[mpls] RtgDir review: https://www.ietf.org/id/draft-ietf-mpls-spring-lsp-ping-06.txt

[Alexander Vainshtein <[e-mail]>]

Hi,
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: https://www.ietf.org/id/draft-ietf-mpls-spring-lsp-ping-06.txt

Reviewer: Alexander (“Sasha”) Vainshtein

Review date: 20.09.2017

IETF LC End Date: Not known

Intended status: Standards Track

Summary:

I have significant concerns about this document and recommend that the Routing ADs discuss these issues further with the authors.

Comments:

I have noticed that there is an earlier review of the same version of  the draft<https://www.ietf.org/mail-archive/web/rtg-dir/current/msg03542.html>, and I fully agree with the previous reviewer regarding the overall impression and specific issues raised. I have tried to avoid reporting the same issues again and focus only on new issues.

This review have been done on a very tight schedule for me, and therefore I could have missed some issues and definitely many nits. The tight schedule has also limited my ability to discuss the draft with the authors in sufficient detail – but some interactions did took place. I would like to thank the authors for their cooperation.

Major Issues:


1.       Section 7.4 “Segment ID Check” of the draft claims to update “the procedure defined in Step 6 of Section 4.4  of [RFC8029]”. However, I have failed to integrate the logic defined by the pseudocode in this section with the logic defined by the referenced pseudocode.

a.       I have suspected that the newly introduced logic should be part of the Egress FEC Validation logic defined in Section 4.4.1 of RFC 8029.  This suspicion has been acknowledged by the authors, and they plan to fix the wrong reference in the next revision of the draft.

b.       From my POV the best way to avoid possible misinterpretation by the implementers would be inclusion of the updated version of the pseudocode (with explicit marking of the new logic) directly in the document. I have suggested this to the authors, but I did not get their response so far.

2.       I agree with the previous reviewer regarding the mismatch between the (presumed) claim and the actual scope covered by the draft. To the best of my understanding, the authors have already agreed to fix this in the next revision of the draft.
Minor Issues:


1.       I wonder why the authors place such an emphasis on Service labels (which, AFAIK, have not been defined in any level of detail anywhere so far) – these labels are mentioned in Sections 4.2, 5 (that mentions them as “Service segments”)  and 8, while so many types of Segment IDs that are already well defined both in the SRPING WG documents and in the extensions of the relevant routing protocols are left out of scope of the draft.  The authors have acknowledged this and plan to fix this issue (probably by removing the Service labels completely).

2.       The description of possible data plane malfunctions in section 4.1 seems to assume that all the nodes in the referenced topology (shown in Figure 1) us the same SRGB. If this assumption (which is not explicitly mentioned) is not correct, the LSP Ping Echo request packets would not be “delivered to their expected destinations but not via the expected path” (as the text in this section claims). I suggest making any such assumptions explicit in the text

3.        I am not sure if “OSPF” and ISIS” (without any reference to Segment Routing) are the proper names for the label distribution protocols in the proposed new IANA registry in Section 10.2. From my POV “OSPF/ISIS SR Extensions” would be more appropriate. The authors disagree with this proposal claiming that OSPF and ISIS are not used for label distribution in any other way. (RFC 8029 that encodes the label distribution protocol in the Downstream  Detailed Mapping TLV did not request a dedicated registry for this purpose, and I think that the authors have been right to request such a registry).

4.       I doubt the draft needs address any FRR issues even in future (as mentioned in the last statement in Section 5), since, to the best of my understanding, any form of FRR:

a.       Is operated locally by the node that is upstream to the failure without passing any information to the initiator of LSP-Ping Request packets

b.       In any case is just a transient state in the (presumably short) interval between the failure being detected by the upstream node and re-convergence of IGP
I also think that references to “future versions” are not quite appropriate in the document that is going to the IETF LC. I recommend removing any such references from the document,

5.       Section 9 “Backward Compatibility with non-Segment Routing devices” defines the behavior for the two slightly different use cases:

a.       LSP Ping Echo Request packets are initiated by an SR-capable node (and therefore their target FEC stack contains SR-related FECs), while the responder is legacy device that is not SR-capable. In this case the proposed solution (respond with the FEC-return-code “Replying router has no mapping for the FEC at stack-depth”) looks as the only possibility since the legacy device cannot parse SR FECs in any case

b.       LSP Ping Echo Request packets are initiated by a legacy device while the responder is SR-capable. The draft defines the same behavior in this case – but, IMHO and FWIW, the responder could provide slightly better diagnostics since it can parse the “old” target FEC and recognize the equivalent Node ID. An additional return code would be required to implement this behavior.
This issue has not been discussed with the authors due to lack of time.
Nits:


1.       The note to the RFC Editor in Section 10 mentions early IANA allocation for some Sub-TLV types defined in the document – but, as of today, these early allocations have already expired (they have been in force until 15.09.17).  I have to admit that I do not fully understand the implications of this fact – e.g., whether the authors may continue to refer to the specific values of parameters (e.g., Sub-TLV 34, 35 etc.) for which early IANA allocation has expired, or should use some other form of reference (e.g., TBDx).

2.       The draft mentions a TBD7 value to be assigned by IANA, but there are no TBD1, …, TBD6. The authors have acknowledged this and plan to fix it. However, I do not know how this can be affected by expiration of early IANA allocations (see above)

3.       I have failed to understand the intention of the authors regarding already mentioned Service labels from the following statement in section 8: “How a node treats Service label is outside the scope of this document and will be included in this or a different document later”. Of course, if the Service labels are removed from the draft, this sentence would hopefully disappear with its internal controversy.

Regards,
Sasha

Office: [phone]
Cell:      [phone]
Email:   [e-mail]


___________________________________________________________________________

This e-mail message is intended for the recipient only and contains information which is 
CONFIDENTIAL and which may be proprietary to ECI Telecom. If you have received this 
transmission in error, please inform us by e-mail, phone or fax, and then delete the original 
and all copies thereof.
___________________________________________________________________________