Skip to main content

Path Computation Element Communication Protocol (PCEP) Extensions for Segment Routing (SR) Policy Candidate Paths
draft-ietf-pce-segment-routing-policy-cp-27

Yes

Roman Danyliw

No Objection

Andy Newton
Erik Kline
Gorry Fairhurst
Jim Guichard
Orie Steele
Paul Wouters

Note: This ballot was opened for revision 20 and is now closed.

Ketan Talaulikar
(was Discuss) Yes
Comment (2025-04-03 for -26) Sent
Thanks for addressing my discuss points/comments and other comments. Following are further comments suggested to improve the updated text that was introduced.

Provided inline in the idnits format output for the v26 of the document.

209	   The term "LSP" in this document represents Candidate Path within an
210	   SR Policy.  In the context of SR Policy for SRv6, the term "LSP" in
211	   this document refers to an SRv6 path, which is represented as a list
212	   of SRv6 segments.

<minor> How about?

CURRENT
In the context of SR Policy for SRv6, the term "LSP" in this document refers to 
an SRv6 path, which is represented as a list of SRv6 segments.

SUGGEST
In the context of SR Policy for SRv6 (refer [RFC9603]), the term "LSP" in this
document refers to an SRv6 path, which is represented as a list of SRv6 segments.


249	   [RFC8697] specifies the mechanism for the capability advertisement of
250	   the Association Types supported by a PCEP speaker by defining an
251	   ASSOC-Type-List TLV to be carried within an OPEN object.  This
252	   capability exchange for the SR Policy Association Type MUST be done
253	   before using the SRPA.  To that aim, a PCEP speaker MUST include the
254	   SRPA Type (6) in the ASSOC-Type-List TLV and MUST receive the same
255	   from the PCEP peer before using the SRPA (Section 6.1).  SRPA MUST be
256	   assigned for all SR Policy LSPs by PCEP speaker originating the LSP
257	   if capability was advertised by both PCEP speakers.

<major> What would be the error reported by the PCEP speaker if it were to
received an SR LSP (say using mechanism in RFC8664) without an SRPA even after
successful capability negotiation? Perhaps there is an existing error that can
be used?

294	   SR Policy Candidate Path Identifier uniquely identifies the SR Policy
295	   Candidate Path within the context of an SR Policy.  SR Policy
296	   Candidate Path Identifier is assigned by PCEP peer originating the
297	   LSP.  Candidate Paths within an SR Policy MUST NOT change their SR
298	   Policy Candidate Path Identifiers for the lifetime of the PCEP
299	   session.  Candidate Paths within an SR Policy MUST NOT carry same SR
300	   Policy Candidate Path Identifiers in their SRPAs.  If the above

<minor> How about?

CURRENT
Candidate Paths within an SR Policy MUST NOT carry same SR Policy Candidate Path Identifiers in their SRPAs.

SUGGEST
Two or more Candidate Paths within an SR Policy MUST NOT carry same SR Policy Candidate Path Identifiers in their SRPAs.


< EoR v26 >
Mohamed Boucadair
(was Discuss) Yes
Comment (2025-04-01 for -24) Sent
Hi Samuel, all,

The changes in [1] fully address my previous DISCUSS/COMMENTs [2]. Thank you for carefully tracking all those. There are some few nits:

* Section 2: s/An LSP set up using/An LSP setup using
* Section 4: s/the ASSOCIATION object Section 6.1 of [RFC8697]/the ASSOCIATION object (Section 6.1 of [RFC8697])
* Section 5.14: 

 (1) s/If this flag is not set/If this flag is set to 0  (more than occurrence in the text)

 (2)

  OLD:
   More flags can be assigned in the future per (Section 6.7)
  
  NEW:  
   More flags can be assigned in the future per (Section 6.7).

* Section 5.4.1: s/As stated in Section 8.1 of [RFC9256]1/As stated in Section 8.1 of [RFC9256]

As a general note (and I'm looking to the routing ADs :-)), I hope we can have an overview document that glue the various SR policy (and beyond) pieces in operations and go through to assess how these can be grafted together.

Cheers,
Med

[1] https://author-tools.ietf.org/iddiff?url2=draft-ietf-pce-segment-routing-policy-cp-24
[2] https://mailarchive.ietf.org/arch/msg/pce/zYWbTajJ943cNxLLySO6kCEaw4Q/
Roman Danyliw
Yes
Andy Newton
No Objection
Deb Cooley
No Objection
Comment (2025-03-30 for -23) Sent
Thanks to Joe Salowey for his secdir reviews.

One very small comment: 

Abstract, para 2:  'bring up'?  what does this mean?  Are we bringing up protocols, like we raise children?  Or does it mean something else?
Erik Kline
No Objection
Gorry Fairhurst
No Objection
Jim Guichard
No Objection
Mike Bishop
No Objection
Comment (2025-03-25 for -23) Sent
Many abbreviated terms in the Introduction aren't defined until Section 2. (PCC, SR-TE, LSPs, etc.) Consider expanding these on first use. Further, several of the abbreviations expanded in Section 2 don't have references to the RFCs where the terms are specified. This makes it difficult to explore further. For example, PCC is defined in RFC5440 which is not referenced by this document until Section 6. Compare this to the Terminology section of RFC8231, which provides explicit sources for each term.

Throughout 3.x, should the names of these elements have "The" or "A" before them? Also, s/consist/consists/.

The sentence in the Acknowledgement lacks a subject. Who would like to thank these people?
Orie Steele
No Objection
Paul Wouters
No Objection
Éric Vyncke
(was Discuss) No Objection
Comment (2025-04-03 for -26) Sent
Thanks for *partially* addressing my blocking DISCUSS (including the last one, which was caused by my misreading of the I-D). I am clearing my DISCUSS.

***BUT***, I still find the text in section 3 very unclear about its applicability to SRv6, the LSP explanation added at the end of section 2 is enough to clear my previous DISCUSS, but ***the whole section can only confuse the readers***.

# Below kept for archiving

No need to act on anything below.

## Archive of DISCUSS (blocking)

As noted in https://www.ietf.org/blog/handling-iesg-ballot-positions/, a DISCUSS ballot is just a request to have a discussion on the following topics:

### Section 3

This section appears to be only about SR-MPLS, i.e., conflicting with the abstract, which explicitly includes SRv6, and the title, which does not mention any restriction to SR-MPLS only. Either add some SRv6 text (my preference) or restrict the I-D to SR-MPLS only.

### Section 4.1

Let's remove any ambiguity and specify the units (octets I guess) and the field coverage (color + endpoint I guess) for the Length field.

### Sections 4.2.1 & 4.2.3

`It is RECOMMENDED that the size of the symbolic name for the SR Policy is limited to 255 bytes. ` why only "RECOMMENDED" when the Length field can only indicate 255 maximum ? Suggest using MUST rather RECOMMENDED here.

*My bad on this one, Med corrected me*

## Archive of COMMENTS (non-blocking)

### Are RFC 8986 policies included ?

This document does not seem to include RFC 8986 (network programming) policies as it only talks about "paths". The document should be clear whether it supports RFC 8986 and, if so, amend the text about "paths".

### Abstract

Suggest deleting `called a headend node` as it brings no value there.

### Section 1

s/Path Setup Type 1 (Segment Routing) /Path Setup Type 1 (SR-MPLS) / ? (and possibly other places, let's remove any ambiguity)

### Sections 4.2.1 and 4.2.3

`Padding is not included in the Length field.` do PCEP receivers know how to handle a Length that does not really represent the length of the TLV ?

### Section 11.2

RFC20 should be normative.

## NITS (non-blocking / cosmetic)

### Section 1

Consider using lower case for `Establishing Relationships Between Sets`

### Section 5.1

Please use balanced quotes in `Type: 71 for "SRPOLICY-CAPABILITY TLV`

### Use of SVG graphics

To make a much nicer HTML rendering, suggest using the aasvg too to generate SVG graphics. It is worth a try ;-)