Sanity Check Time!

[Ed Alcoff - Prod Mtg <[e-mail]>]

OK Folks,

It's time we did a sanity check!  Cheryl and Frank made some
good points.  Pete W. seems a bit frustrated and I'm sure he's
not alone.  And hopefully, we can all hit the d key when we
see Mr. Martillo's comments.

I have been involved in the development, specification and
delivery of SNMP-based products at the same company since
Marshall and Keith wrote RFCs 1065 and 1066 when they worked
here.  Although I haven't written a single line of code, I
know what my customers/potential customers want, and this is
where I'm coming from.

When I look back over the last four years thru my SNMP-colored
glasses, it seems like ancient history because of all the
advancements and proliferation of SNMP products that we have
seen.  At Interop 88, there were no more than 13 (if memory
serves me) vendors showing anything with SNMP.  Things were
a lot simpler then; today's SNMPv2 authors pioneered the way
and they did it with little help from a very small community.

That community has become a something akin to Congress and look
what they get done;-)  Worse yet, we have developers, academics
and users all participating in the SNMP community.  I am
frequently reminded here about the Tower of Babel.  The compromising
that characterizes the IETF process has long been a strength in
bringing real world solutions to market expeditiously, has now
become something of a liability.

As a community, we came to agreement on certain goals:

1) We agreed SNMP has weaknesses and we listed many of them;
2) We agreed to evolve the network management framework to
address these weaknesses;
3) We agreed to work within a certain timeframe;
4) We agreed to use SMP as the basis for SNMPv2.

Anything else, including the proposal from Germany and Chuck's
B&B, involved methodology and design (albeit not necessarily
bad) that is/was beyond the scope of what we agreed to try
and accomplish.  We have lost sight of our original goals
and agreements.

Maybe it's time to refocus; there (listening Karl?) may
be a time when we need to tackle the entire network management
framework from the ground up, but that time is not now.  Our
customers want we agreed to; as they have learned from SNMP,
they need expanded capabilities and it is our job to specify
and deliver what they need.

And Marshall, Keith, Steve and Jeff have done more for the
Internet community than 99.9% of the people that reap the
benefits of their work.  I have oft seen them compromise,
despite believing otherwise and selflessly give of their
time and effort; yes, they have strong personalities and strong
beliefs, and they are right far more often tha anyone else.
Perhaps a little more diplomacy would have been judicious, but
that is difficult when you operate at a level above most others
(in terms of knowledge and experience) and you also have
deadlines, commitments and life to deal with.

Either SNMPv2 goes thru or we live longer with SNMPv1.  I vote
that we go forward or resign ourselves to becoming more
like ISO et al.

Sorry for the length of the post,

Ed Alcoff

std. disc.: These are my views and not those of my employer.