Proposed LDUP WG Meeting Minutes

["Chris Apple" <capple@xxxxxxxxxxxxxxxxxx>]

Please post comments and corrections to the list. The meeting minutes are
due on December 16th to the secretariat. Therefore, we can only accept
comments and corrections posted to the mailing list prior to
1700 US Eastern Time.

LDUP WG Meeting Minutes
November 20, 2002

We initiated a chat conference with remote attendees using
the facilities provided.

Roger G. Harrison volunteered to take notes for the meeting
minutes.

The co-chair in attendance, Chris Apple, spoke on behalf of
remote attendees and displayed the chat conference room window
on the overhead screen while doing so.

Some WG members in the meeting room were also in the chat
conference room and made comments there in response to
comments from remote attendees.

We reviewed the agenda and there were no changes requested.

Chris Apple gave a brief overview of current WG deliverables
status. The requirements document has been published as an RFC
LCUP's WG last call concluded and a substantial number of
issues were raised. The document editors, the remote WG
attendees in this case, are working to resolve the issues.
LCUP will go through a second WG Last Call after it is
revised.

Kurt Zeilenga led a discussion about eventual convergence
for LCUP. The current proposal from WG fails to provide
eventual convergence of updates made to server at client.
Kurt believes that eventual convergence is a strong
requirement based on conversations with authors of LCUP.
Kurt and colleagues have developed a protocol that they
believe does provide convergence, but it is too chatty.
Question: can we solve problem and still meet other LCUP
Requirements. In particular, can we do it without maintaining
session state? The general belief is no.

Current LCUP eventual convergence status: we understand the
problems and have explored ideas to solve them, but the feeling
is that we're not at the end of LCUP effort, we're in the middle.
There are still significant engineering issues to be resolved.

There was some discussion held during this LCUP topic that is
reflected in the chat conference room archive. The discussion
that still needs to occur on this topic is largely in the
technical details. Chris Apple recommended that the discussion
Be moved to the mailing list. Mark Wahl requested that the LCUP
authors publish a revised specification before the next IETF
meeting.

We acknowledged that there have been rumblings of concluding WG.
Chris Apple has talked with several people to get perspectives
and proposed a path to conclusion of work for WG to consider.
The proposal was reviewed by John Strassner and Patrick Faltstrom.

The presentation covered the state of various WG deliverables
and followed that by a proposal from the co-chairs to proceed
with an aim of publishing LCUP as a standards track document
and all other documents as either Informational or Experimental
depending on the nature of their content. There was some
discussion about the usage profile document being more
appropriate as an Informational document than Experimental
as proposed by the co-chairs. The general feeling of the room
was that Informational would be more appropriate for the usage
profile document. The proposal also mentions referencing
X.500 BAC and/or the recent I-Ds written by Kurt Zeilenga
and Steven Legg related to the X.500 administrative model.
The WG chairs also proposed that they would request that
the LDAP Directorate to review those I-Ds from Kurt and
Steven with that context in mind. There was some concern
expressed over giving more weight to the recommendations of
the LDAP Directorate than the rest of the WG. These concerns
were discussed and resolved by the co-chair emphasizing that
Kurt Zeilenga was correct in his observation that the LDAP
Directorate not receive more weight than the WG members
in evaluation of consensus.

It was pointed out by Chris Apple that volunteers are willing
to continue and/or take over editing documents. The document for
which there is no confirmed volunteer for editor is the main
LDUP protocol specification. The current editor of that document,
Tim Hahn recommended that John McMeeking take over. The co-chairs
will approach John about doing this.

The co-chair's proposal also indicates a target date for
completing all documents by June 2003. Completion in the
context of this proposal means the documents are ready to
enter WG last call or have already completed WG last call.
There was some concern expressed by Ed Reed over the normal
use of a WG Last Call in relation to LDUP documents in their
present state because he doesn't believe that this could be
accomplished by that date. In particular, Ed sees the
administrative and access control models as potential deadlocks.
Chris Apple replies that we would attempt to use external
documents for access control and that the administrative
model proposal was on the table in the form of an individually
contributed I-D. Mark Wahl asked if there would be more than
one vendor who is willing to implement in a way that
implementation interoperability can be tested.

Chris Apple responded that we don't have a clear answer and
that this is the primary reason for suggesting experimental
publication for most LDUP documents. Perhaps the best thing
is to publish the experiment and let the market decide the
answers. Ed Reed then proposed that the documents simply be
frozen and published as Experimental. Chris Apple responded that
this was more or less what he hopes to accomplish but wants
to have a last chance to remove gross inconsistencies before
publication.

Mark Wahl made an alternate proposal for concluding LDUP by
suggesting that the WG consider putting drafts back to individual
submission status as was done in LDAPEXT when it was concluding.

Chris Apple concluded the meeting by indicating that both the
proposal from the co-chairs and the proposal from Mark Wahl be
discussed on the mailing list. If consensus can be established
on one of the proposals, the co-chairs will take appropriate
actions to implement that consensus. For the case of the
co-chairs' proposal, this means revising the LDUP WG Charter.
For the case of Mark Wahl's proposal, the WG simply concludes
after attempting to achieve consensus on a revised LCUP
specification and does not attempt to do more work as a
chartered IETF group on other LDUP specifications.

Chris Apple - Principal Architect

DSI Consulting, Inc.

mailto:capple@xxxxxxxxxxxxxxxxxx

http://www.dsi-consulting.com