2.1.32 LDAP Service Deployment (lsd) BOF

Minutes of the LDAP Service Deployment (LSD) BOF

Reported by: Chris Apple and Roland Hedberg (with input from Ryan Moats' meeting notes)

The LSD BOF was attended by 96 people. We discussed the status of current work and changes to the proposed LSD WG charter.

We noted that there seems to be some problem with the mailing list archive. Roland will investigate and fix the problem.

Current work includes the following documents:

The original locating LDAP servers I-D was split into two I-Ds during October, 1997. One draft deals with LDAP clients finding LDAP servers and the other deals with LDAP servers finding other servers. There were few comments that requiring changes to these I-Ds posted to the mailing list. BOF-consensus was to request a last call on the documents after the LSD WG forms officially. The two new I-Ds are:

The minimum white pages schema draft was written and posted to the mailing list, but it was not published as an I-D prior to the meeting. As a result, few people had had the chance to read the document in detail. Regardless, we discussed several issues related to a core concept of the document: attribute labeling. We agreed to discuss issues such as how labels affect DNs, naming attributes, and backwards compatibility. Other issues, such as schema mappings for backwards compatibility with implementations making use of existing white pages schema elements and why this new proposal makes use of attribute labels rather than X.500-style attribute options, were raised and will also be discussed on the mailing list. A request to include the concept of priorities in this schema proposal was made by Peter Jurg and it was agreed that we would discuss this on the list. This concept is likely to be viewed as a type of label. There was also a question about the new syntax definitions presented in this document and whether or not they as well as the entire contents of this document should be merged into the main LDAP schema document produced by the ASID WG. The resolution was that while syntax definitions from this document may eventually be included in an update or revision of the main LDAP schema document, this document would go forward as a proposed standard when it is ready for last call. This document was posted to the mailing list as draft-apple-ldapv3-schema-wp-00.txt and was to be published as an official I-D under the same name after the meeting.

The naming and interconnection guidelines document is currently being written and is targetted for publication of an I-D in January, 1998. This document will explicitly recognize, explore, and clarify issues that are basic to name resolution in a loosely-coupled flock of LDAP servers that make used of two different naming schemes: X.500-naming and dc-naming. How this document relates to the DN requirements document metioned above as well as the IDS naming plan document was flagged as an issue to discuss further on the list.

The schema writer's guidelines is currently being written and is targeted for publication an I-D in January, 1998.

We discussed several proposed changes to the draft LSD WG charter.

Editorial changes to the charter are required to reflect splitting the original "locating LDAP servers" draft into two I-Ds. Other editorial changes need to be made to reflect new target publication dates for other deliverables.

The schema mappings document proposed in the pre-Washington draft charter was extremely difficult to scope in a way that would make it useful for readers. The option of removing this document as a deliverable was presented as a question by the BOF co-chairs and the room was polled for consensus. After changing the proposed charter modification to include retention of the concept of schema mappings in the charter text and to deal with this concept in constructive ways in other, existing documents (such as the minimum schema document and the schema writer's guidelines document), the room was polled again. BOF-consensus was that this would be acceptable. BOF-consensus will be verified on the mailing list.

We discussed the possibility of adding another deliverable to deal with collecting and gaining concensus on requirements for distinguished names. A proposal for this deliverable has been written by Jeff Hodges and published as the I-D draft-hodges-ldap-dir-dn-reqs-00.txt. Also proposed was the possibility of merging this document into the naming and interconnection guidelines document. This will also be discussed on the list, however, BOF-concensus was that it would be a better idea to leave it as a separate document. There were a few questions related to the content of the document which the co-chairs requested be addressed in detail on the mailing list. These questions were related to two very controversial concepts: using URLs as a DN and "why don't you just mandate dc-naming?" To avoid opening a large can of worms any further, Chris Apple requested that we _not_ discuss the second concept, at least not in the BOF room, and preferably not at all since we are, as a WG, proposing to explicitly recognize that there are two naming schemes in use in the real world today.

A few questions unrelated to the proposed LSD WG charter were discussed near the end of the meeting. These questions were related to writing and reviewing LDAP schema for use in PKIX-based applications and services. The first question (from the PKIX WG) was about how review of schema would occur prior to being published in the schema listing service repository. The answer (given by Harald Alvestrand) was that this would be determined in the SCHEMA WG and that when there are specific operational needs, that we might consider publishing standards-track schema at that time. The follow-on question (also from the PKIX WG) was that if the PKIX WG thought the time was right to publish a standards-track PKIX-specific LDAP schema document, would the PKIX group be the right place to do the work. Harald confirmed that this work should be done within the PKIX WG and submitted as a schema listing request to the schema listing service review team (details of how the review team operates is to be decided by the SCHEMA WG).

Another comment was that we need some guidelines about which schema should be developed by the IETF and which should not. Harald's opinion was that if there is an IETF standards track document that requires a schema for deployment, then that schema should be standards track. For things that are not on the standards track and do not have a wide impact, a standards track schema is not necessary. The cases that fall somewhere between these two should be decided on a case-by-case basis.

Harald asked a closing question in which he requested that if there are any additional work items or concepts that this working group should address, that they should be posted and discussed on the mailing list. If concensus indicates so, these new work items or concepts should be added to the charter REALLY SOON.

Roland Hedberg gave a quick overview of the Swedish TISDAG project (a report on the project is currently a proposed work item in our draft charter). TISDAG will use CIP to index Swedish WHOIS++, X.500, LDAP, etc. directories and make them searchable in a slightly-lossy-client-independent way. A pilot will start in the next couple of months and the project report work item will document problems uncovered during the pilot.

A small, unofficial BOF meeting of people interested in LDAP-related piloting was called and they met to discuss what they'd like to see happen. If there was anything about the draft LSD WG charter that needed to change as a result of this piloting discussion, participants were to post suggestions to the LSD mailing list.


None Received

Attendees List

go to list

Previous PageNext Page