[mpls] Updated version of my proposed response ITU SG15 Q12 Liaison Progress on G.8110.1

[Stewart Bryant <stbryant at cisco.com>]

To PWE3 and MPLS WG:

Here is an updated version of my proposed response to 

https://datatracker.ietf.org/public/liaison_detail.cgi?detail_id=312

- Stewart

=================

To etc


The PWE3 and MPLS WGs have studied ITU SG15 Q12 Liaison Progress on G.8110.1
Amendment

https://datatracker.ietf.org/public/liaison_detail.cgi?detail_id=312

and have the following comments on the document:


Section 3.5 Transport MPLS definition

Should be amended to make it clear TMPLS is a true subset 
of IETF PWE3 over MPLS and that in the case of any differences, 
the IETF definition is considered the Normative definition.

Section 4.7.2 and A.3

The PWE3 WG is concerned that the OAM is not based on VCCV. Any true
subset of the IETF PWE3/MPLS design should allow the use of the IETF
preferred OAM mechanism. This does not preclude the ITU specifying an
additional mechanism as an optional extension to enhance VCCV, although
it would be preferable for the ITU to work with the IETF to enhance VCCV
to provide the additional OAM mechanisms that are needed.

The PWE3 WG is not clear how you determine that the client signal is not
present when the client is a packet service.


Section 4.10

"– Minimize the rerouting distance (to avoid physical layer impairments
on the signals)."

We do not understand why in a well engineered network the distance in
the virtual layer providing reroute has impact results in physical layer
impairments

Appendix V

The introductory material to Appendix V states that it will be
re-written at some time in the future to align with the IETF work on
p2mp LSPs. We would therefore request that you liaise the final text 
to us when it is completed.

As it stands the text of Appendix V looks correct, however we do have
two proposals that we would ask you consider


1.  We recommend that Upstream Assigned Labels be used over P2MP
   connections, or at least that you note in the text that generally 
   administration of labels is simplified by using upstream 
   assignment in this scenario.

2.  That you include a reference to RFC 4875 (RSVP-P2MP) as
   this discusses many issues pertaining to P2MP MPLS connections.

Regards

PWE3 and MPLS Co-chairs.

_______________________________________________
mpls mailing list
mpls at lists.ietf.org
https://www1.ietf.org/mailman/listinfo/mpls