Hi Sue, Thanks for your comments. Yes, we are in discussion with other vendors regarding link-bandwidth and we will together put a document that will consider all the aspects that you have mentioned. Indeed, the transitive or non-transitive implementation across vendors poses a challenge to implementation. Yes, we will refresh draft-ietf-idr-link-bandwidth possibly with new authors. That was also one of the points in our offline discussions.
Regarding your comment and alluding to entropy label, we will need further internal discussion. BTW, (2) use-cases, it seems your last sentence is truncated. Maybe a typo. Thanks, --Satya From: BESS <bess-boun...@ietf.org> on behalf of Susan Hares <sha...@ndzh.com> Date: Thursday, July 27, 2023 at 4:36 PM To: BESS <bess@ietf.org> Cc: Andrew Alston <andrew.als...@liquidtelecom.com> Subject: [bess] Review of draft-ietf-bess-ebgp-dmz-03 Bess chairs: The IDR WG was queried for a review of this document. No responses were made. I wrote an IDR chair review is contained on the IDR section of the IETF community wiki: https://wiki.ietf.org/en/group/idr/draft-ietf-bess-ebgp-dmz Summary: The IDR chairs note that authors are discussing non-transitive and transitive extended communities for link bandwidth passed by BGP extended communities. We suggest that these efforts continue. In this process, I have offered additional things the IDR chairs will review in these documents. It is time to ensure “link bandwidth” uses are harmonized across BGP mechanisms (attributes or extended communities) and bgp-ls reporting. As the reviewing IDR chair, I find the publication this document at this time is premature. However, it is a useful input to the process. The IESG while reviewing draft-ietf-idr-entropy-label for publication should consider how extended communities, the router capability attributes, and BGP-LS reporting aligns for link, router, and AS bandwidth. I believe that the chairs of the WGs related to BGP and BGP-LS in IGPs should discuss this topic (e.g. IDR, BESS, Spring, LSVR, Grow, MPLS) Sue Full text draft-ietf-bess-ebgp-dmz-03 IDR Chair review Reviewer: Susan Hares Issues with this draft: 1. Protocol Content Four drafts deal with link bandwidth for a BGP router passed in an extended communities attribute or the entropy attribute outside of BGP-LS reporting. a) draft-ietf-idr-link-bandwidth (a non-transitive extended community attribute) b) draft-ietf-bess-ebgp-dmz-03 (a transitive extended community) c) draft-ietf-entropy-label (router capability attribute) Work is underway by the authors to harmonize the transitive and non-transitive use of the community. Section 6 of draft-ietf-bess-ebgp-dmz-03 indicates a need for a refresh of draft-ietf-idr-link-bandwidth. The IDR chairs suggest this work continues before publishing the use case found in this draft. As part of this work, the authors should consider: a) whether the description is a link, router, or AS bandwidth. b) the ramifications of passing this information as extended community or an attribute, and c) how this relates to the BGP-LS definitions. 2. Use cases The draft presents the following use cases: a) large-scale data centers (RFC7938, section 6.3) unequally weighted ECMP, b) large-scale data centers (RFC79388) equally weighted ECMP, c) external community and top-down Load-balanced community, and d) no-conforming topologies. The descriptions of these cases provide a helpful summary of these use cases. These descriptions help focus the discussions for protocol content. Additional value can be gained for the current protocol discussions by indicating the answers to the questions on protocol content. Whether this 3. English text The English text has spelling errors, grammar errors, and portions that need to be clarified. At this stage, the content needs to be considered before a final check of the text. Before requesting a review of the English text, the authors should use the commonly available tools (such as "Grammarly") to check the text.
_______________________________________________ BESS mailing list BESS@ietf.org https://www.ietf.org/mailman/listinfo/bess
