Re: I-D ACTION:draft-andersson-mpls-g-chng-proc-00.txt

[Stephen Trowbridge <sjtrowbridge@lucent.com>]

Scott,
There is no doubt that liaisons CURRENTLY have no more wieght than
individual IDs (perhaps even less - by the Bala flowchart, the fact
that an idea arrives via liaion may even work to the detriment of
a proposal (yes Loa, I know this is a joke!)).
But it is my opinion that the lack of a liaison process is really
the ROOT CAUSE of difficulties like what we saw in January.

There are some on this thread who seem to feel that the root cause
of the problem is that ITU-T is permitted to use or apply the IETF
protocol to their problem at all.

But I daresay that if we had had a good back-and-forth discussion
between the groups (with liaisons not only being considered seriously,
but responded to), that people would have understood far better what
the ITU-T was trying to do, and that there would not have been the
last minute arguments about whether ITU-T should be able to do it.

I have remarked on this before: some of our members (in ITU-T) are
interested in using the PNNI protcol to implement ASON instead of
the GMPLS protocols. The PNNI protocol also required extensions - it
did not meet the ASON requirements as-is.

I think that the primary reason that the job was so much easier with
PNNI and the ATM forum is not so much that PNNI is any better suited
to solving the problem, but that ATM forum took the liaison statements
seriously. They tried to understand the ITU-T problem and actively
tried to help solve it in the best way with their protocol.

If IETF ccamp had done the same (all the way back to Oct. 2001), the
arguments of January would never have happened.
Regards,
Steve

Scott Bradner wrote:
> 
> Steve,
>         Note that Loa is correct as to the impact of liasions, I-Ds
> or presentations by folk representing otehr SDOs within the IETF
> standards process - they carry just as much weight as the statemnts
> of any individual and there is currently no reason to have them
> treated as having some special status in the change-process ID.
> (see RFC 3356 sec 3.2.2 & RFC 2418 sec 1))  That said, I do think it would be
> good to mention that the use of liasion statements is common in other
> SDOs and that such statements should be not ignored when directed to
> IETF working groups.
> 
>         There are strong arguments that maybe the current assumptions
> should change butthis mailing list is not the place to discuss it - I
> would suggest that this dicussion be redirected to the problem-statement
> mailing list.
> 
> Scott