RE: MPLS OAM & the IETF

["Joel M. Halpern" <joel@stevecrocker.com>]

Having watched some of the discussion in SLC, and much of the discussion on 
this list, a couple of things were and are clear.

There are two different sets of objectives.  Sharam and Neil want a 
comprehensive OAM solution.  In contrast, Ron and others are looking to 
define a simple tool.  It was agreed in Salt Lake that these are different 
tasks.

My understanding is that the CCAMP working group has agreed to undertake 
the task of defining a simple tool for a simple problem.  The ADs have 
agreed that this is within scope.

The debate then should focus on the documents for that simple problem.  One 
should have, at this point, a relatively small number of clear choices:
You can argue that the working group should not define a solution for the 
simple case.  I thought there was consensus to solve this, if there really 
isn't then we should not have a working group document.
You can argue that the description of the problem in the proposed 
requirements draft is completely off base.  In that case, given the stage 
of development, I think you should bring in a draft as to what the 
requirements you do see FOR THE SIMPLE CASE.  A complete draft about what 
the full set of OAM needs for an operator are does not address the question.
In the absence of an alternative draft, and given the consensus to work on 
the simple case, we should take the starting point we have, and fix 
whatever problems there are.  Taking it as a working group document is not 
a vacuous action, but it does not mean "it is done, start last call".  It 
means this is a decent starting point.  (And having a starting point is the 
only way to make progress.)

Yours,
Joel M. Halpern

At 08:05 AM 3/7/02 -0800, Shahram Davari wrote:
>Giles,
>
> > Agreed.
> >
> > I'm not sure which of Scott's options this corresponds to, though!  I
> > don't want to see the IETF working on MPLS OAM as per option 2, but
> > option 1 seems to be a step down the road of allowing the ITU to
> > "co-own" MPLS?
> >
>
>It seems that developing non-IP OAM tools for MPLS is not in the spirit of 
>IETF. However, it seems that many operators want functions more than ping 
>and trace-route for their MPLS networks that are used to carry their 
>legacy traffic.
>
>Y.1711 and future ITU recommendations would possibly use MPLS data-plane 
>to carry the MPLS OAM packets, but that doesn't mean that ITU will co-own 
>the MPLS. Everybody knows that the MPLS data-plane as well as MPLS 
>IP-based control-plane and MPLS IP-based management-plane are owned by IETF.
>
>Besides any need by ITU for extensions to IETF standards will formally go 
>through IETF channels, and therefore be under IETF control.
>
>Therefore I think if there is a clear delineation of work and procedures 
>for the cooperation, then most of your concerns could be addressed.
>
>Yours,
>-Shahram
>