Re: [Lsvr] Mail regarding draft-ietf-lsvr-bgp-spf

[Keyur Patel <[e-mail]>]

Hi Santosh,

We can add the use case in the applicability draft. This bit could be used when you have a RR/Controller based peering model in CLOS (where RR/Controllers are not in the forwarding path). You may want RR/Controller to announce its reachability using H-bit so that they are not being used as a transit.

Regards,
Keyur

From: Lsvr <[e-mail]> on behalf of Santosh P K <[e-mail]>
Date: Thursday, October 3, 2019 at 7:46 AM
To: "Acee Lindem (acee)" <[e-mail]>
Cc: Pushpasis Sarkar <[e-mail]>, "YADLAPALLI, CHAITANYA" <[e-mail]>, "[e-mail]" <[e-mail]>
Subject: Re: [Lsvr] Mail regarding draft-ietf-lsvr-bgp-spf
Resent-From: <[e-mail]>

All,
    If R-bit or H-bit functionality is being introduced then wanted to know should you not callout for use case here? Use-cases in applicability document does not cover this and would like to also understand how this can be used CLOS and non-CLOS/FAT-tree topology..


Thanks
Santosh P K

On Wed, Oct 2, 2019 at 6:10 AM Acee Lindem (acee) <[e-mail]<mailto:[e-mail]>> wrote:
Hi Pushpasis,

From: Pushpasis Sarkar <[e-mail]<mailto:[e-mail]>>
Date: Tuesday, October 1, 2019 at 3:34 PM
To: Acee Lindem <[e-mail]<mailto:[e-mail]>>
Cc: "YADLAPALLI, CHAITANYA" <[e-mail]<mailto:[e-mail]>>, "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>
Subject: Re: [Lsvr] Mail regarding draft-ietf-lsvr-bgp-spf

Hi Acee,

Just saw the latest version of the draft. I wanted to understand what is the exact difference between the values 1 and 2. Just to clarify my doubt let's consider a prefix P that is only originated by node N. Now what will be the reachability of prefix P in the two scenarios (first with SPF Status TLV value set to 1 vs with SPF Status TLV value set to 2). Will P be unreachable in both cases? My understanding is it should still be reachable when the value is set to 2.

If my understanding is correct, then perhaps we need more clarifications on the following text.. especially for the case there is no next link from this node.

"If the current Node NLRI attributes includes the SPF status

          TLV (Section 4.1.2) and the status indicates that the Node

          doesn't support transit, the next link for the current node is
          processed."

If the P is unreachable in the later case too (value set to 2), then I don't see what is the difference between using the values 1 and 2.

In this case, the current node is not unreachable as we’ve already taken it off the candidate list and processed the local prefixes. Optionally, the interface addresses on the current node have also been installed. At this point, we are simply not using any of the links in the SPF graph which will have the effect of preventing transit traffic.

Thanks,
Acee

Thanks
-Pushpasis






On Sun, Sep 29, 2019 at 12:15 PM Acee Lindem (acee) <[e-mail]<mailto:[e-mail]>> wrote:
After discussion with my co-authors and Pushpasis, we are planning on defining an SPF Status TLV for the Node Attribute NLRI analogous to the one defined for Links and Prefixes. However, for the Node Attribute TLV, the status would have an additional value indicating the node should not be used for transit traffic.

                          0 – Reserved
                          1 – Node unreachable with respect to BGP SPF
                          2 – Node does not support transit with respect to BGP SPF
                  3-254 – Undefined
                      255 – Reserved

Comments?

Thanks,
Acee


From: Lsvr <[e-mail]<mailto:[e-mail]>> on behalf of Acee Lindem <[e-mail]<mailto:[e-mail]>>
Date: Thursday, September 26, 2019 at 10:15 AM
To: Pushpasis Sarkar <[e-mail]<mailto:[e-mail]>>
Cc: "YADLAPALLI, CHAITANYA" <[e-mail]<mailto:[e-mail]>>, "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>
Subject: Re: [Lsvr] Mail regarding draft-ietf-lsvr-bgp-spf

Hi Pushpasis,
This OSPFv3 R Bit and IS-IS O bit are basically the same functionality. The node is not used for transit but is used for local prefixes.
Thanks,
Acee

From: Pushpasis Sarkar <[e-mail]<mailto:[e-mail]>>
Date: Thursday, September 26, 2019 at 2:53 AM
To: Acee Lindem <[e-mail]<mailto:[e-mail]>>
Cc: "YADLAPALLI, CHAITANYA" <[e-mail]<mailto:[e-mail]>>, "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>
Subject: Re: [Lsvr] Mail regarding draft-ietf-lsvr-bgp-spf

Hi Chaitanya and Acee,

How about the 'O' bit in Node-Flag-Bits TLV defined in RFC 7752 section 3.3.1.1? I suppose the node can set the 'O' bit when it wants to take itself out from all transit paths. I know the 'O' bit is more related to the scenario when ISIS topology is being exported in BGP-LS. But I suppose we can use that for BGP-LS-SPF as well.

Thanks
-Pushpasis

On Tue, Sep 24, 2019 at 8:35 PM Acee Lindem (acee) <[e-mail]<mailto:[e-mail]>> wrote:
Hi Chaitanya,
I think this is a good idea and will discuss with my co-authors.
Thanks,
Acee

From: "YADLAPALLI, CHAITANYA" <[e-mail]<mailto:[e-mail]>>
Date: Tuesday, September 24, 2019 at 11:02 AM
To: Acee Lindem <[e-mail]<mailto:[e-mail]>>, "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>
Subject: RE: [Lsvr] Mail regarding draft-ietf-lsvr-bgp-spf

Correct like a R-Bit.

I have read this draft and I support it.

Thanks,
Chaitanya


This communication may contain information that is privileged, or confidential.. If you are not the intended recipient, please note that any dissemination, distribution or copying of this communication is strictly prohibited.  Anyone who receives this message in error should notify the sender immediately by telephone or by return e-mail and delete it from his or her computer.



From: Acee Lindem (acee) <[e-mail]<mailto:[e-mail]>>
Sent: Tuesday, September 24, 2019 10:42 AM
To: YADLAPALLI, CHAITANYA <[e-mail]<mailto:[e-mail]>>; [e-mail]<mailto:[e-mail]>
Subject: Re: [Lsvr] Mail regarding draft-ietf-lsvr-bgp-spf

Hi Chaitanya,

Exactly what do you mean by node cost out and what use case are you trying to satisfy. If a node wants to remove itself from the topology, it can simply withdraw its link NLRI. However, are you looking for a mechanism similar to the OSPFv3 R-Bit as a Node NLRI SPF Attribute?

   R-bit
      This bit (the `Router' bit) indicates whether the originator is an
      active router.  If the router bit is clear, then routes that
      transit the advertising node cannot be computed.  Clearing the
      router bit would be appropriate for a multi-homed host that wants
      to participate in routing, but does not want to forward non-
      locally addressed packets.

Thanks,
Acee




From: Lsvr <[e-mail]<mailto:[e-mail]>> on behalf of "YADLAPALLI, CHAITANYA" <[e-mail]<mailto:[e-mail]>>
Date: Tuesday, September 24, 2019 at 10:31 AM
To: "[e-mail]<mailto:[e-mail]>" <[e-mail]<mailto:[e-mail]>>
Subject: [Lsvr] Mail regarding draft-ietf-lsvr-bgp-spf

Hi Authors,
The draft does not explicitly call out mechanisms for node cost out. It would be good to call out mechanisms to cost out a node explicitly.

Thanks,
Chaitanya


Chaitanya Yadlapalli
Network Infrastructure And Services

AT&T Services, Inc.
200 S Laurel Ave, Middletown, NJ 07722
o  732.420.7977  |  [e-mail]<mailto:[e-mail]>

This communication may contain information that is privileged, or confidential.. If you are not the intended recipient, please note that any dissemination, distribution or copying of this communication is strictly prohibited.  Anyone who receives this message in error should notify the sender immediately by telephone or by return e-mail and delete it from his or her computer.


_______________________________________________
Lsvr mailing list
[e-mail]<mailto:[e-mail]>
https://www.ietf.org/mailman/listinfo/lsvr
_______________________________________________
Lsvr mailing list
[e-mail]<mailto:[e-mail]>
https://www.ietf.org/mailman/listinfo/lsvr