RE: Proposed IETF Working Group: sming

["Wijnen, Bert (Bert)" <bwijnen@lucent.com>]

All,
 
Sorry for the cross-posting. PLEASE do ONLY reply to the REPLY-TO
mailing list and not to the mibs, rap or nim lists.

We have a mailing list for the sming discussion: sming@ops.ietf.org
To subscribe, send email to majordomo@psg.com
and in the body: subscribe sming

When I started to initially respond to the email on this topic, I was
mistaken
and mentioned new-work as a mailing list where this could be discussed.
The new-work however is a read-only list for communication between 
(among others) ietf, itu, w3c, etc. The idea is that these organisations
also inform each other of new work that is being planned, so that we
can try to avoid duplication. When I started to see comments, I mistakenly
thought that people were reacting to new-work posting, but it turns out
they were reacting to the posting at IETF-Announce. Normally people
can then send comments to iesg and/or responsible ADs (and that is
true in this case also).

W.r.t to some of the issues that were brought up, here are my answers.

Q1. Why do we not hold a BOF first.
A1. An AD can always initiate new work without having a BOF.
      In this particular case, we have had many requests in the past that
      SMI needed more work. For example, we have approved RFC2856
      as a short term tactical solution under the condition that long term
      solution would be worked on. We also have seen strong support for 
      this type of work in various SNMP WG meetings when new work
      was discussed. At the NIM BOF at the last IETF, it was clear that
      many people where sceptical about complete new Network
      Information Model work, and that it needed much better definition
      on what exactly we want/need to do. However, at that same BOF,
      many people expressed that it would make sense to work on
      SMI (SMIng work from NMRG, SPPI, SMIv2 were all mentioned
      as input and concerns in this area). So that is why we are 
      initiating/proposing this new work (without having a BOF first).

Q2. Can we change the WG name, because it looks too familiar to SMICng
A2. In principle, yes we can change the name. I am not so sure it really
      is an issue... but we're open to suggestions. The one suggestion we've
      seen sofar is SMIv3.

Q3. when the charter says 'The objective is to replace both the SMIv2 and
      the SPPI with a single merged language for defining information for
the
      monitoring, configuration, and provisioning of network devices' what
does
      replace mean? Is the intended work targeting the standards track, 
      and SPPI and SMIv2 will one day become historical? 
      (I actually do not know what is the resolution about the status of
SPPI )
A3. One of the concerns I have as AD (and I guess many more people) is
      that I do not like the idea of having too many languages to do more or
      less the same thing. So "merging SMI and SPPI" seems a good thing.
      SPPI is currently in WG Last Call in RAP WG, and I expect it on my
      desk for consideration as PS soon.
      Potentially, at some point in the future, we may decide to make
      SMIv2 and SPPI historical in favor of this new work. Such a decision
      will be taken via the normal IETF process (i.e. WG deliberations
      and/or IETF Last Call). We currently have SMIv1 and SMIv2 at
      (full) Internet Standard. SMIv1 cannot be used for stds track MIBs
      any more, so at some point, I guess we want to make it historic.

Q4  The third paragraph talks about a 'transport independent information
      model'. How does this relate to NIM? NIM started to discuss about such
a
      model, and seems to have got stuck in a dispute about the language. It
looks
      like smicng has taken a shortcut and decided that it has the answer to
the
      nim dilemma and found the appropriate language that nim could not
agree
      upon.
A4. Yep, the better wording is "language" or "data model" I think. So we'll
change
      the wording a bit. And you are right, I think I came away from the NIM
BOF
      that people do support the idea to work on the language. See also my
      comments above.

Q5. These questions are intended to clarify the relationship between the
      different pieces of work in the area. I think such a work is really
needed -
      do we have the will and bandwidth to execute it?
A5. I hope the above clarifies.
      I think I have heard enough supporting statements that I trust we
      can do the work and succeed. Besides, the SNMPv3 WG will
      need to focus on just "advancing SNMPv3 to full std" and no longer
      talk about all sorts of new work. If more work is needed (like this
      proposed WG), then we need to find other places to do such new
      SNMP related work.

Q6. Should the SMIng work from NMRG be the starting point?
      Is the C-like language something we want/need?
A6. I have followed that work very closely. It is not only theoretical work,
      but it has also been implemented. Then, at the last NMRG meeting,
      I saw good stuff, also with good input/contributions from SPPI
authors.
      So... I think it makes a lot of sense to take that at least as a
starting
      point for the discussions in the WG. 
      Of course the WG can discuss the merits of this as normal. But for
      now, I think we should not just throw away the NMRG work and 
      try to understand why they made the choices that are on the table.
      
It is also my understanding from the last NMRG meeting, that we should
expect revisions of the SMIng documents rsn. So let's see what we 
find in there. Those should be used as the basis.

Hope this helps,
Bert
      
> ----------
> From: 	Dan Romascanu[SMTP:dromasca@avaya.com]
> Sent: 	Tuesday, October 24, 2000 6:15 PM
> To: 	'Steve Coya'; Cucchiara, Joan
> Cc: 	'Jon Saperia'; 'Wijnen, Bert (Bert)'; mibs@ops.ietf.org
> Subject: 	RE: Proposed IETF Working Group: sming
> 
> Steve,
> 
> You seem to be saying that new-work is a list primarily set for
> informational purposes. Jon, Joan, and myself are asking where a PROPOSED
> WG, that did not hold a BOF of itself, and did not establish an e-mail
> list,
> should be announced to and discussed by the IETF community. Bert seems to
> think that 'new-work' might be the place, while your take seems to be
> slightly different.
> 
> Regards,
> 
> Dan
> 
> > -----Original Message-----
> > From:	Steve Coya [SMTP:scoya@ietf.org]
> > Sent:	Tue October 24 2000 18:02
> > To:	Cucchiara, Joan
> > Cc:	'Jon Saperia'; Dan Romascanu; 'Wijnen, Bert (Bert)';
> > mibs@ops.ietf.org
> > Subject:	RE: Proposed IETF Working Group: sming
> > 
> > 
> > Folks,
> > 
> > For quite some time (since January 1998 if not before), the IESG has
> > posted WG Review messages to the IETF-Announcement list and to the
> > new-work list. This is done with a single message (i.e new-work does NOT
> > get advance notice).
> > 
> > The new-work list contains addresses for leaders of other standards
> > organizations. This is one of the mechanisms used by the IESG to insure
> > that we are not about to embark upon a work effort already being done
> > elsewherwe.
> > 
> > Note that these are PROPOSED WG actions, and that the IESG has NOT made
> > any decision on them.
> > 
>