Early Review of draft-ietf-idr-add-paths-10
review-ietf-idr-add-paths-10-rtgdir-early-pignataro-2016-04-28-00

Hello,

I have been selected as the Routing Directorate reviewer for this draft. The
Routing Directorate seeks to review all routing or routing-related drafts as
they pass through IETF last call and IESG review, and sometimes on special
request. The purpose of the
 review is to provide assistance to the Routing ADs. For more information about
 the Routing Directorate, please see ​

http://trac.tools.ietf.org/area/rtg/trac/wiki/RtgDir



Although these comments are primarily for the use of the Routing ADs, it would
be helpful if you could consider them along with any other IETF Last Call
comments that you receive, and strive to resolve them through discussion or by
updating the draft.

Document: draft-ietf-idr-add-paths-13

Reviewer: Carlos Pignataro

Review Date: April 25, 2016

Intended Status: Proposed Standard

https://datatracker.ietf.org/doc/draft-ietf-idr-add-paths/

Summary:

This document is basically ready for publication, but has nits that should be
considered prior to publication.

Comments:

This is an extremely well written document, very focused and with high SNR. I
have a couple totally-non-critical comments and questions below.

Major Issues:

None.

Minor Issues:

I have a couple of questions rather than issues:

2.  How to Identify a Path

...

   A BGP

   speaker that receives a route SHOULD NOT assume that the identifier

   carries any particular semantics; it SHOULD be treated as an opaque

   value.

I was thinking about why “SHOULD NOT” and not “MUST NOT”, and I understand
future proofing, but wondering if there’s another reason.

The path identifier has well-defined semantics: make a path unique for a given
prefix, or make the {identifier; prefix} specify a path among many. Does this
sentence intend to specify that a BGP receiving a route SHOULD NOT assume
particular semantics
 of the numerical value of the field? (such as the lower value means a more
 important route), or SHOULD NOT assume particular semantics of the structure
 of the field? (such as some hierarchy, or MSB with some meaning). Or both?

Maybe, “… assume that the value or structure of the identifier carries …”?

Or maybe it is OK as is and I am reading too much into it?

6.  Applications

   The BGP extension specified in this document can be used by a BGP

   speaker to advertise multiple paths in certain applications.  The

   availability of the additional paths can help reduce or eliminate

   persistent route oscillations [RFC3345].  It can also help with

   optimal routing and routing convergence in a network.  The

   applications are detailed in separate documents.

The final sentence does not seem to add anything, other than questions. I’d
suggest either adding pointers to those separate documents, or removing the
sentence. This is a non-normative section.

7.  Deployment Considerations

   The extension proposed in this document provides a mechanism for a

   BGP speaker to advertise multiple paths over a BGP session.  Care

   needs to be taken in its deployment to ensure consistent routing and

   forwarding in a network, the details of which will be described in

   separate application documents.

Similarly, which documents? These seem like important considerations.

Nits:

4.  ADD-PATH Capability

   The ADD-PATH Capability is a new BGP capability [RFC5492].  The

   Capability Code for this capability is specified in the IANA

   Considerations section of this document.

Why not say "The ADD-PATH Capability is a new BGP capability [RFC5492], with
Capability Code of 69” and simplify it for the reader?

9.  Security Considerations

...

  The use of the ADD-PATH Capability is intended to

   address specific needs related to, for example, eliminating the MED-

   induced route oscillations in a network

   [I-D.ietf-idr-route-oscillation-stop].  While the applications for

   the ADD-PATH Capability are outside the scope of this document, the

   users are encouraged to examine their behavior and potential impact

   by studying the best practices described in

   [I-D.ietf-idr-add-paths-guidelines].

Are these Security Considerations, Applications, or Deployment Considerations?

I hope these help,

— Carlos.