Re: [Idr] draft-ietf-idr-bgp-extended-messages-12 WG LC (5/24 to 6/7)

[Ignas Bagdonas <[e-mail]>]

I have read the draft and support the publication.

Some assorted comments:

- s/RFC4221/RFC4271, and remove reference of RFC4221.

- Section 4, third paragraph talks about the propagation path. How would 
such a path be known in advance of the propagation? This seems contrary 
to a normal multipoint flooding mode of BGP operation.

- Error handling in the environment with old style and extended messages 
capable peers. There could be several scenarios advertising received 
extended messages to an old style peer:
-- Single NLRI itself is longer than 4096 (opaque key-value AFI as an 
example). It is not possible to send it to an old style peer at all, 
should be dropped and logged.
-- Single path attribute is longer than 4096. Removing such attribute 
would leave no place where to put partial indication. Would see dropping 
and logging to stay consistent with the NLRI case being a practical 
approach.
-- The sum of path attributes is above 4096, while each path attribute 
is shorter than 4096. This seems to be the least deterministic case, 
likely it will be optional attributes that exceed the length, with local 
node having no knowledge based on what criteria to remove which 
attribute from propagation. Also this is not specific to extended 
messages. Behaviour today seems to be left to implementation detail. 
Should be logged and possibly dropped, leaving to the implementation 
detail.


Ignas



On 24/05/2016 21:42, Susan Hares wrote:
>
> This begins a 2 week WG LC on draft-ietf-idr-bgp-extended-messages 
> (5/24 to 6/7). The draft can be found at:
>
> https://datatracker.ietf.org/doc/draft-ietf-idr-bgp-extended-messages/
>
> In your comments please provide the following information:
>
> 1)Support/”no-support” of this draft being ready for publication,
>
> 2)You know of an implementation for this draft, please report the 
> implementation to the IDR Wiki
>
> https://trac.tools.ietf.org/wg/idr/trac/wiki/draft-ietf-idr-bgp-extended-implementations 
>
>
> 3)Do you have any concerns about extending the BGP messages.
>
> Sue Hares and John Scudder
>
>
>
> _______________________________________________
> Idr mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/idr