Re: [i2rs] Éric Vyncke's No Objection on draft-ietf-i2rs-yang-l2-network-topology-14: (with COMMENT)

Susan Hares <> Thu, 09 July 2020 14:16 UTC

Return-Path: <>
Received: from localhost (localhost []) by (Postfix) with ESMTP id 42A363A0B8D; Thu, 9 Jul 2020 07:16:36 -0700 (PDT)
X-Virus-Scanned: amavisd-new at
X-Spam-Flag: NO
X-Spam-Score: 0.948
X-Spam-Status: No, score=0.948 tagged_above=-999 required=5 tests=[BAYES_00=-1.9, DOS_OUTLOOK_TO_MX=2.845, SPF_HELO_NONE=0.001, SPF_NONE=0.001, URIBL_BLOCKED=0.001] autolearn=no autolearn_force=no
Received: from ([]) by localhost ( []) (amavisd-new, port 10024) with ESMTP id ij2CA1Sh7FYz; Thu, 9 Jul 2020 07:16:33 -0700 (PDT)
Received: from ( []) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by (Postfix) with ESMTPS id 023FE3A0B87; Thu, 9 Jul 2020 07:16:31 -0700 (PDT)
X-Default-Received-SPF: pass (skip=loggedin (res=PASS)) x-ip-name=;
From: Susan Hares <>
To: "'Eric Vyncke (evyncke)'" <>, 'Qin Wu' <>, "'Eric Vyncke (evyncke)'" <>, 'The IESG' <>, "'Vigoureux, Martin (Nokia - FR/Paris-Saclay)'" <>
References: <> <> <008901d655cd$220cd010$66267030$> <>
In-Reply-To: <>
Date: Thu, 09 Jul 2020 10:16:03 -0400
Message-ID: <003901d655fb$82736080$875a2180$>
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="----=_NextPart_000_003A_01D655D9.FB640A70"
X-Mailer: Microsoft Outlook 14.0
Thread-Index: AQJKQctaPp+7yeQB+bAFp71mi4QzqwNcEY0sANU8y3gCyT/nm6ffmWTg
Content-Language: en-us
X-Antivirus: AVG (VPS 200709-4, 07/09/2020), Outbound message
X-Antivirus-Status: Not-Tested
Archived-At: <>
Subject: Re: [i2rs] Éric Vyncke's No Objection on draft-ietf-i2rs-yang-l2-network-topology-14: (with COMMENT)
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: "Interface to The Internet Routing System \(IRS\)" <>
List-Unsubscribe: <>, <>
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>, <>
X-List-Received-Date: Thu, 09 Jul 2020 14:16:36 -0000


On the I2RS reference,  how brief do you want the reference to be?  

One liner: The short description is - these yang topology models represent networks and topologies for use with network management.    

For the  3 paragraph, See these excerpts from RFC8345.  I've attached an article I sent for publication in the IEEE journal from IETF WG  last September which is 2 paragraph. 

Has Glenn allowed me to send a presentation on the work? 


However the abstract states:

   This document defines an abstract (generic, or base) YANG data model
   for network/service topologies and inventories.  The data model
   serves as a base model that is augmented with technology-specific
   details in other, more specific topology and inventory data models.

The first paragraph of the introduction states:

   This document introduces an abstract (base) YANG [RFC7950] data model
   [RFC3444] to represent networks and topologies.  The data model is
   divided into two parts: The first part of the data model defines a
   network data model that enables the definition of network
   hierarchies, or network stacks (i.e., networks that are layered on
   top of each other) and maintenance of an inventory of nodes contained
   in a network.  The second part of the data model augments the basic
   network data model with information to describe topology information.

   Specifically, it adds the concepts of "links" and
   "termination points" to describe how nodes in a network are connected
   to each other.  Moreover, the data model introduces vertical layering
   relationships between networks that can be augmented to cover both
   network inventories and network/service topologies."

-----Original Message-----
From: Eric Vyncke (evyncke) [] 
Sent: Thursday, July 9, 2020 5:42 AM
To: Susan Hares; 'Qin Wu'; 'Eric Vyncke (evyncke)'; 'The IESG'; Vigoureux, Martin (Nokia - FR/Paris-Saclay)
Subject: Re: [i2rs] Éric Vyncke's No Objection on draft-ietf-i2rs-yang-l2-network-topology-14: (with COMMENT)


Thank you for having requested IEEE feedback and for forwarding the info to all the recipients.

You suggestion to delay the IESG review of this draft until deep IEEE review end of July sounds sensible to me. Up to the responsible AD, Martin, to decide of course.

Finally, if there is a short document describing " the virtual models of I2RS for IETF NM", then I will welcome its reference (I admit my ignorance there).



-----Original Message-----
From: Susan Hares <>
Date: Thursday, 9 July 2020 at 10:44
To: Eric Vyncke <>, 'Qin Wu' <>, "'Eric Vyncke (evyncke)'" <>, 'The IESG' <>
Cc: "" <>, "" <>, "" <>
Subject: RE: [i2rs]  Éric Vyncke's No Objection on draft-ietf-i2rs-yang-l2-network-topology-14: (with COMMENT)

    Eric and Martin: 

    I have received email back from Glenn Parsons regarding the sys-mac-address - who is in charge of the IEEE Yang modules for TSN PAR (time sensitive network). 

    The email started with a rant regarding the lack of coordination over this model occurring at the last minute, and 
    a request to delay final review of this model until the july 20th IEEE 802.1 meeting.   Since the first WG publication of draft-ietf-i2rs-yang-l2-network-topology-00.txt  in 2015 predates the IEEE 802.1 Yang models for the bridge, it seems any coordination for the last 5 years regarding this model has gone to /dev/null.  

    In order to keep harmonious working between IEEE 802.1 and IETF for this I2RS model,  it may be necessary for this review and delay to occur.   Qin, his co-authors, and I have repeatedly tried to make sure this coordination has occurred.  We have received "acks" from internal IETF sources that were supposedly knowledgeable. 

    I recommend that you and Martin communicate with Glenn Parsons in order to foster a good review.  Since the virtual models of I2RS for IETF NM are not well understood in the IESG,  you may need my help or members of the I2RS WG to help explain the basic concept of the I2RS Virtual models.  

    By the way, Glenn did not provide any new data on the sys-mac-address concept.  


    -----Original Message-----
    From: Eric Vyncke (evyncke) [] 
    Sent: Thursday, July 9, 2020 4:33 AM
    To: Qin Wu; Susan Hares; 'Eric Vyncke (evyncke)'; 'The IESG'
    Subject: Re: [i2rs] Éric Vyncke's No Objection on draft-ietf-i2rs-yang-l2-network-topology-14: (with COMMENT)


    Honestly, I do not mind too much about "system-mac-address" (and the LLDP was just an example) but per symmetry with other leaves, adding a 'management-mac-address' could be useful. But, again, I do not mind too much.

    About the link rate... whether it is Sigfox or other data link does not really mater but I keep wondering why not having a minimum lower than 10 kbps is such a problem for the authors/WG... Why not simply augment the number of decimals on this leaf ? It is a decimal64 and by looking at RFC 6020 section 9.3.4 is seems that even using 3 decimals for Mbps (or simply using rate in kbps) will not limit the upper rate. 

    In short, I am still really puzzled why this lower bound on rate is in the model.


    -----Original Message-----
    From: iesg <> on behalf of Qin Wu <>
    Date: Thursday, 9 July 2020 at 08:40
    To: Susan Hares <>, "'Eric Vyncke (evyncke)'" <>, 'The IESG' <>
    Cc: "" <>, "" <>, "" <>
    Subject: RE: [i2rs]  Éric Vyncke's No Objection on draft-ietf-i2rs-yang-l2-network-topology-14: (with COMMENT)

        Hi, Eric:
        Sorry for late reply, your question on wireless is difficult to me,:-)
        Regarding Key editorial additions:
        Sure, we should replace CRC with RCS and fix inconsistency on L2 vs Layer 2.
        Regarding key technical changes:
        I think system-mac-address will be used for sending LLDP frames, we don't need another management MAC address, maybe we use different terminology to talk about the same thing.
        As for speeds for IoT and wireless,
        I am using figure 1 in draft-ietf-lpwan-schc-over-sigfox as an example, I still believe we should not model the link between Device and Sigfox BS (RG) in the Layer 2 network topology, bear in mind, L2 topo focuses network side topology.
        see reference
        Do you think the link speed of wired line between Sigfox BS and Sigfox network should be less than 10 kbps? No?
        Anyway if the wireless can benefit from L2 topology, I hope it could take augmentation approach.

        发件人: Susan Hares [] 
        发送时间: 2020年7月7日 22:20
        收件人: 'Eric Vyncke (evyncke)' <>; Qin Wu <>; 'The IESG' <>
        主题: RE: [i2rs] Éric Vyncke's No Objection on draft-ietf-i2rs-yang-l2-network-topology-14: (with COMMENT)


        Another round of comments are below.   

        As to the  IEEE review, this goes through the IETF-IEEE liaison.     I'm happy to have this yang model reviewed by the IEEE liaison is in 2020.   Qin's review occurred late 2019.  I'm unclear who the 2020 liaison is from IETF to IEEE.  Asking specifically about the Ethernet is useful.  Martin/Alvaro/Deborah may know who to ask. 

        FCS is a more generic term.  I'm sure that Qin will agree to utilize this term. 
        Qin response may come in 2-3 hours. 

        More comments below to your questions.   [Sue2] indicates the next round of answers. 

        Key editorial additions: 
        1) FCS replacing CRC
        2) Inconsistent use of L2 versus Layer 2. 

        Key technical changes: 
        1) management-mac-address - used for sending LLDP frames. 
        2) Speeds for IoT and the Wireless - could we get an expert from Internet on IoT directorate, and Any suggestions on 802.11 person?  
        I have managed a team that wrote code for early 802.11 code in L2 switches, but I have not done that since 2010.   If you do not have a key person, Qin or Donald Eastlake or other 802.11 people can help me. 

        Thanks for going through the model carefully. 

        Susan Hares 

        -----Original Message-----
        From: i2rs [] On Behalf Of Eric Vyncke (evyncke)
        Sent: Tuesday, July 7, 2020 8:39 AM
        To: Qin Wu; Susan Hares; 'The IESG'
        Subject: Re: [i2rs] Éric Vyncke's No Objection on draft-ietf-i2rs-yang-l2-network-topology-14: (with COMMENT)

        Hello Qin

        Long time not talked ☹

        Satisfied to read that IEEE has reviewed the document as this was a major concern of mine. If IEEE reviewers do not mind about using the old 'ethernet' word rather than 'ieee802', then I can only respect your and their choice.

        About the Ethernet frame length, your suggested text is improved but you may want to also specify for IEEE 802.11 whether the 'frame control' and other fields are also counted. I know, L2 is a pain __ Also (and my bad), please use FCS rather than CRC.

        About the rate, AFAIK, SigFox is below 1 kbps and LoRa is just a little faster. So, why not covering those rates ? Except if there is another limitation that I am not aware of

        Sincerely hope it helps


        -----Original Message-----
        From: Qin Wu <>
        Date: Tuesday, 7 July 2020 at 14:22
        To: Eric Vyncke <>, Susan Hares <>, 'The IESG' <>
        Cc: "" <>, "" <>, "" <>
        Subject: RE: [i2rs] Éric Vyncke's No Objection on draft-ietf-i2rs-yang-l2-network-topology-14: (with COMMENT)

            Hi, Eric:
            Thanks Sue for helping clarification. See followup comments inline below.
            发件人: Eric Vyncke (evyncke) [] 
            发送时间: 2020年7月7日 15:36
            收件人: Susan Hares <>; 'The IESG' <>
            主题: Re: [i2rs] Éric Vyncke's No Objection on draft-ietf-i2rs-yang-l2-network-topology-14: (with COMMENT)

            Hello Sue,

            Thank you for your reply; no surprise, you take your document shepherd role seriously.

            About the YANG validations, I was indeed suspecting something about the tool itself rather with the document: thank you for clearing my concerns. Your statement obviously clears my semi DISCUSS.

            Please find below some more comments prefixed with EV> (mainly about IEEE and indeed my lack of knowledge about the IETF topology model)

            All the best,


            -----Original Message-----
            From: Susan Hares <>
            Date: Monday, 6 July 2020 at 22:40
            To: Eric Vyncke <>, 'The IESG' <>
            Cc: "" <>, "" <>, "" <>
            Subject: RE: [i2rs] Éric Vyncke's No Objection on draft-ietf-i2rs-yang-l2-network-topology-14: (with COMMENT)


                <shepherd hat on> 

                Thank you for being concerned about the errors reported by the tools.  These errors have been growing with each revision of the yang tools kit without any change to the Yang.  

                Yang doctors have OKed the draft.  We have asked the Yang Doctors to address this issue and the ADs.    It is hard to fix ghost bugs.   The authors will fix anything that is a real bug rather than a ghost bug. 

                Other comments are below. 
                 I will answer quickly because Authors are in China.   They may correct me - as it is there document. 


                -----Original Message-----
                From: i2rs [] On Behalf Of Éric Vyncke via Datatracker
                Sent: Friday, July 3, 2020 6:21 AM
                To: The IESG
                Subject: [i2rs] Éric Vyncke's No Objection on draft-ietf-i2rs-yang-l2-network-topology-14: (with COMMENT)

                Éric Vyncke has entered the following ballot position for
                draft-ietf-i2rs-yang-l2-network-topology-14: No Objection

                When responding, please keep the subject line intact and reply to all email addresses included in the To and CC lines. (Feel free to cut this introductory paragraph, however.)

                Please refer to
                for more information about IESG DISCUSS and COMMENT positions.

                The document, along with other ballot positions, can be found here:


                Thank you for the work put into this document.

                Please find below a couple on non-blocking COMMENTs (and I would appreciate a reply to each of my COMMENTs because for some of them I was close to ballot a DISCUSS).

                I hope that this helps to improve the document,



                == DISCUSS ==

                == COMMENTS ==

                Generic: is there a reason why the YANG validation results in 19 errors and 4 warnings? Sue Harres (the shepherd) mentions in her write-up that 9 errors are linked to missing IEEE but what about the 10 remaining errors ?

                Has there been a review by IEEE of this YANG model? While the shepherd is extensive and detailed, there is no mention of a coordination with IEEE.

            EV> I still would like to have a confirmation that IEEE has also reviewed this YANG module.
            [Qin]:All IEEE references has been reviewed by IEEE folks during WGLC and cross posted to the netmod mailing list for IEEE folks to chime in by Sue and Rob help relay the discussion between IETF and IEEE.
            This is one of links related to IEEE part, can't remember other links.
            Hope this address your comment.
                -- Section 3 --

                  "The Layer 2 (L2) network topology YANG module is designed to be
                   generic and applicable to Layer 2 networks built with different L2
                Is this statement correct? What about LoraWAN, Sigfox, and other LP-WAN technologies? Or technologies that may be using different MTU sizes on each direction? or having more parameters than this (such as being NBMA that should be captured).

                [Sue:   It is generic abstract representation just like the network topology is a generic abstract representation. 
                Abstract models can be augmented with specific details.  See TEAS augmentation of the abstract network topology model.   As a chair, I delayed this model until it was implemented on a set of topologies.  

                If we can get proposals from running code or people plan running code for yang models,  I will glad work on augmentations for any of these topologies.  The only caveat is if my AD agrees to this work. 

            EV> Ack. I am trusting you about the YANG augment facility

                Should "sys-mac-address? " rather be "management-mac-address? "

                [Sue: My understanding from building switches is that the sys-mac-address may be different than the 
                management-mac-address.   The system processor may have a different chip that the management processor. 
                Therefore, based on the plan implementations - I would have to object to a change. 

            EV> Good point. But, then should there be  'management-mac-address' added to the model? I.e., used as a source for LLDP frames ?

            [Qin]: In most cases I have seen, only management IP address is used as a source for LLDP frames. Do you have other case in mind?
        [Sue2]:   Do you wish us to add a management IP address for LLDP frames?  Most switches do support LLDP. 

                I must admit that I am not familiar with the ietf-topology YANG model, so, the following COMMENTs can be plain wrong :-( ... It is unclear to me the difference between 'node' and 'termination-point'. If not defined in the ietf-topology, then please define before first use (I had to read the YANG module to understand).

                [Sue:  This shows a misunderstand of the base topology model.  May I politely suggest that I will cover this offline with you or that you discuss this with Alvaro or Martin or Rob Wilton [NM/OPS]?  ]

            EV> I had confessed my ignorance before __ so bear with me __

                Why is 'ethernet' used rather than 'ieee802', notably to cover IEEE 802.11 ?
                [Sue: The  IEEE 802.1 is now linking multiple IEEE 802.11 segments using IEEE 802.11 segments.  Donald Eastlake was a part of that process in 802.11 specification.    In my limited understanding, it is correct to use ieee802 as ieee802.1 is provides the linkage.   

                The use of the term Ethernet was vetted by several discussions regarding the 802.1 models this links to.   If the IEEE and the IETF suggest something else  (in their harmonization of this model between groups), the authors will change to.  We've put in the draft what all parties agreed to (i2rs participants and IETF yang experts with IEEE knowledge) suggested.   If you have a concrete alternate proposal the same groups agree to, we will adjust.  We've been trying to do the best to stay aligned to the best common practices of both groups. 

            EV> my point was that the current use of 'ethernet' appears outdated and I would have preferred 'ieee802' (covering a lot of technologies including the good old Ethernet). Nothing blocking on my side, but, I would like that authors/WG have a 2nd thought on this name.
            [Qin]: I understand what you ask is allow ieee802.11 technology series specific parameters can be augmented into L2 topology model. The current model doesn't prevent you from doing that, e..g,
                     choice l2-termination-point-type {
                         "Indicates termination-point type
                          specific attributes.";
                       case ethernet {
                         leaf mac-address {
                           type yang:mac-address;
                             "Interface MAC address.";
                      Case 802.11 {
                   If we rename Ethernet as 'ieee802', do you think all the parameters under etherent case can be reused by IEEE802.11?
                While most termination points have a single MAC address, are we sure that no termination point will ever have more than one MAC address ?
                [Sue: Multiple termination points are possible in the basic topology model.  Again, I suggest a review of the basic topology model concepts may be useful.  ] 

                'rate' leaf is in Mbps and with 2 decimals, i.e., the lowest rate is 10 kbps and this is already higher than some layer-2 links. Any reason to ignore lower rate links ?
                [Sue: Implementation reports gave us the lowest current rate of 10 kbps.  
                If you believe that this lower link boxes will implement yang models, we are glad to adjust this rate.  
                L2 model are difficult because there are so many things out there.]  

            EV> agreed on the diversity of L2... Hence, why limiting the rate to a multiple of 10 kbps? Even if I do not envision IoT devices on low rate link being provisioned by NETCONF/YANG, there could be NMS/OPS/??? systems using this YANG module to model the actual topology for their own purposes.
            [Qin]: Do we need to model IoT device and wireless link as part of network topology? Suppose wireless link should be modelled, data rate is usually in Mbps, see

        [Sue2]:  IoT devices and wireless links are part of the 802.1 model.  As shepherd/chair I had resisted adding 
        these features because we did not have an implementation.   I suggest allowing different data rates at least. 
        Adding the IoT and wireless portions of the model without implementations concerned me. 
        However, Wireless links are part of all switches.  Many switches are deployed for sensors. 

        Eric - if you could provide us a link to an expert in either of these features, we can add this intelligently. 
        Otherwise, it would be better to allow a short draft to augment this model rather than be incorrect In the modeling
        [Sue2 ends] 

                -- Section 4 --
                "leaf maximum-frame-size" please specify whether Ethernet pre-amble, inter-frame gap, and CRC should be included. The text for Ethernet and for PPP are identical, so, why repeating it ?

                [Sue:  I will let the authors take a first pass on why they left out Ethernet pre-amble, inter-frame gap, and CRC.  After they comment, I will provide a comment on the repetition. ] 

            [Qin]: Based on our analysis, maximum-frame-size should include CRC but not include preamble, inter-frame gap,
            we can clarify this in the text. Based on your comment, how about the following change:
                     leaf maximum-frame-size {
                       type uint32;
                         "Maximum L2 frame size that can be transported on 
            			 the data link layer, e.g. Ethernet frame. 
            			 If L2 frame is an Ethernet frame, the maximum frame size
            			 should include CRC and should not include preamble,Start 
            			 Frame Delimiter and Inter frame gap.";
                == NITS ==

                Sometimes 'L2' is used, sometimes 'Layer 2' is used. Not very consistent ;-) I am not an English speaker, but, I believe 'Layer 2 topology' should be written
                'layer-2 topology'

                [Good catch - Sometimes, the obvious aids to clarity get missed after you read a document long enough. ] 
            [Qin]:Thanks Eric.

                i2rs mailing list

        i2rs mailing list