IETF PWE3 WG Response to Addition of ATM and PDH clients to T-MPLS
Date: 09 February 2008

To:
tsbsg15@itu.int
betts01@nortel.com
greg.jones@itu.int

CC:
sob@harvard.edurcallon
dward@cisco.com
rcallon@juniper.net
mpls@lists.ietf.org
jari.arkko@piuha.net
townsley@cisco.com
mankin@psg.com
matthew.bocci@alcatel-lucent.co.uk
pwe3@ietf.org
yoichi.maeda@ntt-at.co.jp
sjtrowbridge@alcatel-lucent.com

swallow@cisco.com
loa@pi.se
danny@arbor.net
stbryant@cisco.com

statements@ietf.org

For Action by ITU-T SG15 Question 12


IETF PWE3 WG Response to Addition of ATM and PDH clients to T-MPLS

Thank you for your liaison of 9th January 2008 entitled
"Addition of ATM and PDH clients to T-MPLS"

<https://datatracker.ietf.org/documents/LIAISON/file510.pdf>

We found this short description of the packet processing more
difficult to read than the descriptions found in either
the relevant RFCs or the corresponding Y series recommendations
as such we found it difficult to review. Rather than defining
the operations in semi-detail we would prefer that you simply
indicate the primary references and modes that you propose to
support and refer the reader there for the description. This
prevents any issue of conflicting definitions. In any case
there needs to be a clear indication that this is a subset
of existing IETF PWE3 and SG13 recommendations
and that their packet processing definitions are authoritative.

We are concerned that you only propose to use the structure
agnostic TDM mode. This mode has limitations which are
addressed by the structure aware modes jointly developed
in the IETF PWE3 Working Group and ITU-T SG13.

Please can we suggest that you add text explaining that the
structure-agnostic mode is not always the idea method of TDM
emulation and preferably include text describing the use of
the structure aware modes as described in Y.1413, RFC 5086
and RFC 5087.

In the case of ATM, please can you be more specific about
the exact type of PW that you intend to support. The
confusion arises as a result of the text at the end of
section 7.1.4 of the document which says:

" For the case of Mode 1

* Cell de-multiplexing according to the VPI value, including
unmatched VPI cell discard
* Remove the VPI field"

Since MODE 1 is defined as the N-to-One cell mode of
[IETF RFC 4717], this this text appears to be in error, and
we believe that you should call up Mode 2.

There are some other items that are unclear in this section.
The paper discusses N-1 mode but then discusses multiple VP.
Is the mode that you wish to use what we call cell relay
mode PW type 0x0003 ?


Regards

Stewart Bryant
Danny McPherson
IETF PWE3 WG Co-chairs