Dear chairs,

I support the adoption of this work.

And I agree with Yingzhen’s comment on the text in section 6. The current text 
about the default behavior of the receiving BGP speaker across AS boundary is 
not clear. If by default it MUST NOT remove the received extended communities, 
then removing the extended communities cannot be set as one of the “default 
behavior”.  It is suggested that the default behavior be specified in the 
draft, then it can say the behavior can be overridden by a knob.

Best regards,
Jie

From: Yingzhen Qu <[email protected]>
Sent: Tuesday, September 2, 2025 2:12 PM
To: Jeffrey Haas <[email protected]>
Cc: idr <[email protected]>; BESS <[email protected]>
Subject: [bess] Re: [Idr] Working Group adoption check, RFC 4360-bis, ending 5 
September, 2025.

I support the adoption of this work.

The following added text in Section 6 is to clarify the behaviors with 
non-transitive extended communities.
"

   If a route has non-transitive extended communities, those communities

   SHOULD NOT be propagated across an Autonomous System boundary and

   SHOULD be removed from the route.  However, non-transitive extended

   communities SHOULD NOT be removed when advertising the route within

   the same BGP AS Confederation(as defined in [RFC5065]).  As part of

   configuration or BGP protocol extensions, BGP speakers MAY attach

   non-transitive extended communities to routes advertised across

   Autonomous System boundaries.



   By default, when a BGP speaker receives routes with non-transitive

   extended communities across Autonomous System or Confederation

   Member-AS boundaries, it MUST NOT remove these extended communities.

   This default behavior MAY be configurable.  The BGP speaker SHOULD

   also allow local policies to match against or remove these extended

   communities.
"

I'd like to understand the reason for using "SHOULD NOT/SHOULD" in the first 
paragraph while "MUST NOT" in the second paragraph.
Also "This default behavior MAY be configurable.", I think it should be that 
the draft specifies the default behavior, and there might be configuration 
knobs to change the behavior.

Thanks,
Yingzhen



On Fri, Aug 22, 2025 at 12:23 PM Jeffrey Haas 
<[email protected]<mailto:[email protected]>> wrote:
IDR, BESS,

During the work driven by draft-ietf-idr-link-bandwidth, the issue of 
originating non-transitive was brought up and partially discussed in the use 
case work for draft-ietf-bess-ebgp-dmz.  As discussed during IDR sessions at 
IETFs 122 and 123, the preferred solution for addressing the ambiguities in 
non-transitivity was to do a small -bis for RFC 4360.  Nat Kao has kindly 
agreed to be our editor to move this process along. This document, and issues 
vs. it, will be managed in the IDR github.[1]

Since this is IDR chair commissioned work to address this gap, it's our 
intention to adopt this work.  However, the chairs would like to provide a 
review period to OBJECT to adoption.  That said, if you'd like to offer support 
for the work, or other technical comments, please do so in this thread!

This adoption check ends on 5 September.  Please note this overlaps the US 
Labor Day holiday and consider that in the timing of your request, in case 
that's relevant.

The scope of the commissioned work is:

- Address open errata vs. RFC 4360
- Address the origination and reception of non-transitive routes across eBGP 
boundaries.

The current text of the draft currently addresses these items.

As part of reviewing this problem, the IETF archives show that there was prior 
work covering this issue in draft-decraene-idr-rfc4360-clarification-00 [2].  
We've made sure to acknowledge those prior efforts in the -bis and would 
request review from those authors on this -bis.

-- Jeff (for the IDR Chairs)

[1] https://github.com/ietf-wg-idr/draft-ietf-idr-rfc4360-bis
[2] Bruno and company are to be commended for pressing this issue for several 
years.  While prior IDR mail threads seem to suggest "this works fine was the 
answer", the fact that we had non-transitive behaviors as a point of contention 
in the BESS LBW work means it's past time to enshrine fixing the original 
criticisms in an RFC update.


Begin forwarded message:

From: [email protected]<mailto:[email protected]>
Subject: I-D Action: draft-chairs-idr-rfc4360-bis-00.txt
Date: August 22, 2025 at 2:46:40 PM EDT
To: <[email protected]<mailto:[email protected]>>
Reply-To: [email protected]<mailto:[email protected]>

Internet-Draft draft-chairs-idr-rfc4360-bis-00.txt is now available.

  Title:   BGP Extended Communities Attribute
  Author:  Nat Kao
  Name:    draft-chairs-idr-rfc4360-bis-00.txt
  Pages:   13
  Dates:   2025-08-22

Abstract:

  This document describes the "extended community" BGP-4 attribute.
  This attribute provides a mechanism for labeling information carried
  in BGP-4.  These labels can be used to control the distribution of
  this information, or for other applications.

  This document obsoletes [RFC4360].

The IETF datatracker status page for this Internet-Draft is:
https://datatracker.ietf.org/doc/draft-chairs-idr-rfc4360-bis/

There is also an HTMLized version available at:
https://datatracker.ietf.org/doc/html/draft-chairs-idr-rfc4360-bis-00

Internet-Drafts are also available by rsync at:
rsync.ietf.org::internet-drafts


_______________________________________________
I-D-Announce mailing list -- [email protected]<mailto:[email protected]>
To unsubscribe send an email to 
[email protected]<mailto:[email protected]>

_______________________________________________
Idr mailing list -- [email protected]<mailto:[email protected]>
To unsubscribe send an email to [email protected]<mailto:[email protected]>
_______________________________________________
BESS mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to