IETF#61 LDAPBIS minutes

["Kurt D. Zeilenga" <Kurt@OpenLDAP.org>]

LDAP Revision WG (LDAPBIS)

TUESDAY, November 9 at 1300-1400
================================

CHAIRS:
	RL "Bob" Morgan <rlmorgan@washington.edu>
	Kurt Zeilenga <kurt@OpenLDAP.org> 

Minutes:

These minutes were compiled by Kurt from Jabber session
notes taken by Ludo.

The chair called the meeting to order shortly after 1300.

The chair provided a WG status summary.

Multiple WG documents have or soon will be progressed to the
IESG for consideration, including draft-ietf-ldapbis-models, -dn,
-url, and -filter.  The latter two require minor revision before
progression, to be done immediately after the meeting.

There is one known outstanding issue with -syntax, uniqueMemberMatch.
As this rule is owned by the ITU-T, guidance (or at least, "no objection")
is desirable.  As ITU-T has not responded to informal requests, will
try formal channels.  Chairs note that the Editor has been asked
to proposal a technical solution to this problem.  This solution
will be presented to the WG and ITU-T consideration.

There are no known issues with -user-schema.  Will likely be
progressed to the IESG soon.

There are a number of issues requiring further work with
-protocol and -authmeth.

There are no issues with -roadmap, but it will be LC'ed last.

The chairs indicated their intent to drive WG to deliver on
existing charter items quickly.  There appeared to be generally
support this could be done, and should be done.  

Roger presented a summary of changes (since last meeting) to the
authmeth I-D (see presentation).  There are five open issues.
 1) TLS mail thread
 2) TLS vs External mail thread
 3) Rework for new terminology
 4) Return explicit result code for Bind with no name or password
 5) Rename unauthenticated mechanism 
Regarding the issue 4, Jim noted that issue had already been
discussed and we should leave the semantics undefined.  The WG
appears to support this approach.
Regarding the issue 5, Kurt noted that aside from there being no
good alternative and that renaming was technically unnecessary.
Issue 3 was discussed separately (see below).  The other issues
were not discussed in any depth.   Roger estimated he would need
2-3 weeks to produce the next revision, but needed feedback on the
current release.  Chair suggested waiting for consensus on terminology
to producing next revision.  Chair noted intent to proceed with WGLC
after giving WG two weeks to consider next revision.

Kurt presented a terminology proposal (see mailing list archives)
for the LDAP session and layers within that session.  The basic
proposal included definitions for "LDAP session", "LDAP stream"
(replacing "LDAP Exchange"), "SASL layer", "TLS layer" and
"connection".  Roger noted that the proposal would allow the
term "LDAP association" to be dropped from authmeth I-D.  Ted
noted that, unless explicitly defined, "LDAP stream" might be
viewed oddly by some.  Jim suggested replacing "LDAP stream"
with "LDAP PDU layer" or "LDAP message layer".  The WG appeared
to favor the latter.  The modified proposal appeared to be
supported by consensus.  Will be taken to list for further
discussion and consideration.

Jim presented a summary of changes (since last meeting) to the
protocol I-D (see presentation).  Remaining issues include:
 - Extensibility of ASN1: defer to extensibility mechanisms
   defined in ASN1. (don't want to add own semantic). 
 - Update with recommended connection terminology
 - Removing some imperatives: referral following (removing MUSTs
   instructing clients how to consume fields)
 - Commonize adherence to data model text.
 - Result code for invalidated association. Need text to suggest
   Notice of Disconnection ? Need to prescribe specific error ? 
 - Undo some changes to strongAuthRequired in regards to
   Notice of Disconnection.
While the WG rehashed some of the list discussions, the general
sediment as to how these issues should be resolved appears
consistent with that on the list.  The Editor is to prepare
a revision addressing these issues and that revision will likely
be then be progressed to the IESG.

The chairs briefly went over next steps.  They are: protocol
progressed by end of November, authmeth WGLC in December,
then Roadmap in January.

There were no open mic discussions.

The session was concluded shortly before 1400.