Early Review of draft-ietf-idr-bgpls-inter-as-topology-ext-23
review-ietf-idr-bgpls-inter-as-topology-ext-23-rtgdir-early-richardson-2026-04-15-00

RtgDir Early review: draft-ietf-idr-bgpls-inter-as-topology-ext-23

I have been selected to do a routing directorate “early” review of this draft.
​https://datatracker.ietf.org/doc/draft-ietf-idr-bgpls-inter-as-topology-ext-23

As this document is approaching WGLC, this early rtg directorate review is to notice
any issues that might hang the WGLC process.

Document: draft-ietf-idr-bgpls-inter-as-topology-ext-23
Reviewer: Michael Richardson
Review Date: 2026-04-15
Intended Status: Standards Track

Summary:  Document seems in good shape

I have some minor concerns about this document that I think should be resolved before WGLC.

Comments:

This document is well written.
While I am familiar with PCE and other SDN concepts, its' not a core part of
my work.  I was unaware of the contents of RF9552, and I reviewed that
briefly.

I found that the problem description was well presented and I did not have a
problem understanding what problem this document is trying to solve.

The following are minor issues, mostly editorial.

1. Figure 1 has some issues at the S2 and T1 box, with the BGP-LS lines. The
ascii art suggests that the left-side BGP-LS connects to "ISP-A"
box, while right-side BGP-lS seems to connect to T3.  I'm not sure if that
was intended.  I looked if this was translated to SVG, but as the document
seems to use XML as it's reference source, so doing aasvg might be hard.
But please consider kramdown, as that would make your life much easier.

2. Section 6.2, says we can use any stable, routable IPv4.
   My understanding is that many larger intranets do not use routable IPv4.
   Maybe the goal is to have a stable unique v4.
   Is an IPv4 ASBR *required* for an IPv6-only Enterprise/ISP?

   Meanwhile, 6.3 seems to make IPv6 ID optional, deferring to using the
   peer's IPv4, even though IPv6 GUA ought to be far easier to get.  Maybe
   there shoud be parity here: v6 can replace v4 ID?

4. Section 5.
   I found the reference for the "Local Node Descriptor" mentioning the TLV
   by number (256) odd, and I couldn't follow that reference easily.
   The reference is was wrong: it should be 5.2.1.2.
   The TLVs which are taken from 9552 are sometimes by name, and
   sometimes by number, and maybe it shoild be consistently by name?

Nits are editorial or layout items.
  I didn't run the idnits tools, I'm sure the shepherd will.