2/3 (25 over 36, if my count is correct) of the listed WG I-D documents are pending in RFC Editor queue, under reviews, etc. - the important issue is that one of the argument commonly used is "GMPLS beside signaling is still at the draft level" e.g. GMPLS UNI and other routing docs (meaning not stable, even if this not the case) but this is a REAL argument playing against GMPLS development and deployment => so, if an in-depth effort is not triggered this situation is going to persist for a long time, i would suggest we put this real issue over the next CCAMP meeting boiler plate 2. CCAMP WG recharter -> we are discussing this "re-charter" since quite some time now and there are no stated milestones beside closing items we have initiated since then (or introduce here and there some informational document trying to clarify achievements). As illustrative examples, pending items such as I/W and MRN (and derivates), are impacted by this delayed process: -> for I/W: there is a need for a DT-like approach since there are several I-Ds popping up and there is no real common WG vision, a 3-step approach (10 page framework/problem statement - solution/ mechanism - and then applicability) would be more than advisabe here as there is a need to clean up the orientations and come with a real focused stepwise process; with the current situation we are missing the increasing demand with confusing vision on migration mechanisms and this is also playing against GMPLS deployment -> for MRN and derivates: the real problem is the scope, as long as the scope is left open - and this is a WG responsibility - people have the tendancy to consider it as a moving target instead of stabilizing it and so at the end delaying instead of achieving stepwise progress that the industry requires the major resulting issue here (if this can not be addressed rapidly) will soon become "if the CCAMP WG is not willing to open its charter and address real community concerns what is the longer term role of this WG ?" and i am afraid to say here that this is going to leave us with two major options either split effort in smaller scoped WGs or do like with SIP and split into a base WG and an investigation WG. so, in brief, i would like to ask you and Kireeti, to seriously consider these issues and let us know how this inconfortable situation can be tackled and what are the actions to undertake in order to make it evolving rapidly thanks, - dimitri. Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 15 Jul 2005 09:19:19 +0000 To: ccamp@ops.ietf.org MIME-Version: 1.0 From: Dimitri.Papadimitriou@alcatel.be Subject: http://www.ietf.org/internet-drafts/draft-otani-ccamp-gmpls-cspf-constraints-00 Date: Fri, 15 Jul 2005 11:15:42 +0200 Message-ID: <OFA2541A06.F719EEC0-ONC125703F.0032DFAE-C125703F.0032E08E@netfr.alcatel.fr> MIME-Version: 1.0 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: base64 PFA+PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPmF1dGhvcnMsIGlzIHRoZSBwdXJwb3Nl IG9mIHRoZSBhYm92ZSBtZW50aW9uZWQgZHJhZnQ6PC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJN b25vc3BhY2UsQ291cmllciI+Jm5ic3A7MS4gcHJlc2NyaXB0aXZlIChldmVuIGlmIGluZm9ybWF0 aW9uYWwpPC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+Jm5ic3A7 Mi4gZGVzY3JpcHRpdmUgZm9yIGd1aWRpbmcgb3BlcmF0aW9ucyAoQkNQKTwvRk9OVD48L1A+PFA+ PEZPTlQgRkFDRT0iTW9ub3NwYWNlLENvdXJpZXIiPiZuYnNwOzMuIGluZm9ybWF0aXZlIChzdXJ2 ZXktbGlrZSk8QlI+PC9GT05UPjwvUD48UD48Rk9OVCBGQUNFPSJNb25vc3BhY2UsQ291cmllciI+ ZnJvbSB3aGF0IGkgcmVhZCwgaXQgaXMgbW9yZSAxIGxpa2UgLSBpcyBteSB1bmRlcnN0YW5kaW5n IGNvcnJlY3QgPzwvRk9OVD48L1A+ Envelope-to: ccamp-data@psg.com Delivery-date: Fri, 15 Jul 2005 00:52:51 +0000 Mime-Version: 1.0 (Apple Message framework v733) Content-Transfer-Encoding: 7bit Message-Id: <1F2E17DC-224E-4F09-A634-9595BDBE1175@cisco.com> Content-Type: text/plain; charset=US-ASCII; format=flowed To: ccamp@ops.ietf.org From: "Thomas D. Nadeau" <tnadeau@cisco.com> Subject: no more milestones? Date: Thu, 14 Jul 2005 17:50:34 -0700 Why the WG is stuck without any milestones and unable to take on more work? Are there any plans to extend the charter? --Tom Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 14 Jul 2005 23:52:38 +0000 Message-ID: <42D6FAB3.9050804@kddilabs.jp> Date: Fri, 15 Jul 2005 08:52:19 +0900 From: Tomohiro Otani <otani@kddilabs.jp> User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; ja-JP; rv:1.7.2) Gecko/20040803 MIME-Version: 1.0 To: ccamp@ops.ietf.org Subject: GMPLS OSPF-TE MIB Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi everyone, We posted a new draft about GMPLS OSPF-TE MIB as follows. http://www.ietf.org/internet-drafts/draft-otani-ccamp-gmpls-ospf-mib-00.txt We welcome your comments and feedbacks, pls. And please make a comment whether this is within ccamp WG or not (OSPF-WG?). We believe that this is a very significant definition of OSPF-TE MIB, being complementary to other GMPLS related MIBs (TE/LSR/LMP mibs) in order to deploy GMPLS control plane in the actual environment with simple but reliable network management. With best regards, tomo ------------------------------------ Tomohiro Otani KDDI R&D Laboratories, Inc. Optical network architecture lab. 2-1-15 Ohara Kamifukuoka Saitama, 356-8502, Japan TEL: +81-49-278-7357 FAX: +81-49-278-7811 E-mail: otani@kddilabs.jp ------------------------------------ Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 14 Jul 2005 18:52:13 +0000 Message-ID: <067701c588a5$45c7f1c0$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <ccamp@ops.ietf.org> Subject: Draft agenda for Paris Date: Thu, 14 Jul 2005 19:53:01 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi, We have come up with the following draft agenda. Before you all shout that your I-D is not on the list, as yourselves - Has the I-D been discussed much on the mailing list? - What would I like to see cut to make space for my I-D? - What benefit will the WG gain from discussing the I-D in the meeting. Recall, the purpose of an agenda slot is to allow the floor to discuss your I-D. It is not for you to present the contents or your I-D because (obviously) your I-D is so well written that everyone already understands the contents. A slide template and hints can be found at http://www.olddog.co.uk/ccamp.htm Please email me your final set of slides by close of business Monday 1st August. Thanks, Adrian === CCAMP Agenda - IETF-63 Paris Administrivia [10 minutes] (10) ITU-T Liaison report [10 minutes] (20) Status of WG [5 minutes] (25) - Remaining milestones - Drafts with RFC Editor - Drafts with AD - Drafts still in working group WG drafts Status and Plans (70) - Inter-domain - Stitching (Arthi) [5 minutes] draft-ietf-ccamp-lsp-stitching-01.txt - Inter-domain Signaling (Arthi) [5 minutes] draft-ietf-ccamp-inter-domain-rsvp-te-00.txt - PD-Path (JP) [5 minutes] draft-ietf-ccamp-inter-domain-pd-path-comp-00.txt - GMPLS Addressing (Richard) [10 minutes] draft-ietf-ccamp-gmpls-addressing-01.txt - ASON/GMPLS Lexicography (Igor) [10 minutes] draft-ietf-ccamp-gmpls-ason-lexicography-03.txt - ASON routing evaluation (Dimitri) [5 minutes] draft-ietf-ccamp-gmpls-ason-routing-eval-00.txt - ASON signaling (Dimitri) [5 minutes] draft-ietf-ccamp-gmpls-rsvp-te-ason-03.txt Working Group Work Plan (105) - Options for progress (chairs) [5 minutes] - Potential new work (chairs) [15 minutes] - Charter discussion (chairs and ADs) [15 minutes] Layer Two Switching (125) - Framework (Dimitri) [10 minutes] draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt - Next steps (chairs) [10 minutes] Other drafts (150) - Advertising node TE capabilities (JP) [5 minutes] draft-vasseur-ccamp-te-node-cap draft-vasseur-ccamp-automesh - Advanced constraints draft-otani-ccamp-gmpls-cspf-constraints (Tomohiro) [5 minutes] draft-ashwood-ccamp-gmpls-constraint-reqts-00.txt (Don) [5 mins] - VCAT/LCAS draft-bernstein-ccamp-gmpls-vcat-lcas-00.txt (Greg) [5 mins] draft-imajuku-ccamp-gmpls-vcat-lagr-req-00.txt (Wataru) [5 mins] Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 14 Jul 2005 08:46:44 +0000 Date: Thu, 14 Jul 2005 09:43:37 +0000 To: "Ccamp" <ccamp@ops.ietf.org> From: "Kireeti" <kireeti@juniper.net> Subject: Re: Message-ID: <sopfyfouabjzmphzerl@ops.ietf.org> MIME-Version: 1.0 Content-Type: multipart/mixed; boundary="--------vmfswhkcqaueuwndugfu" ----------vmfswhkcqaueuwndugfu Content-Type: text/plain; charset="UTF-8"; format="flowed" +----------------------------------------------------+ Panda GateDefender has detected malicious content (Virus) in the following file: [Dog.cpl] W32/Bagle.AH.worm The file has been deleted to protect the network. 07/14/2005 08:37 +0100 www.pandasoftware.com +----------------------------------------------------+ ----------vmfswhkcqaueuwndugfu Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: 7bit <html><body> >Lovely animals<br><br> <br> </body></html> ----------vmfswhkcqaueuwndugfu-- Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 14 Jul 2005 03:25:08 +0000 Thread-Topic: End-to-end L2-LSP thread-index: AcWII4lsKzFN0JPAQs66qP37pnL8MA== Reply-To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> From: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> To: <richard.spencer@bt.com>, <ccamp@ops.ietf.org> Cc: Bcc: Subject: RE: End-to-end L2-LSP Date: Thu, 14 Jul 2005 12:24:26 +0900 Comment: GQ19@|@ZEk=E?,18?x, BcN1b<z:P<.F@, 4c4g Message-ID: <6b78301c58823$896e5770$8310fe81@email1> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Class: urn:content-classes:message Dear Richard, See in-line please, >Thank you for taking the time to respond to my questions. I anticipated >that your response would include "over-provisioning is the best answer" >to decrease the chance of congestion on non-GMPLS switches. >This is what is done in the LAN environment today in conjunction with CoS, >and this is why I believe there will be two opposing views on how >far L2-LSP connections should be extended down to the user. One view >is that over provisioning and CoS work fine in the enterprise network today, >and therefore the L2-LSP should only be extended down to the point >at which the enterprise network connects to the SP network, >e.g. Ethernet DSLAM. I think the other view will be that over provisioning >and CoS in the enterprise network don't work and therefore L2-LSPs >should be provided end-to-end using only GMPLS capable switches. >I wouldn't have thought there would be much call for mixed >GMPLS/non-GMPLS enterprise networks (except for during migration) >due to the opposing views, e.g. if over provisioning is good enough for >one switch, why isn't it good enough for the other? Over-provisioned network is non-guaranteed network. It can't prevent occasional congestion at any rate. Further, network condition in private network is very diverse. You can't predict minimum service level inside every private network. Suppose that an ISP want to provide high-quality, video-phone service for enterprise users. Nothing can be guaranteed inside the over-provisioned customer network. If the service level is not satisfactory, customers will complain to service provider even if the main cause is in his network or destination network. We need some structure that coordinate service quality in both private sector, as well as in public sector consistently. An end-to-end signal is an effective means to communicate between user terminal, private switches and provider switches, and concert service quality at each different operations network. This is a goal that RSVP (RFC2205) pursued, however ineffective in local network because Ethernet industry didn't support RSVP over Ethernet (RFC2814, 2815, 2816). Now we have another opportunity to realize the goal using L2-LSP. Therefore, I believe end-to-end LSP is necessary even in enterprise network for the purpose of providing guaranteed commercial service. >I think where most people will agree is that using connections >in the SPs access network does make sense in order to provide efficient >link utilisation and per connection policing to ensure one customers >traffic can not affect another's. Regarding the ability to set up new >connections based on a request from the user (rather than constraining >RSVP signalling to inside the SP network), I think this will come >down to enterprise customer demand for improved service performance. >If a customer wasn't getting the performance they needed from their >SP and was offered a new service that allowed them to setup new L2-LSP >connections on demand that would offer better performance (e.g. for voice >calls or video conferences), then upgrading the user devices to support >the new service could well be an attractive proposition. If on the other Yes, this will be one of prospective service newly enabled using L2-LSP. >hand a customer buys a best effort service and considers the performance >to be 'good enough', then there would be no reason to go to the expense >of upgrading the user devices to support the set up of L2-LSPs. Ever increasing besteffort bandwidth will eventually not satisfy both providers and customers for many reasons. Service providers will have difficulty in collecting expenses enough to invest on increased bandwidth. This will in turn gradually downgrade overall communication quality, and applications requiring real-time service will never satisfy customers. We already see such symptom in some low-price service providers. I believe new service structure that can provide difference of service quality will offer benefit to customers, providers as well as vendors in the end. I also look forward to see you at Paris and continue helpful discussion. Sincerely Jaihyung Dr. Jaihyung Cho ETRI, Korea phone : 042) 860-5514 oversea: +82-42-860-5514 fax: +82-42-861-5550 -----?? ???----- ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com> ?? ??: 2005-07-14 ?? 1:10:30 ?? ??: "jaihyung@etri.re.kr" <jaihyung@etri.re.kr>, "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> ??: ??: RE: End-to-end L2-LSP Jaihyung, Thank you for taking the time to respond to my questions. I anticipated that your response would include "over-provisioning is the best answer" to decrease the chance of congestion on non-GMPLS switches. This is what is done in the LAN environment today in conjunction with CoS, and this is why I believe there will be two opposing views on how far L2-LSP connections should be extended down to the user. One view is that over provisioning and CoS work fine in the enterprise network today, and therefore the L2-LSP should only be extended down to the point at which the enterprise network connects to the SP network, e.g. Ethernet DSLAM. I think the other view will be that over provisioning and CoS in the enterprise network don't work and therefore L2-LSPs should be provided end-to-end using only GMPLS capable switches. I wouldn't have thought there would be much call for mixed GMPLS/non-GMPLS enterprise networks (except for during migration) due to the opposing views, e.g. if over provisioning is good enough for one switch, why isn't it good enough for the other? I think where most people will agree is that using connections in the SPs access network does make sense in order to provide efficient link utilisation and per connection policing to ensure one customers traffic can not affect another's. Regarding the ability to set up new connections based on a request from the user (rather than constraining RSVP signalling to inside the SP network), I think this will come down to enterprise customer demand for improved service performance. If a customer wasn't getting the performance they needed from their SP and was offered a new service that allowed them to setup new L2-LSP connections on demand that would offer better performance (e.g. for voice calls or video conferences), then upgrading the user devices to support the new service could well be an attractive proposition. If on the other hand a customer buys a best effort service and considers the performance to be 'good enough', then there would be no reason to go to the expense of upgrading the user devices to support the set up of L2-LSPs. I look forward to the discussions in Paris. Regards, Richard > -----Original Message----- > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > Sent: 13 July 2005 15:14 > To: Spencer,R,Richard,XDE73 R; ccamp@ops.ietf.org > Subject: RE: End-to-end L2-LSP > > > > > Dear Richard Spencer > > see in-line > > >Jaihyung, > > > >Perhaps I'm becoming too much of a pessimist;-) > > oh, not at all, > you are reasonable enough to dig up unclear knowledge > please go ahead until you finally get satisfactory answer. > > > >Ignoring the commercial aspects, please can you help clarify > a few technical > >questions I have in order to help me gain a better > understanding of the proposal? > > > >1. You say that one of the key reasons for using GMPLS in > the enterprise > >network is to provide end-to-end QoS control. However, you > also say that not > >all the switches will need to support GMPLS, only those at > bottle necks. > >Please can you expand on this to explain how you can have end-to-end > >QoS control when the L2-LSP is not end-to-end? > > > The contextual meaning of 'End-to-end LSP' is that GMPLS label is > assigned from source terminal and removed at destination terminal, > because RSVP-TE message is exchanged from host to host. > It may not perfectly guarantee QoS when the GMPLS deployment > is not complete, however it will at least be good enough > for end-user applications to tell those GMPLS implemented L2 switches > what frame must be given priority service when congestion occur. > > > >As being discussed > >on the PWE3 mailing list, the performance a service receives > is only as > >good as the worst performing server layer. Therefore, if there is a > >standard 802.1 workgroup switch between the user and the > L2-LSP that is > >experiencing congestion (e.g. due to a large amount of > peer-to-peer traffic), > >then the QoS that the user receives will only be as good as > that provided > >by the congested workgroup switch. > > > of course, ETH-GMPLS does not prevent congestion inevitably occurring > in those non-GMPLS deployed part. > > > >The only way I can see a network using a mixture of GMPLS > and non-GMPLS > >switches guaranteeing end-to-end QoS would be to design the > network in > >such a way that the non-GMPLS switches could never become congested. > > > Yes, that will be the way wise network designer improve their network. > > > >To do this would require that RSVP requests be used for ALL > traffic flows > >to ensure that there was enough bandwidth available, along > with strict > >per L2-LSP policing on the user devices (e.g. PC). > >However, the obvious flaw in this approach is that someone could > >just come along and connect > >an end device to a non-GMPLS switch and start transmitting traffic > >(without performing CAC/policing) causing the switch to > become congested. > >Therefore, please can you explain how end-to-end QoS can be > guaranteed > >in a mixed non-GMPLS and GMPLS switching environment? > > > Of course traffic came from non controlled ingress can not be > prevented > in non-GMPLS node. There's no magic. Only the > over-provisioning is the > best answer. However, once the aggregated traffic pass 'ANY' one of > GMPLS enabled node, those uncontrolled traffic will suffer > queuing discipline. > > > >2. Section 5.1 currently states "When the customer initiates > data transmission, > >the access switch maps the user flow into the L2 LSP. > Mapping procedure is > >subject to implementation choices, and is out of the scope of this > >document." In order to be able to map any traffic, first of > all the switch needs > >to know what type of traffic (e.g. based on source/dest MAC, > VLAN ID, 802.1p > >bits, etc) should be mapped to the L2-LSP. In Figure 1, if > an L2-LSP is set up > >from the DSLAM to the edge router based on a request from > the user (e.g. via > >a PC or VoIP phone), then won't the user also need to signal > what type of > >traffic should to be mapped to the L2-LSP? > > > Thank you for pointing out important thing. > That's the most unfortunate part I reluctantly agreed because the > work condition in the framework DT was not to span L2-LSP > beyond provider network. > As you noticed, mapping user traffic at provider access is awkward. > A smooth method will be for user terminal aware of the L2 label > to use by receiving RSVP-TE message. > However, there was strong argument that RSVP-TE must not > go beyond provider network because the DT is not chartered > to work in customer network. > That's why I requested amendment of CCAMP charter > to include user terminal explicitly in CCAMP work scope, so that > avoid such silly argument in future. > > > >3. In section 5.1, the text states that the aggregation > switch between the > >DSLAM and edge router may be a 802.1 switch, i.e. it doesn't > have to support > >GMPLS. To set up a connection, first of all there needs to > be a data plane > >capable of carrying control plane traffic between GMPLS > peers. Secondly, > >the GMPLS peers need reachability information in order to > forward control > >messages onto the next peer (i.e. a static or IGP route). If > the GMPLS peers > >(e.g. the DSLAM and edge router in Figure 1) are not > directly connected, > >then the GMPLS control packets will need to be forwarded by > the 802.1 switch. > >Would the idea be to have say a "control traffic VLAN" > configured between the > >DSLAM and edge router via the transit 802.1 switch so that > control packets > >could be forwarded transparently by the 802.1 switch? > > > Although it doesn't preclude use of VLAN, I think such > special configuration > for RSVP-TE control message is not necessary in legacy switches. > Legacy switches only need to pass RSVP-TE messages as they > pass normal data frames. Only the GMPLS implemented switches > need to distinguish frames containing RSVP-TE messages and > snoop them. > Detail of implementation is, of course, for future discussion. > > > >4. If we assume that control messages can be transparently forwarded > >by the intermediate 802.1 aggregation switch then how are > user frames > >forwarded in the data plane by the 802.1 switch? > > > Legacy switches do not distinguish control/data frames. > > > >If the DSLAM sets up a L2-LSP > >to the edge router using say VLAN ID 99, how does the switch > know that any > >packets coming in with a VLAN ID of 99 should be forwarded > out of the port > >connecting to the edge router? > > > oh, you misunderstood the L2-LSP. > The question you are asking depends on how L2 label is encoded. > there was hot debate in the DT about the issue how L2 label > is encoded. > My proposal is to use part of MAC address for label and the > other people > argued to use VLAN ID for label. The debate was not concluded > because the goal of the L2SC framework document is not to select > implementation method but to draw necessary work scope. > I believe this issue will be dealt in depth after Paris meeting. > So, please reserve your question until then and ask to those > VLAN proponents. > > > >5. During connection setup, I take it that the GMPLS peers > assume that > >resources are available within the intermediate 802.1 switch > to support > >the connection (as its transparent). However, as there is no > >CAC/policing being performed by the 802.1 switch > >(the L2-LSPs are transparent), and due to the any-to-any and > broadcast > >based nature of Ethernet (i.e. its CL properties), then how can this > >assumption be made? This comes back to my original question, how > >can end-to-end QoS be guaranteed in an environment where there is > >a mixture of 802.1 and GMPLS enabled switches? > > > yes, you came back to the question I answered in above ;-) > so, please see above. > GMPLS enabled nodes must be positioned at points congestion > is likely to be expected. > > Thanks > > Jaihyung > > > ============================================ > > > > -----?? ???----- > ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com> > ?? ??: 2005-07-13 ?? 8:53:13 > ?? ??: "jaihyung@etri.re.kr" <jaihyung@etri.re.kr>, > "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> > ??: > ??: RE: End-to-end L2-LSP > > > > > Jaihyung, > > Perhaps I'm becoming too much of a pessimist;-) Ignoring the > commercial aspects, please can you help clarify a few > technical questions I have in order to help me gain a better > understanding of the proposal? > > 1. You say that one of the key reasons for using GMPLS in the > enterprise network is to provide end-to-end QoS control. > However, you also say that not all the switches will need to > support GMPLS, only those at bottle necks. Please can you > expand on this to explain how you can have end-to-end QoS > control when the L2-LSP is not end-to-end? As being discussed > on the PWE3 mailing list, the performance a service receives > is only as good as the worst performing server layer. > Therefore, if there is a standard 802.1 workgroup switch > between the user and the L2-LSP that is experiencing > congestion (e.g. due to a large amount of peer-to-peer > traffic), then the QoS that the user receives will only be as > good as that provided by the congested workgroup switch. > > The only way I can see a network using a mixture of GMPLS and > non-GMPLS switches guaranteeing end-to-end QoS would be to > design the network in such a way that the non-GMPLS switches > could never become congested. To do this would require that > RSVP requests be used for ALL traffic flows to ensure that > there was enough bandwidth available, along with strict per > L2-LSP policing on the user devices (e.g. PC). However, the > obvious flaw in this approach is that someone could just come > along and connect an end device to a non-GMPLS switch and > start transmitting traffic (without performing CAC/policing) > causing the switch to become congested. Therefore, please can > you explain how end-to-end QoS can be guaranteed in a mixed > non-GMPLS and GMPLS switching environment? > > 2. Section 5.1 currently states "When the customer initiates > data transmission, the access switch maps the user flow into > the L2 LSP. Mapping procedure is subject to implementation > choices, and is out of the scope of this document." In order > to be able to map any traffic, first of all the switch needs > to know what type of traffic (e.g. based on source/dest MAC, > VLAN ID, 802.1p bits, etc) should be mapped to the L2-LSP. In > Figure 1, if an L2-LSP is set up from the DSLAM to the edge > router based on a request from the user (e.g. via a PC or > VoIP phone), then won't the user also need to signal what > type of traffic should to be mapped to the L2-LSP? > > 3. In section 5.1, the text states that the aggregation > switch between the DSLAM and edge router may be a 802.1 > switch, i.e. it doesn't have to support GMPLS. To set up a > connection, first of all there needs to be a data plane > capable of carrying control plane traffic between GMPLS > peers. Secondly, the GMPLS peers need reachability > information in order to forward control messages onto the > next peer (i.e. a static or IGP route). If the GMPLS peers > (e.g. the DSLAM and edge router in Figure 1) are not directly > connected, then the GMPLS control packets will need to be > forwarded by the 802.1 switch. Would the idea be to have say > a "control traffic VLAN" configured between the DSLAM and > edge router via the transit 802.1 switch so that control > packets could be forwarded transparently by the 802.1 switch? > > 4. If we assume that control messages can be transparently > forwarded by the intermediate 802.1 aggregation switch then > how are user frames forwarded in the data plane by the 802.1 > switch? If the DSLAM sets up a L2-LSP to the edge router > using say VLAN ID 99, how does the switch know that any > packets coming in with a VLAN ID of 99 should be forwarded > out of the port connecting to the edge router? > > 5. During connection setup, I take it that the GMPLS peers > assume that resources are available within the intermediate > 802.1 switch to support the connection (as its transparent). > However, as there is no CAC/policing being performed by the > 802.1 switch (the L2-LSPs are transparent), and due to the > any-to-any and broadcast based nature of Ethernet (i.e. its > CL properties), then how can this assumption be made? This > comes back to my original question, how can end-to-end QoS be > guaranteed in an environment where there is a mixture of > 802.1 and GMPLS enabled switches? > > Thanks, > Richard > > > -----Original Message----- > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > > Sent: 12 July 2005 09:21 > > To: Spencer,R,Richard,XDE73 R; ccamp@ops.ietf.org > > Subject: RE: End-to-end L2-LSP > > > > > > > > > > Dear Richard, > > > > Yes, I believe it is not only commercially profitable but it > > may also grow total > > revenue of service operators by introducing new services. > > > > For example, think the example of IP phone you referred. > > How do you think you can provide reliable VoIP service and > > safely authenticate phone users and charge fees as counted by > > packet usage > > using simple Ethernet based xDSL network ? > > > > There's hardly new service that providers may offer using > > Ethernet based > > architecture, and collect sufficient profit to investment. > > > > A sophisticated ATM switch might do the work if it is not > expensive. > > However, my intension is modifiying existing design of > > Ethernet hardware > > with a little bit additional cost, and provide the level of > > service that > > ATM switches might offer. > > > > We have initial design of the Ethernet-MPLS hardware. I don't think > > GMPLS enabled Ethernet switch will be so complex. The market price > > of GMPLS enabled switch will not be so expensive than normal > > Ethernet switches. > > > > Further, it only requires upgrade of switches in some > > bottleneck points of network. > > Read the framework documents of L2SC. It doesn't require > > entire replacement > > of switches to introduce new service. > > > > Every year, ISPs purchase new access switches and replaces > > old machines. > > With a bit additional cost, ISPs may use the GMPLS enabled > switch for > > normal Ethernet switching in initial deployment. > > > > When the GMPLS function is enabled in some upgraded part of > network, > > ISPs may provide high quality, reliable VoIP service to > xDSL customers > > using RSVP-TE enabled VoIP phone. > > The high quality video-phone service enabled by GMPLS will be > > competable to legacy voice only phone service. > > > > > > There are lots of things I can argue about commercial aspects > > of this technique. > > However I don't think it is not appropriate to discuss this > > issue in this mailing list. > > All I can say is, > > yes, I believe that it may be commercially viable. > > > > > > Thanks > > > > Jaihyung > > > > ================================================ > > > > Do you really think the benefits of extended GMPLS down to > > the end user could ever justify the costs involved in > > upgrading LAN switches and things like IP phones (simple low > > cost commodity items) to support GMPLS signalling/routing? > > > > Your proposal may be technically possible (though it would > > require new hardware and significant software upgrades), but > > commercially I would consider this to be a non-starter. > > > > Regards, > > Richard > > > > > -----Original Message----- > > > From: owner-ccamp@ops.ietf.org > [mailto:owner-ccamp@ops.ietf.org]On > > > Behalf Of CHO, JAI HYUNG > > > Sent: 12 July 2005 01:10 > > > To: ccamp@ops.ietf.org > > > Subject: End-to-end L2-LSP > > > > > > > > > > > > Dear all, > > > > > > Greetings. > > > > > > I am a co-editor of the Ethernet-GMPLS framework document. > > > (draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt) > > > > > > I am writing this to explain the motivation behind the > > > proposal of scenario-1, and to raise the issue of > > > end-to-end LSP to be included in CCAMP charter. > > > > > > An important goal intended in scenario-1 is > > > QoS control over end-to-end packet delivery using LSP. > > > In other words, the purpose of scenario-1 is extending coverage > > > of GMPLS control including customer network and end-user device, > > > such as PC, VoIP phone, IP-TV, or even future 4G mobile phone. > > > End-user applications will have a means to request resources > > > to network using GMPLS signal. > > > > > > Unfortunately, this view is not included in the > > > framework because there was strong opinion that > > > L2-LSP must not span beyond provider network because > > > CCAMP is not chartered to work on customer > > > network. I am wondering if this is right argument. > > > Perhaps the charter can be amended to include this work. > > > > > > I believe there is little technical reason to preclude > > > LSP being used in customer network. It is a reasonable > > > assumption that private company may also use GMPLS > > > implemented Ethernet switch, once the L2SC work is successful. > > > Currently, CCAMP is the only place discussing the matter of > > > Ethernet and LSP. There is no other WG dealing similar issue. > > > If CCAMP wants to work on Ethernet, > > > I think the coverage of common GMPLS control must be > > > extended to include all area where Ethernet is deployed. > > > Currently, Ethernet is dominantly deployed in first-mile > > > network where access network as well as customer network > > > and user hosts are usually included. Therefore, user host must > > > be considered as initiation and termination point of L2-LSP. > > > > > > Another important reason to include user host in CCAMP > > > work scope is to achieve consistent end-to-end control of QoS. > > > End-to-end QoS has long been desired goal since RSVP was > > > first designed. This goal can be best achieved using > > > end-to-end GMPLS signal and L2-LSP. > > > Currently, RSVP (RFC2205) is available in many user hosts. > > > There is only a little difference between RSVP and RSVP-TE. > > > If the RSVP-TE signal of provider network can reach user host > > > across Ethernet based private network, an LSP can be established > > > from application to application via the provider backbone. > > > Service providers will be able to control application flows > > > in L2-LSP level as well as in aggregated LSP pipe of similar > > > application type. A variety of new services will be enabled using > > > the enhanced capability of distinguishing and controlling each > > > individual application flow. > > > > > > I personally believe that end-to-end L2-LSP may offer > > > innovative solutions for servicing IP-TV as well as fast > mobility. > > > There have been some experimental approaches combining > > > MPLS and IP mobility. However such potential of the new > > > application can only be explored when end-user host is > > > included in the scope of CCAMP work area. > > > > > > Therefore, I hereby sincerely request people's attention > > > and support on this perspective of end-to-end LSP. > > > If there is enough number of people agree on this view, > > > the WG can request the necessary update of the charter. > > > > > > Thank you. > > > > > > Jaihyung Cho > > > > > > > > > > > > ETRI, Korea > > > phone : 042) 860-5514 > > > oversea: +82-42-860-5514 > > > fax: +82-42-861-5550 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 13 Jul 2005 16:11:35 +0000 content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: End-to-end L2-LSP Date: Wed, 13 Jul 2005 17:10:30 +0100 Message-ID: <B5E87B043D4C514389141E2661D255EC074C4F5A@i2km41-ukdy.domain1.systemhost.net> Thread-Topic: End-to-end L2-LSP Thread-Index: AcWHtSqnmWn31nu6RQGbw8KrVOPrAAAApXPg From: <richard.spencer@bt.com> To: <jaihyung@etri.re.kr>, <ccamp@ops.ietf.org> Jaihyung, Thank you for taking the time to respond to my questions. I anticipated = that your response would include "over-provisioning is the best answer" = to decrease the chance of congestion on non-GMPLS switches. This is what = is done in the LAN environment today in conjunction with CoS, and this = is why I believe there will be two opposing views on how far L2-LSP = connections should be extended down to the user. One view is that over = provisioning and CoS work fine in the enterprise network today, and = therefore the L2-LSP should only be extended down to the point at which = the enterprise network connects to the SP network, e.g. Ethernet DSLAM. = I think the other view will be that over provisioning and CoS in the = enterprise network don't work and therefore L2-LSPs should be provided = end-to-end using only GMPLS capable switches. I wouldn't have thought = there would be much call for mixed GMPLS/non-GMPLS enterprise networks = (except for during migration) due to the opposing views, e.g. if over = provisioning is good enough for one switch, why isn't it good enough for = the other? I think where most people will agree is that using connections in the = SPs access network does make sense in order to provide efficient link = utilisation and per connection policing to ensure one customers traffic = can not affect another's. Regarding the ability to set up new = connections based on a request from the user (rather than constraining = RSVP signalling to inside the SP network), I think this will come down = to enterprise customer demand for improved service performance. If a = customer wasn't getting the performance they needed from their SP and = was offered a new service that allowed them to setup new L2-LSP = connections on demand that would offer better performance (e.g. for = voice calls or video conferences), then upgrading the user devices to = support the new service could well be an attractive proposition. If on = the other hand a customer buys a best effort service and considers the = performance to be 'good enough', then there would be no reason to go to = the expense of upgrading the user devices to support the set up of = L2-LSPs. I look forward to the discussions in Paris. Regards, Richard > -----Original Message----- > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > Sent: 13 July 2005 15:14 > To: Spencer,R,Richard,XDE73 R; ccamp@ops.ietf.org > Subject: RE: End-to-end L2-LSP >=20 >=20 > =20 > =20 > Dear Richard Spencer=20 > =20 > see in-line > =20 > >Jaihyung,=20 > >=20 > >Perhaps I'm becoming too much of a pessimist;-)=20 > =20 > oh, not at all,=20 > you are reasonable enough to dig up unclear knowledge=20 > please go ahead until you finally get satisfactory answer. > =20 >=20 > >Ignoring the commercial aspects, please can you help clarify=20 > a few technical=20 > >questions I have in order to help me gain a better=20 > understanding of the proposal?=20 > >=20 > >1. You say that one of the key reasons for using GMPLS in=20 > the enterprise=20 > >network is to provide end-to-end QoS control. However, you=20 > also say that not=20 > >all the switches will need to support GMPLS, only those at=20 > bottle necks.=20 > >Please can you expand on this to explain how you can have end-to-end=20 > >QoS control when the L2-LSP is not end-to-end?=20 > =20 >=20 > The contextual meaning of 'End-to-end LSP' is that GMPLS label is=20 > assigned from source terminal and removed at destination terminal,=20 > because RSVP-TE message is exchanged from host to host.=20 > It may not perfectly guarantee QoS when the GMPLS deployment=20 > is not complete, however it will at least be good enough=20 > for end-user applications to tell those GMPLS implemented L2 switches=20 > what frame must be given priority service when congestion occur. > =20 >=20 > >As being discussed=20 > >on the PWE3 mailing list, the performance a service receives=20 > is only as=20 > >good as the worst performing server layer. Therefore, if there is a=20 > >standard 802.1 workgroup switch between the user and the=20 > L2-LSP that is=20 > >experiencing congestion (e.g. due to a large amount of=20 > peer-to-peer traffic),=20 > >then the QoS that the user receives will only be as good as=20 > that provided=20 > >by the congested workgroup switch. > =20 >=20 > of course, ETH-GMPLS does not prevent congestion inevitably occurring=20 > in those non-GMPLS deployed part.=20 > =20 >=20 > >The only way I can see a network using a mixture of GMPLS=20 > and non-GMPLS=20 > >switches guaranteeing end-to-end QoS would be to design the=20 > network in=20 > >such a way that the non-GMPLS switches could never become congested. > =20 >=20 > Yes, that will be the way wise network designer improve their network. > =20 >=20 > >To do this would require that RSVP requests be used for ALL=20 > traffic flows=20 > >to ensure that there was enough bandwidth available, along=20 > with strict=20 > >per L2-LSP policing on the user devices (e.g. PC).=20 > >However, the obvious flaw in this approach is that someone could=20 > >just come along and connect=20 > >an end device to a non-GMPLS switch and start transmitting traffic=20 > >(without performing CAC/policing) causing the switch to=20 > become congested.=20 > >Therefore, please can you explain how end-to-end QoS can be=20 > guaranteed=20 > >in a mixed non-GMPLS and GMPLS switching environment? > =20 >=20 > Of course traffic came from non controlled ingress can not be=20 > prevented=20 > in non-GMPLS node. There's no magic. Only the=20 > over-provisioning is the=20 > best answer. However, once the aggregated traffic pass 'ANY' one of=20 > GMPLS enabled node, those uncontrolled traffic will suffer=20 > queuing discipline. > =20 >=20 > >2. Section 5.1 currently states "When the customer initiates=20 > data transmission,=20 > >the access switch maps the user flow into the L2 LSP.=20 > Mapping procedure is=20 > >subject to implementation choices, and is out of the scope of this=20 > >document." In order to be able to map any traffic, first of=20 > all the switch needs=20 > >to know what type of traffic (e.g. based on source/dest MAC,=20 > VLAN ID, 802.1p=20 > >bits, etc) should be mapped to the L2-LSP. In Figure 1, if=20 > an L2-LSP is set up=20 > >from the DSLAM to the edge router based on a request from=20 > the user (e.g. via=20 > >a PC or VoIP phone), then won't the user also need to signal=20 > what type of=20 > >traffic should to be mapped to the L2-LSP? > =20 >=20 > Thank you for pointing out important thing.=20 > That's the most unfortunate part I reluctantly agreed because the=20 > work condition in the framework DT was not to span L2-LSP=20 > beyond provider network.=20 > As you noticed, mapping user traffic at provider access is awkward.=20 > A smooth method will be for user terminal aware of the L2 label=20 > to use by receiving RSVP-TE message.=20 > However, there was strong argument that RSVP-TE must not=20 > go beyond provider network because the DT is not chartered=20 > to work in customer network.=20 > That's why I requested amendment of CCAMP charter=20 > to include user terminal explicitly in CCAMP work scope, so that=20 > avoid such silly argument in future. > =20 >=20 > >3. In section 5.1, the text states that the aggregation=20 > switch between the=20 > >DSLAM and edge router may be a 802.1 switch, i.e. it doesn't=20 > have to support=20 > >GMPLS. To set up a connection, first of all there needs to=20 > be a data plane=20 > >capable of carrying control plane traffic between GMPLS=20 > peers. Secondly,=20 > >the GMPLS peers need reachability information in order to=20 > forward control=20 > >messages onto the next peer (i.e. a static or IGP route). If=20 > the GMPLS peers=20 > >(e.g. the DSLAM and edge router in Figure 1) are not=20 > directly connected,=20 > >then the GMPLS control packets will need to be forwarded by=20 > the 802.1 switch.=20 > >Would the idea be to have say a "control traffic VLAN"=20 > configured between the=20 > >DSLAM and edge router via the transit 802.1 switch so that=20 > control packets=20 > >could be forwarded transparently by the 802.1 switch? > =20 >=20 > Although it doesn't preclude use of VLAN, I think such=20 > special configuration=20 > for RSVP-TE control message is not necessary in legacy switches.=20 > Legacy switches only need to pass RSVP-TE messages as they=20 > pass normal data frames. Only the GMPLS implemented switches=20 > need to distinguish frames containing RSVP-TE messages and=20 > snoop them.=20 > Detail of implementation is, of course, for future discussion. > =20 >=20 > >4. If we assume that control messages can be transparently forwarded=20 > >by the intermediate 802.1 aggregation switch then how are=20 > user frames=20 > >forwarded in the data plane by the 802.1 switch?=20 > =20 >=20 > Legacy switches do not distinguish control/data frames. > =20 >=20 > >If the DSLAM sets up a L2-LSP=20 > >to the edge router using say VLAN ID 99, how does the switch=20 > know that any=20 > >packets coming in with a VLAN ID of 99 should be forwarded=20 > out of the port=20 > >connecting to the edge router? > =20 >=20 > oh, you misunderstood the L2-LSP.=20 > The question you are asking depends on how L2 label is encoded.=20 > there was hot debate in the DT about the issue how L2 label=20 > is encoded.=20 > My proposal is to use part of MAC address for label and the=20 > other people=20 > argued to use VLAN ID for label. The debate was not concluded=20 > because the goal of the L2SC framework document is not to select=20 > implementation method but to draw necessary work scope.=20 > I believe this issue will be dealt in depth after Paris meeting.=20 > So, please reserve your question until then and ask to those=20 > VLAN proponents. > =20 >=20 > >5. During connection setup, I take it that the GMPLS peers=20 > assume that=20 > >resources are available within the intermediate 802.1 switch=20 > to support=20 > >the connection (as its transparent). However, as there is no=20 > >CAC/policing being performed by the 802.1 switch=20 > >(the L2-LSPs are transparent), and due to the any-to-any and=20 > broadcast=20 > >based nature of Ethernet (i.e. its CL properties), then how can this=20 > >assumption be made? This comes back to my original question, how=20 > >can end-to-end QoS be guaranteed in an environment where there is=20 > >a mixture of 802.1 and GMPLS enabled switches? > =20 >=20 > yes, you came back to the question I answered in above ;-)=20 > so, please see above.=20 > GMPLS enabled nodes must be positioned at points congestion=20 > is likely to be expected. > =20 > Thanks > =20 > Jaihyung > =20 >=20 > = =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=20 >=20 > =20 >=20 > -----?? ???-----=20 > ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com>=20 > ?? ??: 2005-07-13 ?? 8:53:13=20 > ?? ??: "jaihyung@etri.re.kr" <jaihyung@etri.re.kr>,=20 > "ccamp@ops.ietf.org" <ccamp@ops.ietf.org>=20 > ??:=20 > ??: RE: End-to-end L2-LSP=20 >=20 >=20 >=20 >=20 > Jaihyung,=20 >=20 > Perhaps I'm becoming too much of a pessimist;-) Ignoring the=20 > commercial aspects, please can you help clarify a few=20 > technical questions I have in order to help me gain a better=20 > understanding of the proposal?=20 >=20 > 1. You say that one of the key reasons for using GMPLS in the=20 > enterprise network is to provide end-to-end QoS control.=20 > However, you also say that not all the switches will need to=20 > support GMPLS, only those at bottle necks. Please can you=20 > expand on this to explain how you can have end-to-end QoS=20 > control when the L2-LSP is not end-to-end? As being discussed=20 > on the PWE3 mailing list, the performance a service receives=20 > is only as good as the worst performing server layer.=20 > Therefore, if there is a standard 802.1 workgroup switch=20 > between the user and the L2-LSP that is experiencing=20 > congestion (e.g. due to a large amount of peer-to-peer=20 > traffic), then the QoS that the user receives will only be as=20 > good as that provided by the congested workgroup switch.=20 >=20 > The only way I can see a network using a mixture of GMPLS and=20 > non-GMPLS switches guaranteeing end-to-end QoS would be to=20 > design the network in such a way that the non-GMPLS switches=20 > could never become congested. To do this would require that=20 > RSVP requests be used for ALL traffic flows to ensure that=20 > there was enough bandwidth available, along with strict per=20 > L2-LSP policing on the user devices (e.g. PC). However, the=20 > obvious flaw in this approach is that someone could just come=20 > along and connect an end device to a non-GMPLS switch and=20 > start transmitting traffic (without performing CAC/policing)=20 > causing the switch to become congested. Therefore, please can=20 > you explain how end-to-end QoS can be guaranteed in a mixed=20 > non-GMPLS and GMPLS switching environment?=20 >=20 > 2. Section 5.1 currently states "When the customer initiates=20 > data transmission, the access switch maps the user flow into=20 > the L2 LSP. Mapping procedure is subject to implementation=20 > choices, and is out of the scope of this document." In order=20 > to be able to map any traffic, first of all the switch needs=20 > to know what type of traffic (e.g. based on source/dest MAC,=20 > VLAN ID, 802.1p bits, etc) should be mapped to the L2-LSP. In=20 > Figure 1, if an L2-LSP is set up from the DSLAM to the edge=20 > router based on a request from the user (e.g. via a PC or=20 > VoIP phone), then won't the user also need to signal what=20 > type of traffic should to be mapped to the L2-LSP?=20 >=20 > 3. In section 5.1, the text states that the aggregation=20 > switch between the DSLAM and edge router may be a 802.1=20 > switch, i.e. it doesn't have to support GMPLS. To set up a=20 > connection, first of all there needs to be a data plane=20 > capable of carrying control plane traffic between GMPLS=20 > peers. Secondly, the GMPLS peers need reachability=20 > information in order to forward control messages onto the=20 > next peer (i.e. a static or IGP route). If the GMPLS peers=20 > (e.g. the DSLAM and edge router in Figure 1) are not directly=20 > connected, then the GMPLS control packets will need to be=20 > forwarded by the 802.1 switch. Would the idea be to have say=20 > a "control traffic VLAN" configured between the DSLAM and=20 > edge router via the transit 802.1 switch so that control=20 > packets could be forwarded transparently by the 802.1 switch?=20 >=20 > 4. If we assume that control messages can be transparently=20 > forwarded by the intermediate 802.1 aggregation switch then=20 > how are user frames forwarded in the data plane by the 802.1=20 > switch? If the DSLAM sets up a L2-LSP to the edge router=20 > using say VLAN ID 99, how does the switch know that any=20 > packets coming in with a VLAN ID of 99 should be forwarded=20 > out of the port connecting to the edge router?=20 >=20 > 5. During connection setup, I take it that the GMPLS peers=20 > assume that resources are available within the intermediate=20 > 802.1 switch to support the connection (as its transparent).=20 > However, as there is no CAC/policing being performed by the=20 > 802.1 switch (the L2-LSPs are transparent), and due to the=20 > any-to-any and broadcast based nature of Ethernet (i.e. its=20 > CL properties), then how can this assumption be made? This=20 > comes back to my original question, how can end-to-end QoS be=20 > guaranteed in an environment where there is a mixture of=20 > 802.1 and GMPLS enabled switches?=20 >=20 > Thanks,=20 > Richard=20 >=20 > > -----Original Message-----=20 > > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr]=20 > > Sent: 12 July 2005 09:21=20 > > To: Spencer,R,Richard,XDE73 R; ccamp@ops.ietf.org=20 > > Subject: RE: End-to-end L2-LSP=20 > >=20 > >=20 > >=20 > >=20 > > Dear Richard,=20 > >=20 > > Yes, I believe it is not only commercially profitable but it=20 > > may also grow total=20 > > revenue of service operators by introducing new services.=20 > >=20 > > For example, think the example of IP phone you referred.=20 > > How do you think you can provide reliable VoIP service and=20 > > safely authenticate phone users and charge fees as counted by=20 > > packet usage=20 > > using simple Ethernet based xDSL network ?=20 > >=20 > > There's hardly new service that providers may offer using=20 > > Ethernet based=20 > > architecture, and collect sufficient profit to investment.=20 > >=20 > > A sophisticated ATM switch might do the work if it is not=20 > expensive.=20 > > However, my intension is modifiying existing design of=20 > > Ethernet hardware=20 > > with a little bit additional cost, and provide the level of=20 > > service that=20 > > ATM switches might offer.=20 > >=20 > > We have initial design of the Ethernet-MPLS hardware. I don't think=20 > > GMPLS enabled Ethernet switch will be so complex. The market price=20 > > of GMPLS enabled switch will not be so expensive than normal=20 > > Ethernet switches.=20 > >=20 > > Further, it only requires upgrade of switches in some=20 > > bottleneck points of network.=20 > > Read the framework documents of L2SC. It doesn't require=20 > > entire replacement=20 > > of switches to introduce new service.=20 > >=20 > > Every year, ISPs purchase new access switches and replaces=20 > > old machines.=20 > > With a bit additional cost, ISPs may use the GMPLS enabled=20 > switch for=20 > > normal Ethernet switching in initial deployment.=20 > >=20 > > When the GMPLS function is enabled in some upgraded part of=20 > network,=20 > > ISPs may provide high quality, reliable VoIP service to=20 > xDSL customers=20 > > using RSVP-TE enabled VoIP phone.=20 > > The high quality video-phone service enabled by GMPLS will be=20 > > competable to legacy voice only phone service.=20 > >=20 > >=20 > > There are lots of things I can argue about commercial aspects=20 > > of this technique.=20 > > However I don't think it is not appropriate to discuss this=20 > > issue in this mailing list.=20 > > All I can say is,=20 > > yes, I believe that it may be commercially viable.=20 > >=20 > >=20 > > Thanks=20 > >=20 > > Jaihyung=20 > >=20 > > = =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=20 > >=20 > > Do you really think the benefits of extended GMPLS down to=20 > > the end user could ever justify the costs involved in=20 > > upgrading LAN switches and things like IP phones (simple low=20 > > cost commodity items) to support GMPLS signalling/routing?=20 > >=20 > > Your proposal may be technically possible (though it would=20 > > require new hardware and significant software upgrades), but=20 > > commercially I would consider this to be a non-starter.=20 > >=20 > > Regards,=20 > > Richard=20 > >=20 > > > -----Original Message-----=20 > > > From: owner-ccamp@ops.ietf.org=20 > [mailto:owner-ccamp@ops.ietf.org]On=20 > > > Behalf Of CHO, JAI HYUNG=20 > > > Sent: 12 July 2005 01:10=20 > > > To: ccamp@ops.ietf.org=20 > > > Subject: End-to-end L2-LSP=20 > > >=20 > > >=20 > > > =20 > > > Dear all,=20 > > >=20 > > > Greetings.=20 > > >=20 > > > I am a co-editor of the Ethernet-GMPLS framework document.=20 > > > (draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt)=20 > > >=20 > > > I am writing this to explain the motivation behind the=20 > > > proposal of scenario-1, and to raise the issue of=20 > > > end-to-end LSP to be included in CCAMP charter.=20 > > >=20 > > > An important goal intended in scenario-1 is=20 > > > QoS control over end-to-end packet delivery using LSP.=20 > > > In other words, the purpose of scenario-1 is extending coverage=20 > > > of GMPLS control including customer network and end-user device,=20 > > > such as PC, VoIP phone, IP-TV, or even future 4G mobile phone.=20 > > > End-user applications will have a means to request resources=20 > > > to network using GMPLS signal.=20 > > >=20 > > > Unfortunately, this view is not included in the=20 > > > framework because there was strong opinion that=20 > > > L2-LSP must not span beyond provider network because=20 > > > CCAMP is not chartered to work on customer=20 > > > network. I am wondering if this is right argument.=20 > > > Perhaps the charter can be amended to include this work.=20 > > >=20 > > > I believe there is little technical reason to preclude=20 > > > LSP being used in customer network. It is a reasonable=20 > > > assumption that private company may also use GMPLS=20 > > > implemented Ethernet switch, once the L2SC work is successful.=20 > > > Currently, CCAMP is the only place discussing the matter of=20 > > > Ethernet and LSP. There is no other WG dealing similar issue.=20 > > > If CCAMP wants to work on Ethernet,=20 > > > I think the coverage of common GMPLS control must be=20 > > > extended to include all area where Ethernet is deployed.=20 > > > Currently, Ethernet is dominantly deployed in first-mile=20 > > > network where access network as well as customer network=20 > > > and user hosts are usually included. Therefore, user host must=20 > > > be considered as initiation and termination point of L2-LSP.=20 > > >=20 > > > Another important reason to include user host in CCAMP=20 > > > work scope is to achieve consistent end-to-end control of QoS.=20 > > > End-to-end QoS has long been desired goal since RSVP was=20 > > > first designed. This goal can be best achieved using=20 > > > end-to-end GMPLS signal and L2-LSP.=20 > > > Currently, RSVP (RFC2205) is available in many user hosts.=20 > > > There is only a little difference between RSVP and RSVP-TE.=20 > > > If the RSVP-TE signal of provider network can reach user host=20 > > > across Ethernet based private network, an LSP can be established=20 > > > from application to application via the provider backbone.=20 > > > Service providers will be able to control application flows=20 > > > in L2-LSP level as well as in aggregated LSP pipe of similar=20 > > > application type. A variety of new services will be enabled using=20 > > > the enhanced capability of distinguishing and controlling each=20 > > > individual application flow.=20 > > >=20 > > > I personally believe that end-to-end L2-LSP may offer=20 > > > innovative solutions for servicing IP-TV as well as fast=20 > mobility.=20 > > > There have been some experimental approaches combining=20 > > > MPLS and IP mobility. However such potential of the new=20 > > > application can only be explored when end-user host is=20 > > > included in the scope of CCAMP work area.=20 > > >=20 > > > Therefore, I hereby sincerely request people's attention=20 > > > and support on this perspective of end-to-end LSP.=20 > > > If there is enough number of people agree on this view,=20 > > > the WG can request the necessary update of the charter.=20 > > >=20 > > > Thank you.=20 > > > =20 > > > Jaihyung Cho=20 > > > =20 > > > =20 > > > =20 > > > ETRI, Korea=20 > > > phone : 042) 860-5514=20 > > > oversea: +82-42-860-5514=20 > > > fax: +82-42-861-5550=20 > > >=20 > > >=20 > > >=20 > > >=20 > > >=20 > > >=20 > > >=20 > >=20 > >=20 > >=20 > >=20 >=20 >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 13 Jul 2005 14:14:55 +0000 Thread-Topic: End-to-end L2-LSP thread-index: AcWHtSqnmWn31nu6RQGbw8KrVOPrAA== Reply-To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> From: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> To: <richard.spencer@bt.com>, <ccamp@ops.ietf.org> Cc: Bcc: Subject: RE: End-to-end L2-LSP Date: Wed, 13 Jul 2005 23:14:22 +0900 Comment: GQ19@|@ZEk=E?,18?x, BcN1b<z:P<.F@, 4c4g Message-ID: <66fc801c587b5$2aac6a60$8310fe81@email1> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Class: urn:content-classes:message Dear Richard Spencer see in-line >Jaihyung, > >Perhaps I'm becoming too much of a pessimist;-) oh, not at all, you are reasonable enough to dig up unclear knowledge please go ahead until you finally get satisfactory answer. >Ignoring the commercial aspects, please can you help clarify a few technical >questions I have in order to help me gain a better understanding of the proposal? > >1. You say that one of the key reasons for using GMPLS in the enterprise >network is to provide end-to-end QoS control. However, you also say that not >all the switches will need to support GMPLS, only those at bottle necks. >Please can you expand on this to explain how you can have end-to-end >QoS control when the L2-LSP is not end-to-end? The contextual meaning of 'End-to-end LSP' is that GMPLS label is assigned from source terminal and removed at destination terminal, because RSVP-TE message is exchanged from host to host. It may not perfectly guarantee QoS when the GMPLS deployment is not complete, however it will at least be good enough for end-user applications to tell those GMPLS implemented L2 switches what frame must be given priority service when congestion occur. >As being discussed >on the PWE3 mailing list, the performance a service receives is only as >good as the worst performing server layer. Therefore, if there is a >standard 802.1 workgroup switch between the user and the L2-LSP that is >experiencing congestion (e.g. due to a large amount of peer-to-peer traffic), >then the QoS that the user receives will only be as good as that provided >by the congested workgroup switch. of course, ETH-GMPLS does not prevent congestion inevitably occurring in those non-GMPLS deployed part. >The only way I can see a network using a mixture of GMPLS and non-GMPLS >switches guaranteeing end-to-end QoS would be to design the network in >such a way that the non-GMPLS switches could never become congested. Yes, that will be the way wise network designer improve their network. >To do this would require that RSVP requests be used for ALL traffic flows >to ensure that there was enough bandwidth available, along with strict >per L2-LSP policing on the user devices (e.g. PC). >However, the obvious flaw in this approach is that someone could >just come along and connect >an end device to a non-GMPLS switch and start transmitting traffic >(without performing CAC/policing) causing the switch to become congested. >Therefore, please can you explain how end-to-end QoS can be guaranteed >in a mixed non-GMPLS and GMPLS switching environment? Of course traffic came from non controlled ingress can not be prevented in non-GMPLS node. There's no magic. Only the over-provisioning is the best answer. However, once the aggregated traffic pass 'ANY' one of GMPLS enabled node, those uncontrolled traffic will suffer queuing discipline. >2. Section 5.1 currently states "When the customer initiates data transmission, >the access switch maps the user flow into the L2 LSP. Mapping procedure is >subject to implementation choices, and is out of the scope of this >document." In order to be able to map any traffic, first of all the switch needs >to know what type of traffic (e.g. based on source/dest MAC, VLAN ID, 802.1p >bits, etc) should be mapped to the L2-LSP. In Figure 1, if an L2-LSP is set up >from the DSLAM to the edge router based on a request from the user (e.g. via >a PC or VoIP phone), then won't the user also need to signal what type of >traffic should to be mapped to the L2-LSP? Thank you for pointing out important thing. That's the most unfortunate part I reluctantly agreed because the work condition in the framework DT was not to span L2-LSP beyond provider network. As you noticed, mapping user traffic at provider access is awkward. A smooth method will be for user terminal aware of the L2 label to use by receiving RSVP-TE message. However, there was strong argument that RSVP-TE must not go beyond provider network because the DT is not chartered to work in customer network. That's why I requested amendment of CCAMP charter to include user terminal explicitly in CCAMP work scope, so that avoid such silly argument in future. >3. In section 5.1, the text states that the aggregation switch between the >DSLAM and edge router may be a 802.1 switch, i.e. it doesn't have to support >GMPLS. To set up a connection, first of all there needs to be a data plane >capable of carrying control plane traffic between GMPLS peers. Secondly, >the GMPLS peers need reachability information in order to forward control >messages onto the next peer (i.e. a static or IGP route). If the GMPLS peers >(e.g. the DSLAM and edge router in Figure 1) are not directly connected, >then the GMPLS control packets will need to be forwarded by the 802.1 switch. >Would the idea be to have say a "control traffic VLAN" configured between the >DSLAM and edge router via the transit 802.1 switch so that control packets >could be forwarded transparently by the 802.1 switch? Although it doesn't preclude use of VLAN, I think such special configuration for RSVP-TE control message is not necessary in legacy switches. Legacy switches only need to pass RSVP-TE messages as they pass normal data frames. Only the GMPLS implemented switches need to distinguish frames containing RSVP-TE messages and snoop them. Detail of implementation is, of course, for future discussion. >4. If we assume that control messages can be transparently forwarded >by the intermediate 802.1 aggregation switch then how are user frames >forwarded in the data plane by the 802.1 switch? Legacy switches do not distinguish control/data frames. >If the DSLAM sets up a L2-LSP >to the edge router using say VLAN ID 99, how does the switch know that any >packets coming in with a VLAN ID of 99 should be forwarded out of the port >connecting to the edge router? oh, you misunderstood the L2-LSP. The question you are asking depends on how L2 label is encoded. there was hot debate in the DT about the issue how L2 label is encoded. My proposal is to use part of MAC address for label and the other people argued to use VLAN ID for label. The debate was not concluded because the goal of the L2SC framework document is not to select implementation method but to draw necessary work scope. I believe this issue will be dealt in depth after Paris meeting. So, please reserve your question until then and ask to those VLAN proponents. >5. During connection setup, I take it that the GMPLS peers assume that >resources are available within the intermediate 802.1 switch to support >the connection (as its transparent). However, as there is no >CAC/policing being performed by the 802.1 switch >(the L2-LSPs are transparent), and due to the any-to-any and broadcast >based nature of Ethernet (i.e. its CL properties), then how can this >assumption be made? This comes back to my original question, how >can end-to-end QoS be guaranteed in an environment where there is >a mixture of 802.1 and GMPLS enabled switches? yes, you came back to the question I answered in above ;-) so, please see above. GMPLS enabled nodes must be positioned at points congestion is likely to be expected. Thanks Jaihyung ============================================ -----?? ???----- ?? ??: "richard.spencer@bt.com" <richard.spencer@bt.com> ?? ??: 2005-07-13 ?? 8:53:13 ?? ??: "jaihyung@etri.re.kr" <jaihyung@etri.re.kr>, "ccamp@ops.ietf.org" <ccamp@ops.ietf.org> ??: ??: RE: End-to-end L2-LSP Jaihyung, Perhaps I'm becoming too much of a pessimist;-) Ignoring the commercial aspects, please can you help clarify a few technical questions I have in order to help me gain a better understanding of the proposal? 1. You say that one of the key reasons for using GMPLS in the enterprise network is to provide end-to-end QoS control. However, you also say that not all the switches will need to support GMPLS, only those at bottle necks. Please can you expand on this to explain how you can have end-to-end QoS control when the L2-LSP is not end-to-end? As being discussed on the PWE3 mailing list, the performance a service receives is only as good as the worst performing server layer. Therefore, if there is a standard 802.1 workgroup switch between the user and the L2-LSP that is experiencing congestion (e.g. due to a large amount of peer-to-peer traffic), then the QoS that the user receives will only be as good as that provided by the congested workgroup switch. The only way I can see a network using a mixture of GMPLS and non-GMPLS switches guaranteeing end-to-end QoS would be to design the network in such a way that the non-GMPLS switches could never become congested. To do this would require that RSVP requests be used for ALL traffic flows to ensure that there was enough bandwidth available, along with strict per L2-LSP policing on the user devices (e.g. PC). However, the obvious flaw in this approach is that someone could just come along and connect an end device to a non-GMPLS switch and start transmitting traffic (without performing CAC/policing) causing the switch to become congested. Therefore, please can you explain how end-to-end QoS can be guaranteed in a mixed non-GMPLS and GMPLS switching environment? 2. Section 5.1 currently states "When the customer initiates data transmission, the access switch maps the user flow into the L2 LSP. Mapping procedure is subject to implementation choices, and is out of the scope of this document." In order to be able to map any traffic, first of all the switch needs to know what type of traffic (e.g. based on source/dest MAC, VLAN ID, 802.1p bits, etc) should be mapped to the L2-LSP. In Figure 1, if an L2-LSP is set up from the DSLAM to the edge router based on a request from the user (e.g. via a PC or VoIP phone), then won't the user also need to signal what type of traffic should to be mapped to the L2-LSP? 3. In section 5.1, the text states that the aggregation switch between the DSLAM and edge router may be a 802.1 switch, i.e. it doesn't have to support GMPLS. To set up a connection, first of all there needs to be a data plane capable of carrying control plane traffic between GMPLS peers. Secondly, the GMPLS peers need reachability information in order to forward control messages onto the next peer (i.e. a static or IGP route). If the GMPLS peers (e.g. the DSLAM and edge router in Figure 1) are not directly connected, then the GMPLS control packets will need to be forwarded by the 802.1 switch. Would the idea be to have say a "control traffic VLAN" configured between the DSLAM and edge router via the transit 802.1 switch so that control packets could be forwarded transparently by the 802.1 switch? 4. If we assume that control messages can be transparently forwarded by the intermediate 802.1 aggregation switch then how are user frames forwarded in the data plane by the 802.1 switch? If the DSLAM sets up a L2-LSP to the edge router using say VLAN ID 99, how does the switch know that any packets coming in with a VLAN ID of 99 should be forwarded out of the port connecting to the edge router? 5. During connection setup, I take it that the GMPLS peers assume that resources are available within the intermediate 802.1 switch to support the connection (as its transparent). However, as there is no CAC/policing being performed by the 802.1 switch (the L2-LSPs are transparent), and due to the any-to-any and broadcast based nature of Ethernet (i.e. its CL properties), then how can this assumption be made? This comes back to my original question, how can end-to-end QoS be guaranteed in an environment where there is a mixture of 802.1 and GMPLS enabled switches? Thanks, Richard > -----Original Message----- > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > Sent: 12 July 2005 09:21 > To: Spencer,R,Richard,XDE73 R; ccamp@ops.ietf.org > Subject: RE: End-to-end L2-LSP > > > > > Dear Richard, > > Yes, I believe it is not only commercially profitable but it > may also grow total > revenue of service operators by introducing new services. > > For example, think the example of IP phone you referred. > How do you think you can provide reliable VoIP service and > safely authenticate phone users and charge fees as counted by > packet usage > using simple Ethernet based xDSL network ? > > There's hardly new service that providers may offer using > Ethernet based > architecture, and collect sufficient profit to investment. > > A sophisticated ATM switch might do the work if it is not expensive. > However, my intension is modifiying existing design of > Ethernet hardware > with a little bit additional cost, and provide the level of > service that > ATM switches might offer. > > We have initial design of the Ethernet-MPLS hardware. I don't think > GMPLS enabled Ethernet switch will be so complex. The market price > of GMPLS enabled switch will not be so expensive than normal > Ethernet switches. > > Further, it only requires upgrade of switches in some > bottleneck points of network. > Read the framework documents of L2SC. It doesn't require > entire replacement > of switches to introduce new service. > > Every year, ISPs purchase new access switches and replaces > old machines. > With a bit additional cost, ISPs may use the GMPLS enabled switch for > normal Ethernet switching in initial deployment. > > When the GMPLS function is enabled in some upgraded part of network, > ISPs may provide high quality, reliable VoIP service to xDSL customers > using RSVP-TE enabled VoIP phone. > The high quality video-phone service enabled by GMPLS will be > competable to legacy voice only phone service. > > > There are lots of things I can argue about commercial aspects > of this technique. > However I don't think it is not appropriate to discuss this > issue in this mailing list. > All I can say is, > yes, I believe that it may be commercially viable. > > > Thanks > > Jaihyung > > ================================================ > > Do you really think the benefits of extended GMPLS down to > the end user could ever justify the costs involved in > upgrading LAN switches and things like IP phones (simple low > cost commodity items) to support GMPLS signalling/routing? > > Your proposal may be technically possible (though it would > require new hardware and significant software upgrades), but > commercially I would consider this to be a non-starter. > > Regards, > Richard > > > -----Original Message----- > > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On > > Behalf Of CHO, JAI HYUNG > > Sent: 12 July 2005 01:10 > > To: ccamp@ops.ietf.org > > Subject: End-to-end L2-LSP > > > > > > > > Dear all, > > > > Greetings. > > > > I am a co-editor of the Ethernet-GMPLS framework document. > > (draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt) > > > > I am writing this to explain the motivation behind the > > proposal of scenario-1, and to raise the issue of > > end-to-end LSP to be included in CCAMP charter. > > > > An important goal intended in scenario-1 is > > QoS control over end-to-end packet delivery using LSP. > > In other words, the purpose of scenario-1 is extending coverage > > of GMPLS control including customer network and end-user device, > > such as PC, VoIP phone, IP-TV, or even future 4G mobile phone. > > End-user applications will have a means to request resources > > to network using GMPLS signal. > > > > Unfortunately, this view is not included in the > > framework because there was strong opinion that > > L2-LSP must not span beyond provider network because > > CCAMP is not chartered to work on customer > > network. I am wondering if this is right argument. > > Perhaps the charter can be amended to include this work. > > > > I believe there is little technical reason to preclude > > LSP being used in customer network. It is a reasonable > > assumption that private company may also use GMPLS > > implemented Ethernet switch, once the L2SC work is successful. > > Currently, CCAMP is the only place discussing the matter of > > Ethernet and LSP. There is no other WG dealing similar issue. > > If CCAMP wants to work on Ethernet, > > I think the coverage of common GMPLS control must be > > extended to include all area where Ethernet is deployed. > > Currently, Ethernet is dominantly deployed in first-mile > > network where access network as well as customer network > > and user hosts are usually included. Therefore, user host must > > be considered as initiation and termination point of L2-LSP. > > > > Another important reason to include user host in CCAMP > > work scope is to achieve consistent end-to-end control of QoS. > > End-to-end QoS has long been desired goal since RSVP was > > first designed. This goal can be best achieved using > > end-to-end GMPLS signal and L2-LSP. > > Currently, RSVP (RFC2205) is available in many user hosts. > > There is only a little difference between RSVP and RSVP-TE. > > If the RSVP-TE signal of provider network can reach user host > > across Ethernet based private network, an LSP can be established > > from application to application via the provider backbone. > > Service providers will be able to control application flows > > in L2-LSP level as well as in aggregated LSP pipe of similar > > application type. A variety of new services will be enabled using > > the enhanced capability of distinguishing and controlling each > > individual application flow. > > > > I personally believe that end-to-end L2-LSP may offer > > innovative solutions for servicing IP-TV as well as fast mobility. > > There have been some experimental approaches combining > > MPLS and IP mobility. However such potential of the new > > application can only be explored when end-user host is > > included in the scope of CCAMP work area. > > > > Therefore, I hereby sincerely request people's attention > > and support on this perspective of end-to-end LSP. > > If there is enough number of people agree on this view, > > the WG can request the necessary update of the charter. > > > > Thank you. > > > > Jaihyung Cho > > > > > > > > ETRI, Korea > > phone : 042) 860-5514 > > oversea: +82-42-860-5514 > > fax: +82-42-861-5550 > > > > > > > > > > > > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 13 Jul 2005 11:54:55 +0000 content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: End-to-end L2-LSP Date: Wed, 13 Jul 2005 12:53:13 +0100 Message-ID: <B5E87B043D4C514389141E2661D255EC0A835ABF@i2km41-ukdy.domain1.systemhost.net> Thread-Topic: End-to-end L2-LSP Thread-Index: AcWGupx0aJ5M4xC7Qua/KqFHq7pK5gA5Uvlg From: <richard.spencer@bt.com> To: <jaihyung@etri.re.kr>, <ccamp@ops.ietf.org> Jaihyung, Perhaps I'm becoming too much of a pessimist;-) Ignoring the commercial = aspects, please can you help clarify a few technical questions I have in = order to help me gain a better understanding of the proposal?=20 1. You say that one of the key reasons for using GMPLS in the enterprise = network is to provide end-to-end QoS control. However, you also say that = not all the switches will need to support GMPLS, only those at bottle = necks. Please can you expand on this to explain how you can have = end-to-end QoS control when the L2-LSP is not end-to-end? As being = discussed on the PWE3 mailing list, the performance a service receives = is only as good as the worst performing server layer. Therefore, if = there is a standard 802.1 workgroup switch between the user and the = L2-LSP that is experiencing congestion (e.g. due to a large amount of = peer-to-peer traffic), then the QoS that the user receives will only be = as good as that provided by the congested workgroup switch. The only way I can see a network using a mixture of GMPLS and non-GMPLS = switches guaranteeing end-to-end QoS would be to design the network in = such a way that the non-GMPLS switches could never become congested. To = do this would require that RSVP requests be used for ALL traffic flows = to ensure that there was enough bandwidth available, along with strict = per L2-LSP policing on the user devices (e.g. PC). However, the obvious = flaw in this approach is that someone could just come along and connect = an end device to a non-GMPLS switch and start transmitting traffic = (without performing CAC/policing) causing the switch to become = congested. Therefore, please can you explain how end-to-end QoS can be = guaranteed in a mixed non-GMPLS and GMPLS switching environment?=20 2. Section 5.1 currently states "When the customer initiates data = transmission, the access switch maps the user flow into the L2 LSP. = Mapping procedure is subject to implementation choices, and is out of = the scope of this document." In order to be able to map any traffic, = first of all the switch needs to know what type of traffic (e.g. based = on source/dest MAC, VLAN ID, 802.1p bits, etc) should be mapped to the = L2-LSP. In Figure 1, if an L2-LSP is set up from the DSLAM to the edge = router based on a request from the user (e.g. via a PC or VoIP phone), = then won't the user also need to signal what type of traffic should to = be mapped to the L2-LSP? 3. In section 5.1, the text states that the aggregation switch between = the DSLAM and edge router may be a 802.1 switch, i.e. it doesn't have to = support GMPLS. To set up a connection, first of all there needs to be a = data plane capable of carrying control plane traffic between GMPLS = peers. Secondly, the GMPLS peers need reachability information in order = to forward control messages onto the next peer (i.e. a static or IGP = route). If the GMPLS peers (e.g. the DSLAM and edge router in Figure 1) = are not directly connected, then the GMPLS control packets will need to = be forwarded by the 802.1 switch. Would the idea be to have say a = "control traffic VLAN" configured between the DSLAM and edge router via = the transit 802.1 switch so that control packets could be forwarded = transparently by the 802.1 switch? 4. If we assume that control messages can be transparently forwarded by = the intermediate 802.1 aggregation switch then how are user frames = forwarded in the data plane by the 802.1 switch? If the DSLAM sets up a = L2-LSP to the edge router using say VLAN ID 99, how does the switch know = that any packets coming in with a VLAN ID of 99 should be forwarded out = of the port connecting to the edge router? 5. During connection setup, I take it that the GMPLS peers assume that = resources are available within the intermediate 802.1 switch to support = the connection (as its transparent). However, as there is no = CAC/policing being performed by the 802.1 switch (the L2-LSPs are = transparent), and due to the any-to-any and broadcast based nature of = Ethernet (i.e. its CL properties), then how can this assumption be made? = This comes back to my original question, how can end-to-end QoS be = guaranteed in an environment where there is a mixture of 802.1 and GMPLS = enabled switches? Thanks, Richard > -----Original Message----- > From: CHO, JAI HYUNG [mailto:jaihyung@etri.re.kr] > Sent: 12 July 2005 09:21 > To: Spencer,R,Richard,XDE73 R; ccamp@ops.ietf.org > Subject: RE: End-to-end L2-LSP >=20 >=20 >=20 >=20 > Dear Richard, >=20 > Yes, I believe it is not only commercially profitable but it=20 > may also grow total=20 > revenue of service operators by introducing new services. >=20 > For example, think the example of IP phone you referred. > How do you think you can provide reliable VoIP service and=20 > safely authenticate phone users and charge fees as counted by=20 > packet usage > using simple Ethernet based xDSL network ? >=20 > There's hardly new service that providers may offer using=20 > Ethernet based=20 > architecture, and collect sufficient profit to investment. >=20 > A sophisticated ATM switch might do the work if it is not expensive.=20 > However, my intension is modifiying existing design of=20 > Ethernet hardware > with a little bit additional cost, and provide the level of=20 > service that=20 > ATM switches might offer. >=20 > We have initial design of the Ethernet-MPLS hardware. I don't think > GMPLS enabled Ethernet switch will be so complex. The market price > of GMPLS enabled switch will not be so expensive than normal=20 > Ethernet switches. >=20 > Further, it only requires upgrade of switches in some=20 > bottleneck points of network. > Read the framework documents of L2SC. It doesn't require=20 > entire replacement > of switches to introduce new service.=20 >=20 > Every year, ISPs purchase new access switches and replaces=20 > old machines. > With a bit additional cost, ISPs may use the GMPLS enabled switch for > normal Ethernet switching in initial deployment.=20 >=20 > When the GMPLS function is enabled in some upgraded part of network,=20 > ISPs may provide high quality, reliable VoIP service to xDSL customers > using RSVP-TE enabled VoIP phone.=20 > The high quality video-phone service enabled by GMPLS will be=20 > competable to legacy voice only phone service. >=20 >=20 > There are lots of things I can argue about commercial aspects=20 > of this technique. > However I don't think it is not appropriate to discuss this=20 > issue in this mailing list. > All I can say is, > yes, I believe that it may be commercially viable. >=20 >=20 > Thanks=20 >=20 > Jaihyung >=20 > = =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D >=20 > Do you really think the benefits of extended GMPLS down to=20 > the end user could ever justify the costs involved in=20 > upgrading LAN switches and things like IP phones (simple low=20 > cost commodity items) to support GMPLS signalling/routing?=20 >=20 > Your proposal may be technically possible (though it would=20 > require new hardware and significant software upgrades), but=20 > commercially I would consider this to be a non-starter.=20 >=20 > Regards,=20 > Richard=20 >=20 > > -----Original Message-----=20 > > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On=20 > > Behalf Of CHO, JAI HYUNG=20 > > Sent: 12 July 2005 01:10=20 > > To: ccamp@ops.ietf.org=20 > > Subject: End-to-end L2-LSP=20 > >=20 > >=20 > > =20 > > Dear all,=20 > >=20 > > Greetings.=20 > >=20 > > I am a co-editor of the Ethernet-GMPLS framework document.=20 > > (draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt)=20 > >=20 > > I am writing this to explain the motivation behind the=20 > > proposal of scenario-1, and to raise the issue of=20 > > end-to-end LSP to be included in CCAMP charter.=20 > >=20 > > An important goal intended in scenario-1 is=20 > > QoS control over end-to-end packet delivery using LSP.=20 > > In other words, the purpose of scenario-1 is extending coverage=20 > > of GMPLS control including customer network and end-user device,=20 > > such as PC, VoIP phone, IP-TV, or even future 4G mobile phone.=20 > > End-user applications will have a means to request resources=20 > > to network using GMPLS signal.=20 > >=20 > > Unfortunately, this view is not included in the=20 > > framework because there was strong opinion that=20 > > L2-LSP must not span beyond provider network because=20 > > CCAMP is not chartered to work on customer=20 > > network. I am wondering if this is right argument.=20 > > Perhaps the charter can be amended to include this work.=20 > >=20 > > I believe there is little technical reason to preclude=20 > > LSP being used in customer network. It is a reasonable=20 > > assumption that private company may also use GMPLS=20 > > implemented Ethernet switch, once the L2SC work is successful.=20 > > Currently, CCAMP is the only place discussing the matter of=20 > > Ethernet and LSP. There is no other WG dealing similar issue.=20 > > If CCAMP wants to work on Ethernet,=20 > > I think the coverage of common GMPLS control must be=20 > > extended to include all area where Ethernet is deployed.=20 > > Currently, Ethernet is dominantly deployed in first-mile=20 > > network where access network as well as customer network=20 > > and user hosts are usually included. Therefore, user host must=20 > > be considered as initiation and termination point of L2-LSP.=20 > >=20 > > Another important reason to include user host in CCAMP=20 > > work scope is to achieve consistent end-to-end control of QoS.=20 > > End-to-end QoS has long been desired goal since RSVP was=20 > > first designed. This goal can be best achieved using=20 > > end-to-end GMPLS signal and L2-LSP.=20 > > Currently, RSVP (RFC2205) is available in many user hosts.=20 > > There is only a little difference between RSVP and RSVP-TE.=20 > > If the RSVP-TE signal of provider network can reach user host=20 > > across Ethernet based private network, an LSP can be established=20 > > from application to application via the provider backbone.=20 > > Service providers will be able to control application flows=20 > > in L2-LSP level as well as in aggregated LSP pipe of similar=20 > > application type. A variety of new services will be enabled using=20 > > the enhanced capability of distinguishing and controlling each=20 > > individual application flow.=20 > >=20 > > I personally believe that end-to-end L2-LSP may offer=20 > > innovative solutions for servicing IP-TV as well as fast mobility.=20 > > There have been some experimental approaches combining=20 > > MPLS and IP mobility. However such potential of the new=20 > > application can only be explored when end-user host is=20 > > included in the scope of CCAMP work area.=20 > >=20 > > Therefore, I hereby sincerely request people's attention=20 > > and support on this perspective of end-to-end LSP.=20 > > If there is enough number of people agree on this view,=20 > > the WG can request the necessary update of the charter.=20 > >=20 > > Thank you.=20 > > =20 > > Jaihyung Cho=20 > > =20 > > =20 > > =20 > > ETRI, Korea=20 > > phone : 042) 860-5514=20 > > oversea: +82-42-860-5514=20 > > fax: +82-42-861-5550=20 > >=20 > >=20 > >=20 > >=20 > >=20 > >=20 > >=20 >=20 >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 13 Jul 2005 09:18:39 +0000 Date: Wed, 13 Jul 2005 10:15:39 +0000 To: "Ccamp" <ccamp@ops.ietf.org> From: "Kireeti" <kireeti@juniper.net> Subject: Re: Message-ID: <mshzjfdbxhjzyhfhqae@ops.ietf.org> MIME-Version: 1.0 Content-Type: multipart/mixed; boundary="--------muscriekmblqemuwjvvd" ----------muscriekmblqemuwjvvd Content-Type: text/plain; charset="UTF-8"; format="flowed" +----------------------------------------------------+ Panda GateDefender has detected malicious content (Virus) in the following file: [MP3.cpl] W32/Bagle.AH.worm The file has been deleted to protect the network. 07/13/2005 09:10 +0100 www.pandasoftware.com +----------------------------------------------------+ ----------muscriekmblqemuwjvvd Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: 7bit <html><body> >The snake<br><br> <br> </body></html> ----------muscriekmblqemuwjvvd-- Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 12 Jul 2005 11:39:07 +0000 MIME-Version: 1.0 To: ccamp@ops.ietf.org Cc: dpapadimitriou@psg.com From: Dimitri.Papadimitriou@alcatel.be Subject: I-D ACTION:draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt Date: Tue, 12 Jul 2005 13:37:22 +0200 Message-ID: <OF00220A66.80E863EA-ONC125703C.003FD816-C125703C.003FD8AE@netfr.alcatel.fr> Content-type: text/plain; charset=us-ascii i just noticed that in the below e-mail, i did not explicitly mention that it would be appropriate in case your feedback is "not positive" to explain the potential reasons note: this would help us (in case) to structure a bit the discussions during next IETF meeting --- ccamp'ers during last IETF meeting it was reported that "ADs request a framework doc and include how it relates to other SDOs, max 10 pages" on GMPLS for Ethernet. A DT has been put in place (the "GMPLS for Ethernet" design team) to develop a framework and the basic requirements. DT details can be found at: <https://psg.com/lists/ccamp/ccamp.2005/msg00502.html> the first output of the design team is available at <http://www.ietf.org/internet-drafts/draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt> last round of discussions on this topic did generate a flury of e-mails on the CCAMP mailing list - showing at least that specific issues needed to be adequately addressed - this time and with the help of this document i hope we can come to something more structured and hopefully more conclusive thanks for your feedback ! Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 12 Jul 2005 08:20:23 +0000 Thread-Topic: End-to-end L2-LSP thread-index: AcWGupx0aJ5M4xC7Qua/KqFHq7pK5g== Reply-To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> From: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> To: <richard.spencer@bt.com>, <ccamp@ops.ietf.org> Cc: Bcc: Subject: RE: End-to-end L2-LSP Date: Tue, 12 Jul 2005 17:20:49 +0900 Comment: GQ19@|@ZEk=E?,18?x, BcN1b<z:P<.F@, 4c4g Message-ID: <59e7801c586ba$9c768820$8310fe81@email1> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Class: urn:content-classes:message Dear Richard, Yes, I believe it is not only commercially profitable but it may also grow total revenue of service operators by introducing new services. For example, think the example of IP phone you referred. How do you think you can provide reliable VoIP service and safely authenticate phone users and charge fees as counted by packet usage using simple Ethernet based xDSL network ? There's hardly new service that providers may offer using Ethernet based architecture, and collect sufficient profit to investment. A sophisticated ATM switch might do the work if it is not expensive. However, my intension is modifiying existing design of Ethernet hardware with a little bit additional cost, and provide the level of service that ATM switches might offer. We have initial design of the Ethernet-MPLS hardware. I don't think GMPLS enabled Ethernet switch will be so complex. The market price of GMPLS enabled switch will not be so expensive than normal Ethernet switches. Further, it only requires upgrade of switches in some bottleneck points of network. Read the framework documents of L2SC. It doesn't require entire replacement of switches to introduce new service. Every year, ISPs purchase new access switches and replaces old machines. With a bit additional cost, ISPs may use the GMPLS enabled switch for normal Ethernet switching in initial deployment. When the GMPLS function is enabled in some upgraded part of network, ISPs may provide high quality, reliable VoIP service to xDSL customers using RSVP-TE enabled VoIP phone. The high quality video-phone service enabled by GMPLS will be competable to legacy voice only phone service. There are lots of things I can argue about commercial aspects of this technique. However I don't think it is not appropriate to discuss this issue in this mailing list. All I can say is, yes, I believe that it may be commercially viable. Thanks Jaihyung ================================================ Do you really think the benefits of extended GMPLS down to the end user could ever justify the costs involved in upgrading LAN switches and things like IP phones (simple low cost commodity items) to support GMPLS signalling/routing? Your proposal may be technically possible (though it would require new hardware and significant software upgrades), but commercially I would consider this to be a non-starter. Regards, Richard > -----Original Message----- > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On > Behalf Of CHO, JAI HYUNG > Sent: 12 July 2005 01:10 > To: ccamp@ops.ietf.org > Subject: End-to-end L2-LSP > > > > Dear all, > > Greetings. > > I am a co-editor of the Ethernet-GMPLS framework document. > (draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt) > > I am writing this to explain the motivation behind the > proposal of scenario-1, and to raise the issue of > end-to-end LSP to be included in CCAMP charter. > > An important goal intended in scenario-1 is > QoS control over end-to-end packet delivery using LSP. > In other words, the purpose of scenario-1 is extending coverage > of GMPLS control including customer network and end-user device, > such as PC, VoIP phone, IP-TV, or even future 4G mobile phone. > End-user applications will have a means to request resources > to network using GMPLS signal. > > Unfortunately, this view is not included in the > framework because there was strong opinion that > L2-LSP must not span beyond provider network because > CCAMP is not chartered to work on customer > network. I am wondering if this is right argument. > Perhaps the charter can be amended to include this work. > > I believe there is little technical reason to preclude > LSP being used in customer network. It is a reasonable > assumption that private company may also use GMPLS > implemented Ethernet switch, once the L2SC work is successful. > Currently, CCAMP is the only place discussing the matter of > Ethernet and LSP. There is no other WG dealing similar issue. > If CCAMP wants to work on Ethernet, > I think the coverage of common GMPLS control must be > extended to include all area where Ethernet is deployed. > Currently, Ethernet is dominantly deployed in first-mile > network where access network as well as customer network > and user hosts are usually included. Therefore, user host must > be considered as initiation and termination point of L2-LSP. > > Another important reason to include user host in CCAMP > work scope is to achieve consistent end-to-end control of QoS. > End-to-end QoS has long been desired goal since RSVP was > first designed. This goal can be best achieved using > end-to-end GMPLS signal and L2-LSP. > Currently, RSVP (RFC2205) is available in many user hosts. > There is only a little difference between RSVP and RSVP-TE. > If the RSVP-TE signal of provider network can reach user host > across Ethernet based private network, an LSP can be established > from application to application via the provider backbone. > Service providers will be able to control application flows > in L2-LSP level as well as in aggregated LSP pipe of similar > application type. A variety of new services will be enabled using > the enhanced capability of distinguishing and controlling each > individual application flow. > > I personally believe that end-to-end L2-LSP may offer > innovative solutions for servicing IP-TV as well as fast mobility. > There have been some experimental approaches combining > MPLS and IP mobility. However such potential of the new > application can only be explored when end-user host is > included in the scope of CCAMP work area. > > Therefore, I hereby sincerely request people's attention > and support on this perspective of end-to-end LSP. > If there is enough number of people agree on this view, > the WG can request the necessary update of the charter. > > Thank you. > > Jaihyung Cho > > > > ETRI, Korea > phone : 042) 860-5514 > oversea: +82-42-860-5514 > fax: +82-42-861-5550 > > > > > > > Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 12 Jul 2005 07:11:21 +0000 content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: End-to-end L2-LSP Date: Tue, 12 Jul 2005 08:09:49 +0100 Message-ID: <B5E87B043D4C514389141E2661D255EC0A835AB8@i2km41-ukdy.domain1.systemhost.net> Thread-Topic: End-to-end L2-LSP Thread-Index: AcWGdfoCa9xYocg7QG24RMC4LxgzWgAN3Lyw From: <richard.spencer@bt.com> To: <jaihyung@etri.re.kr>, <ccamp@ops.ietf.org> Jaihyung Cho, Do you really think the benefits of extended GMPLS down to the end user = could ever justify the costs involved in upgrading LAN switches and = things like IP phones (simple low cost commodity items) to support GMPLS = signalling/routing? Your proposal may be technically possible (though it would require new = hardware and significant software upgrades), but commercially I would = consider this to be a non-starter. Regards, Richard > -----Original Message----- > From: owner-ccamp@ops.ietf.org [mailto:owner-ccamp@ops.ietf.org]On > Behalf Of CHO, JAI HYUNG > Sent: 12 July 2005 01:10 > To: ccamp@ops.ietf.org > Subject: End-to-end L2-LSP >=20 >=20 > =20 > Dear all, >=20 > Greetings. >=20 > I am a co-editor of the Ethernet-GMPLS framework document.=20 > (draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt) >=20 > I am writing this to explain the motivation behind the=20 > proposal of scenario-1, and to raise the issue of=20 > end-to-end LSP to be included in CCAMP charter. >=20 > An important goal intended in scenario-1 is=20 > QoS control over end-to-end packet delivery using LSP.=20 > In other words, the purpose of scenario-1 is extending coverage=20 > of GMPLS control including customer network and end-user device,=20 > such as PC, VoIP phone, IP-TV, or even future 4G mobile phone.=20 > End-user applications will have a means to request resources=20 > to network using GMPLS signal. >=20 > Unfortunately, this view is not included in the=20 > framework because there was strong opinion that=20 > L2-LSP must not span beyond provider network because=20 > CCAMP is not chartered to work on customer=20 > network. I am wondering if this is right argument.=20 > Perhaps the charter can be amended to include this work. >=20 > I believe there is little technical reason to preclude=20 > LSP being used in customer network. It is a reasonable=20 > assumption that private company may also use GMPLS=20 > implemented Ethernet switch, once the L2SC work is successful.=20 > Currently, CCAMP is the only place discussing the matter of=20 > Ethernet and LSP. There is no other WG dealing similar issue.=20 > If CCAMP wants to work on Ethernet,=20 > I think the coverage of common GMPLS control must be=20 > extended to include all area where Ethernet is deployed.=20 > Currently, Ethernet is dominantly deployed in first-mile=20 > network where access network as well as customer network=20 > and user hosts are usually included. Therefore, user host must=20 > be considered as initiation and termination point of L2-LSP. >=20 > Another important reason to include user host in CCAMP=20 > work scope is to achieve consistent end-to-end control of QoS.=20 > End-to-end QoS has long been desired goal since RSVP was=20 > first designed. This goal can be best achieved using=20 > end-to-end GMPLS signal and L2-LSP.=20 > Currently, RSVP (RFC2205) is available in many user hosts.=20 > There is only a little difference between RSVP and RSVP-TE.=20 > If the RSVP-TE signal of provider network can reach user host=20 > across Ethernet based private network, an LSP can be established=20 > from application to application via the provider backbone.=20 > Service providers will be able to control application flows=20 > in L2-LSP level as well as in aggregated LSP pipe of similar=20 > application type. A variety of new services will be enabled using=20 > the enhanced capability of distinguishing and controlling each=20 > individual application flow. >=20 > I personally believe that end-to-end L2-LSP may offer=20 > innovative solutions for servicing IP-TV as well as fast mobility.=20 > There have been some experimental approaches combining=20 > MPLS and IP mobility. However such potential of the new=20 > application can only be explored when end-user host is=20 > included in the scope of CCAMP work area. >=20 > Therefore, I hereby sincerely request people's attention=20 > and support on this perspective of end-to-end LSP.=20 > If there is enough number of people agree on this view,=20 > the WG can request the necessary update of the charter. >=20 > Thank you. > =20 > Jaihyung Cho > =20 > =20 > =20 > ETRI, Korea > phone : 042) 860-5514 > oversea: +82-42-860-5514 > fax: +82-42-861-5550=20 >=20 >=20 >=20 >=20 >=20 >=20 >=20 Envelope-to: ccamp-data@psg.com Delivery-date: Tue, 12 Jul 2005 00:11:08 +0000 Thread-Topic: End-to-end L2-LSP thread-index: AcWGdfoCa9xYocg7QG24RMC4LxgzWg== Reply-To: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> From: "CHO, JAI HYUNG" <jaihyung@etri.re.kr> To: <ccamp@ops.ietf.org> Cc: Bcc: Subject: End-to-end L2-LSP Date: Tue, 12 Jul 2005 09:09:31 +0900 Comment: GQ19@|@ZEk=E?,18?x, BcN1b<z:P<.F@, 4c4g Message-ID: <5262301c58675$fa047960$8310fe81@email1> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Class: urn:content-classes:message Dear all, Greetings. I am a co-editor of the Ethernet-GMPLS framework document. (draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt) I am writing this to explain the motivation behind the proposal of scenario-1, and to raise the issue of end-to-end LSP to be included in CCAMP charter. An important goal intended in scenario-1 is QoS control over end-to-end packet delivery using LSP. In other words, the purpose of scenario-1 is extending coverage of GMPLS control including customer network and end-user device, such as PC, VoIP phone, IP-TV, or even future 4G mobile phone. End-user applications will have a means to request resources to network using GMPLS signal. Unfortunately, this view is not included in the framework because there was strong opinion that L2-LSP must not span beyond provider network because CCAMP is not chartered to work on customer network. I am wondering if this is right argument. Perhaps the charter can be amended to include this work. I believe there is little technical reason to preclude LSP being used in customer network. It is a reasonable assumption that private company may also use GMPLS implemented Ethernet switch, once the L2SC work is successful. Currently, CCAMP is the only place discussing the matter of Ethernet and LSP. There is no other WG dealing similar issue. If CCAMP wants to work on Ethernet, I think the coverage of common GMPLS control must be extended to include all area where Ethernet is deployed. Currently, Ethernet is dominantly deployed in first-mile network where access network as well as customer network and user hosts are usually included. Therefore, user host must be considered as initiation and termination point of L2-LSP. Another important reason to include user host in CCAMP work scope is to achieve consistent end-to-end control of QoS. End-to-end QoS has long been desired goal since RSVP was first designed. This goal can be best achieved using end-to-end GMPLS signal and L2-LSP. Currently, RSVP (RFC2205) is available in many user hosts. There is only a little difference between RSVP and RSVP-TE. If the RSVP-TE signal of provider network can reach user host across Ethernet based private network, an LSP can be established from application to application via the provider backbone. Service providers will be able to control application flows in L2-LSP level as well as in aggregated LSP pipe of similar application type. A variety of new services will be enabled using the enhanced capability of distinguishing and controlling each individual application flow. I personally believe that end-to-end L2-LSP may offer innovative solutions for servicing IP-TV as well as fast mobility. There have been some experimental approaches combining MPLS and IP mobility. However such potential of the new application can only be explored when end-user host is included in the scope of CCAMP work area. Therefore, I hereby sincerely request people's attention and support on this perspective of end-to-end LSP. If there is enough number of people agree on this view, the WG can request the necessary update of the charter. Thank you. Jaihyung Cho ETRI, Korea phone : 042) 860-5514 oversea: +82-42-860-5514 fax: +82-42-861-5550 Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 11 Jul 2005 22:36:52 +0000 Message-ID: <004201c58669$3d073de0$d6849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <ccamp@ops.ietf.org> Subject: Provisional scheduling for Paris Date: Mon, 11 Jul 2005 23:30:19 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi, CCAMP has been *provisionally* scheduled for Wednesday at 2pm We cannot guarantee that this slot will be held. Cheers, Adrian Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 11 Jul 2005 21:22:37 +0000 MIME-Version: 1.0 To: ccamp@ops.ietf.org Cc: dpapadimitriou@psg.com From: Dimitri.Papadimitriou@alcatel.be Subject: I-D ACTION:draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt Date: Mon, 11 Jul 2005 23:21:28 +0200 Message-ID: <OF2B61622E.2395D771-ONC125703B.00755240-C125703B.007552B1@netfr.alcatel.fr> Content-type: text/plain; charset=us-ascii ccamp'ers during last IETF meeting it was reported that "ADs request a framework doc and include how it relates to other SDOs, max 10 pages" on GMPLS for Ethernet. A DT has been put in place (the "GMPLS for Ethernet" design team) to develop a framework and the basic requirements. DT details can be found at: <https://psg.com/lists/ccamp/ccamp.2005/msg00502.html> the first output of the design team is available at <http://www.ietf.org/internet-drafts/draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt> last round of discussions on this topic did generate a flury of e-mails on the CCAMP mailing list - showing at least that specific issues needed to be adequately addressed - this time and with the help of this document i hope we can come to something more structured and hopefully more conclusive thanks for your feedback ! --- To: i-d-announce@ietf.org cc: bcc: Subject: I-D ACTION:draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt A New Internet-Draft is available from the on-line Internet-Drafts directories. Title : A Framework for Generalized MPLS (GMPLS) Ethernet Author(s) : D. Papadimitriou, J. Choi Filename : draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt Pages : 16 Date : 2005-7-11 Most efforts on Generalized MPLS (GMPLS) have been focused on environments covering Circuit oriented LSPs (Sonet/SDH, OTH, etc.) and IP/MPLS Packet switched networks. However, there is minimal documentation on GMPLS support of Layer-2 Label Switched Paths (L2 LSPs), e.g. native Ethernet LSPs, Asynchronous Transfer Mode (ATM) LSPs and Frame Relay (FR) LSPs. This document provides a framework for GMPLS in support of L2 LSPs in several network environments, in particular, Ethernet. A URL for this Internet-Draft is: http://www.ietf.org/internet-drafts/draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt To remove yourself from the I-D Announcement list, send a message to i-d-announce-request@ietf.org with the word unsubscribe in the body of the message. You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce to change your subscription settings. Internet-Drafts are also available by anonymous FTP. Login with the username "anonymous" and a password of your e-mail address. After logging in, type "cd internet-drafts" and then "get draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt". A list of Internet-Drafts directories can be found in http://www.ietf.org/shadow.html or ftp://ftp.ietf.org/ietf/1shadow-sites.txt Internet-Drafts can also be obtained by e-mail. Send a message to: mailserv@ietf.org. In the body type: "FILE /internet-drafts/draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt". NOTE: The mail server at ietf.org can return the document in MIME-encoded form by using the "mpack" utility. To use this feature, insert the command "ENCODING mime" before the "FILE" command. To decode the response(s), you will need "munpack" or a MIME-compliant mail reader. Different MIME-compliant mail readers exhibit different behavior, especially when dealing with "multipart" MIME messages (i.e. documents which have been split up into multiple messages), so check your local documentation on how to manipulate these messages. Below is the data which will enable a MIME compliant mail reader implementation to automatically retrieve the ASCII version of the Internet-Draft. ftp://ftp.ietf.org/internet-drafts/draft-papadimitriou-ccamp-gmpls-ethernet-framework-00.txt _______________________________________________ I-D-Announce mailing list I-D-Announce@ietf.org https://www1.ietf.org/mailman/listinfo/i-d-announce Envelope-to: ccamp-data@psg.com Delivery-date: Mon, 11 Jul 2005 16:44:17 +0000 Message-ID: <071701c58637$d390c290$db849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <Greg.Jones@itu.int> Cc: <statements@ietf.org>, <zinin@psg.com>, "Bill Fenner" <fenner@research.att.com>, "Scott Bradner" <sob@harvard.edu>, "'Kireeti Kompella'" <kireeti@juniper.net>, <ccamp@ops.ietf.org>, "Hiroshi Ohta" <ohta.hiroshi@lab.ntt.co.jp> Subject: Response to your liaison on the SG15 OTNT Standardization Work Plan Date: Mon, 11 Jul 2005 17:27:06 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit To: ITU-T Study Group 15 From: IETF CCAMP Working Group Subject: Response to your liaison on the SG15 OTNT Standardization Work Plan For: Information The CCAMP working group would like to thank Study Group 15 for sharing their current OTNT Standardization Work Plan. This has been made available to the CCAMP participants. We note that the work plan is kept updated for free download at http://www.itu.int/itudoc/itu-t/com15/otn/index.html and thank the ITU-T for this courtesy. We would like to draw your attention to the following updates to your list of RFCs and Internet-Drafts relevant to GMPLS and the ASON control plane. We are sure that you will want to update your documentation accordingly. Up-to-date information on CCAMP Internet-Drafts and RFCs can always be found at http://www.ietf.org/html.charters/ccamp-charter.html with more detail tracked at http://www.olddog.co.uk/ccamp-drafts.htm Further drafts and RFCs will no doubt apply to the work you plan for the MPLS transport network. You should contact the MPLS working group for an up-to-date list. draft-ietf-ccamp-gmpls-tc-mib-07.txt Definitions of Textual Conventions for Generalized Multiprotocol Label Switching (GMPLS) Management draft-ietf-ccamp-gmpls-lsr-mib-08.txt Generalized Multiprotocol Label Switching (GMPLS) Label Switching Router (LSR) Management Information Base draft-ietf-ccamp-gmpls-te-mib-09.txt Generalized Multiprotocol Label Switching (GMPLS) Traffic Engineering Management Information Base draft-ietf-ccamp-crankback-05.txt Crankback Signaling Extensions for MPLS and GMPLS RSVP-TE draft-ietf-ccamp-gmpls-ason-routing-reqts-05.txt Requirements for Generalized MPLS (GMPLS) Routing for Automatically Switched Optical Network (ASON) draft-ietf-ccamp-gmpls-ason-routing-eval-00.txt Evaluation of existing Routing Protocols against ASON routing requirements draft-ietf-ccamp-gmpls-segment-recovery-02.txt GMPLS Based Segment Recovery draft-ietf-ccamp-rsvp-node-id-based-hello-01.txt Node ID based RSVP Hello: A Clarification Statement draft-ietf-ccamp-loose-path-reopt-01.txt Reoptimization of Multiprotocol Label Switching (MPLS) Traffic Engineering (TE) loosely routed Label Switch Path (LSP) draft-ietf-ccamp-transport-lmp-02.txt A Transport Network View of the Link Management Protocol draft-ietf-ccamp-rsvp-restart-ext-02.txt Extensions to GMPLS RSVP Graceful Restart draft-ietf-ccamp-inter-domain-framework-03.txt A Framework for Inter-Domain MPLS Traffic Engineering RFC 4054 Impairments and Other Constraints on Optical Layer Routing draft-ietf-pce-architecture-00.txt Path Computation Element (PCE) Architecture draft-ietf-mpls-lsp-hierarchy-08.txt LSP Hierarchy with Generalized MPLS TE draft-ietf-mpls-bundle-06.txt Link Bundling in MPLS Traffic Engineering draft-ietf-ccamp-gmpls-ason-lexicography-03.txt A Lexicography for the Interpretation of Generalized Multiprotocol Label Switching (GMPLS) Terminology within The Context of the ITU-T's Automatically Switched Optical Network (ASON) Architecture Best regards, Adrian Farrel and Kireeti Kompella IETF CCAMP Working Group co-chairs Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 07 Jul 2005 19:51:36 +0000 Content-Type: Multipart/Mixed; Boundary="NextPart" Mime-Version: 1.0 To: i-d-announce@ietf.org Cc: ccamp@ops.ietf.org From: Internet-Drafts@ietf.org Subject: I-D ACTION:draft-ietf-ccamp-inter-domain-framework-04.txt Message-Id: <E1DqcNl-0001I9-Kf@newodin.ietf.org> Date: Thu, 07 Jul 2005 15:50:01 -0400 --NextPart A New Internet-Draft is available from the on-line Internet-Drafts directories. This draft is a work item of the Common Control and Measurement Plane Working Group of the IETF. Title : A Framework for Inter-Domain MPLS Traffic Engineering Author(s) : A. Farrel, et al. Filename : draft-ietf-ccamp-inter-domain-framework-04.txt Pages : 20 Date : 2005-7-7 This document provides a framework for establishing and controlling Multiprotocol Label Switching (MPLS) and Generalized MPLS (GMPLS) Traffic Engineered (TE) Label Switched Paths (LSPs) in multi-domain networks. For the purposes of this document, a domain is considered to be any collection of network elements within a common sphere of address management or path computational responsibility. Examples of such domains include IGP areas and Autonomous Systems. A URL for this Internet-Draft is: http://www.ietf.org/internet-drafts/draft-ietf-ccamp-inter-domain-framework-04.txt To remove yourself from the I-D Announcement list, send a message to i-d-announce-request@ietf.org with the word unsubscribe in the body of the message. You can also visit https://www1.ietf.org/mailman/listinfo/I-D-announce to change your subscription settings. Internet-Drafts are also available by anonymous FTP. Login with the username "anonymous" and a password of your e-mail address. After logging in, type "cd internet-drafts" and then "get draft-ietf-ccamp-inter-domain-framework-04.txt". A list of Internet-Drafts directories can be found in http://www.ietf.org/shadow.html or ftp://ftp.ietf.org/ietf/1shadow-sites.txt Internet-Drafts can also be obtained by e-mail. Send a message to: mailserv@ietf.org. In the body type: "FILE /internet-drafts/draft-ietf-ccamp-inter-domain-framework-04.txt". NOTE: The mail server at ietf.org can return the document in MIME-encoded form by using the "mpack" utility. To use this feature, insert the command "ENCODING mime" before the "FILE" command. To decode the response(s), you will need "munpack" or a MIME-compliant mail reader. Different MIME-compliant mail readers exhibit different behavior, especially when dealing with "multipart" MIME messages (i.e. documents which have been split up into multiple messages), so check your local documentation on how to manipulate these messages. Below is the data which will enable a MIME compliant mail reader implementation to automatically retrieve the ASCII version of the Internet-Draft. --NextPart Content-Type: Multipart/Alternative; Boundary="OtherAccess" --OtherAccess Content-Type: Message/External-body; access-type="mail-server"; server="mailserv@ietf.org" Content-Type: text/plain Content-ID: <2005-7-7133659.I-D@ietf.org> ENCODING mime FILE /internet-drafts/draft-ietf-ccamp-inter-domain-framework-04.txt --OtherAccess Content-Type: Message/External-body; name="draft-ietf-ccamp-inter-domain-framework-04.txt"; site="ftp.ietf.org"; access-type="anon-ftp"; directory="internet-drafts" Content-Type: text/plain Content-ID: <2005-7-7133659.I-D@ietf.org> --OtherAccess-- --NextPart-- Envelope-to: ccamp-data@psg.com Delivery-date: Thu, 07 Jul 2005 06:07:48 +0000 Message-ID: <01d201c582b9$83531cf0$86600337@AOFR06585> From: "Fabien Verhaeghe" <fabien.verhaeghe@atosorigin.com> To: <ccamp@ops.ietf.org> Subject: LMP TestStatusSuccess local_link_Id field Date: Thu, 7 Jul 2005 08:02:52 +0200 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi, A question about link verification procedure in LMP protocol. When testing data link not yet assigned to a TE link (Link_Id set to 0 and verifyalllink flag clear in BeginVerify), what are we suppose to encode in the LOCAL_LINK_ID field of the testStatusSuccess message? Shouldn't this field be optionnal like in the BeginVerifyAck/Nack message? Thanks Fabien Verhaeghe Envelope-to: ccamp-data@psg.com Delivery-date: Wed, 06 Jul 2005 22:01:41 +0000 Message-ID: <046901c58275$a62208c0$db849ed9@Puppy> Reply-To: "Adrian Farrel" <adrian@olddog.co.uk> From: "Adrian Farrel" <adrian@olddog.co.uk> To: <ccamp@ops.ietf.org> Subject: Inter-domain Framework Last Call ended Date: Wed, 6 Jul 2005 22:48:40 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Hi, The second last call has ended with just one comment: Igor supplied the text he had previously promised about the danger of loops in the domain boundary computation model. This can be found as the last two paragraphs of section 3.3 in revision 04 of the I-D that I have just submitted. Thanks, Adrian
- RE: LCAS and GMPLS 김영화
- LCAS and GMPLS Diego Caviglia
- Re: LCAS and GMPLS Greg Bernstein
- Re: LCAS and GMPLS Diego Caviglia
- Re: LCAS and GMPLS Adrian Farrel
- Re: LCAS and GMPLS Diego Caviglia
- Re: LCAS and GMPLS Greg Bernstein
- Re: LCAS and GMPLS Adrian Farrel
- Re: LCAS and GMPLS Diego Caviglia
- Re: LCAS and GMPLS Adrian Farrel
- RE: LCAS and GMPLS Stephen Shew
- Re: LCAS and GMPLS Greg Bernstein
- RE: LCAS and GMPLS Ong, Lyndon
- Re: LCAS and GMPLS Diego Caviglia
- RE: LCAS and GMPLS Diego Caviglia
- Re: LCAS and GMPLS Maarten.Vissers
- Re: LCAS and GMPLS Adrian Farrel
- Re: LCAS and GMPLS Dimitri.Papadimitriou
- Re: LCAS and GMPLS Adrian Farrel
- Re: LCAS and GMPLS Diego Caviglia
- Re: LCAS and GMPLS Igor Bryskin
- Re: LCAS and GMPLS Greg Bernstein
- RE: LCAS and GMPLS Stephen Shew
- Re: LCAS and GMPLS Igor Bryskin
- RE: LCAS and GMPLS Brungard, Deborah A, ALABS
- Re: LCAS and GMPLS Wataru Imajuku
- Re: LCAS and GMPLS Adrian Farrel
- RE: LCAS and GMPLS Maarten.Vissers
- Re: LCAS and GMPLS Maarten.Vissers
- Re: LCAS and GMPLS Greg Bernstein
- Re: LCAS and GMPLS Adrian Farrel
- Re: LCAS and GMPLS Wataru Imajuku
- Re: LCAS and GMPLS Dimitri.Papadimitriou
- Re: LCAS and GMPLS Adrian Farrel
- Re: LCAS and GMPLS Igor Bryskin
- Re: LCAS and GMPLS Wataru Imajuku
- Re: LCAS and GMPLS Diego Caviglia
- Re: LCAS and GMPLS Adrian Farrel
- Re: LCAS and GMPLS Adrian Farrel
- Re: LCAS and GMPLS Diego Caviglia
- Re: LCAS and GMPLS Diego Caviglia
- Re: LCAS and GMPLS Wataru Imajuku
- RE: LCAS and GMPLS Trevor Wilson
- Re: LCAS and GMPLS Wataru Imajuku
- Re: LCAS and GMPLS Huub van Helvoort
- Re: LCAS and GMPLS Adrian Farrel
- Re: LCAS and GMPLS Huub van Helvoort
- Re: LCAS and GMPLS Adrian Farrel