RE: Second incoming communication from the OIF

[Kireeti Kompella <kireeti@juniper.net>]

Hi Lyndon,

On Tue, 23 Nov 2004, Ong, Lyndon wrote:

> I'm reading into your email a concern with the use of "ASON" in the title
> of the project.  The interworking project is likely to cover ITU-T
> G.7713.2 as the ASON signaling standard at the UNI and E-NNI, taking into
> account experiences that members have had implementing and testing
> protocols.

Our concern is with the project as a whole.  There are several
deficiencies in G.7713.2 and the OIF UNI; these have been pointed out
many times by CCAMP, and liaised to the ITU, but we are not aware of
any significant changes that have been made to the OIF UNI or to
G.7713.2 to address these problems.  In view of this, it seems
premature to expend effort attempting to build networks that interwork
the OIF UNI with GMPLS since future work to fix the OIF UNI is likely
to have considerable impact on any interworking specifications.  The
real problem (namely, the ASON signaling requirements as well as a UNI
and E-NNI) have also been addressed by CCAMP; to continue work on
G.7713.2, and in fact to use it as the ASON signaling standard at the
UNI and E-NNI seems to us very shortsighted indeed.

The communication says, "We expect the OIF, ITU-T and IETF will
continue to work together to minimize or eliminate differences between
control plane signaling protocols."

Certainly a laudable goal.  Persisting with G.7713.2 and the OIF UNI
without any willingness to adapt these specifications clashes with
this goal.  Working together implies converging -- the IETF took a big
step in this direction by acceding to the ASON signaing requirements
as stated; it doesn't seem that the ITU-T or OIF are willing to take
similar steps.

> As you note, GMPLS is being extended to support ASON
> as well.  It sounds reasonable to comment that some clarification of
> the project definition may be helpful.

And very importantly, to state the need for more work on this front,
rather than beginning work on convergence at once.

> This also raises a good point about GMPLS continuing to evolve beyond
> RFC 3473.  As the work on GMPLS extensions to support ASON functionality
> continues to develop, this will need to be taken into account.

At what stage?  It seems that instead of converging, the object seems
to develop a full suite of competing protocols, and only then take the
GMPLS suite into account.

Kireeti.
-------