Re: [Teas] RtgDir review: draft-ietf-teas-p2mp-loose-path-reopt-07.txt

["Rakesh Gandhi (rgandhi)" <[e-mail]>]

Hi Joel,

Thank you for the detailed review of the document. Please see inline <RG> for replies..




On 2016-11-23, 4:10 PM, "Joel M. Halpern" <[e-mail]> wrote:

>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-teas-p2mp-loose-path-reopt-07.txt
>Reviewer: Joel M. Halpern
>Review Date: 23-November-2016
>IETF LC End Date: N/A
>Intended Status: Standards Track
>
>Summary: I have some moderate concerns about this document that I think 
>should be resolved before publication is approved.
>
>Comments:
>
>Major:
>     The use of SHOULD and MAY in section 4.1 seems to lead to a device 
>which ostensibly supports this document, but does the wrong things.
>     First, with regard to the SHOULDs, in the absence of any indication 
>as to why it would not do this, it appears that the SHOULD is really 
>"MUST if the device supports this document" which is what MUST in a 
>document actually means.
>     Section 4.2 first bullet says that a mid-point LSR "SHOULD" check 
>for a preferable P2MP-TE LSP Tree.  But if it doesn't, it is not 
>supporting this document.  As written, it could decide to ignore the 
>message, even though it claims to support this RFC.
>     Looking at the handling when a preferable P2MP-TE LSP tree is 
>found, according to the document, the LSR MAY send the PathErr response. 
>  My assumption is that if it does not send the PathErr, it MUST 
>propagate the request.  If it does not do either one, the protocol does 
>not function.  It seems likely that if this is really intended to be 
>optional (MAY), the document would be improved my giving implementors 
>some hint as to when it is desirable or undesirable to send the message.
>     Then in the third bullet, it is only a SHOULD to pass on the 
>request.  Thus, a device which supports this mechanism, but chooses not 
>to pass on the request, is compliant to this document while preventing 
>other devices from properly supporting the mechanism.



<RG> Ok, how about following text in Section 4.1?

-----------
A mid-point LSR that expands loose next-hop(s) for one or more S2L
   sub-LSP path(s) does the following upon receiving a Path message with
   the "P2MP-TE Tree Re-evaluation Request" flag set:

   o  The mid-point LSR MUST check for a preferable P2MP-TE LSP tree by
      re-evaluating all S2L sub-LSP(s) that are expanded paths of the
      loose next-hops of the P2MP-TE LSP.  

   o  If a preferable P2MP-TE LSP tree is found, the mid-point LSR MUST
      send an RSVP PathErr with the Notify error code 25 defined in
      [RFC3209] and sub-code "Preferable P2MP-TE Tree Exists (value
      TBA2)" defined in this document to the ingress node.  The mid-
      point LSR, in turn, SHOULD NOT propagate the "P2MP-TE Tree Re-
      evaluation Request" flag in the subsequent RSVP Path messages sent
      downstream for the re-evaluated P2MP-TE LSP.

   o  If no preferable tree for P2MP-TE LSP can be found, the mid-point
      LSR that expands loose next-hop(s) for one or more S2L sub-LSP
      path(s) MUST propagate the request downstream by setting the
      "P2MP-TE Tree Re-evaluation Request" flag in the LSP_ATTRIBUTES
      Object of the RSVP Path message.

   The sending of an RSVP PathErr with the Notify error code and
   "Preferable P2MP-TE Tree Exists" sub-code to the ingress node
   notifies the ingress node of the existence of a preferable P2MP-TE
   LSP tree and upon receiving this PathErr, the ingress node MUST
   trigger re-optimization of the LSP using the MBB method with a
 
different LSP-ID.
---------------



>
>Minor:
>     The abstract is much too long.  Much of the content of the abstract 
>belongs in the introduction.  Even teh second paragraph has too much 
>detail for an abstract.
>
>Editorial:
>     In the last paragraph of the introduction, it says that this 
>document "proposes" solutions.  Given we are now in the position of 
>evaluating publication as a Proposed Standard, I would say that this 
>document "defines" solutions.
>


<RG> Ok, how about following Abstract?

----------
Re-optimization of a Point-to-Multipoint (P2MP) Traffic Engineered
   (TE) Label Switched Path (LSP) may be triggered based on the need to
   re-optimize an individual source-to-leaf (S2L) sub-LSP or a set of
   S2L sub-LSPs, both using Sub-Group-Based Re-optimization method, or
   the entire P2MP-TE LSP tree using the Make-Before-Break (MBB) method.
    This document discusses the application of the existing mechanisms
   for path re-optimization of loosely routed Point-to-Point (P2P) TE
   LSPs to the P2MP-TE LSPs, identifies issues in doing so and defines
   procedures to address them.  When re-optimizing a large number of S2L
   sub-LSPs in a tree using the Sub-Group-Based Re-optimization method,
   the S2L sub-LSP descriptor list may need to be semantically
   fragmented.  This document defines the notion of a fragment
   identifier to help recipient nodes unambiguously reconstruct the
   fragmented S2L sub-LSP descriptor list.
----------



Thanks,
Rakesh (for authors and contributors)



>