Re: STDCONF WG charter proposal

At 12:49 PM 9/26/2002 -0400, Margaret Wasserman wrote:

>Hi Andy,
>
>In general, I think that your proposal is excellent.  I have some
>general questions/thoughts, though...
>
>Do you plan to hold a BOF for this effort in Atlanta?  Or are you
>hoping to move directly to a WG?

I was hoping to move directly to WG.  Maybe Bert and Randy can
provide some guidance here, as to whether a BOF is needed first.


>How could we avoid dead-lock on the overall approach for configuration?

good question.  I am hoping to get (and keep) operators
involved in the process.  We need to recognize that not
all network environments are the same. We need to resist
the temptation to attempt a one-size-fits-all solution.


>Are you expecting this to be a requirements-driven effort, or a
>solutions-driven effort?  Maybe some of both?

I think it's important that we create a comprehensive road-map
for NM standards development in the IETF.  We should be willing
to let go of old baggage if it doesn't meet the needs of
most operators.  We should determine what the next generation
NM technology should be, and plan for coordinated phased deployment 
over the next few years.  It may turn out that the recommendations 
of the XMLCONF BOF (e.g., transport first) should be carried out,
but this should be done as part of a well thought out plan.

How do we lower device management costs? How do we improve
device API consistency and stability? How do we handle
multi-box transactions? How do we create standard schema
that covers all common features? How do we get vendors
to implement these standards in a timely fashion? How
do we co-exist with the CLI? How do we integrate with
DB driven provisioning systems?

I'm sure there's lots more of these questions to answer.
That would be the first job of the STDCONF WG.  The actual
technology development may be done in other WGs. Some
technology may already be done -- we don't want to
reinvent pieces that are already done.

I think the WG should focus on technology solutions
that are available now (or could be available soon).
IMO, that makes it a job for the IETF, not the IRTF.


>Margaret

Andy