home *** CD-ROM | disk | FTP | other *** search
-
-
- CURRENT_MEETING_REPORT_
-
-
-
- Reported by Amatzi Ben-Artzi/ 3Com
-
- Minutes of February 22, 1989
-
- The Lan Manager MIB working group met Wednesday, 2/22/89 in Sunnyvale,
- CA for our first meeting. The meeting was very productive and generated
- a long list of output and action items. Below is a summery of the
- meeting and major decisions reached.
-
- The minutes cover seven topics:
-
- 1. Introduction
- 2. The Group's Objectives
- 3. To Proxy or Not?
- 4. Initial Documents and Editors
- 5. Relationship to the MIB WG
- 6. Mailing List
- 7. Timetable
-
- Introduction to the Lan Manager.
-
- Amatzia gave a short introduction on the Lan Manager. The emphasis is
- on the management interoperability issues between the Lan Manager as a
- standard in the workgroup environment and the standards being developed
- in the ``enterprise" environment. As both are based on the TCP/IP it is
- important that they can cooperate. Microsoft offered to provide a more
- extensive overview of the LanManager if people will find it useful.
- Please send me feedback on this one!!
-
- Group's Objective
-
- The objective of the group is to define the MIB (and relevant related
- mechanisms) needed to allow management overlap between the workgroup
- environment (Lan Manager based) and the enterprise environment (based on
- TCP/IP management).
-
- We found that it translated into four basic areas:
-
-
- 1. Define a set of management information out of the existing Lan
- Manager objects to allow for useful management from a TCP/IP based
- manager.
- 2. Define extensions to the TCP/SMI where appropriate.
- 3. Develop requirements for additional network management information,
- as needed, and work to extend the Lan Manager interfaces to support
- such information.
- 4. Define the mechanisms of exchange of management information between
- clients and servers so that proxies can be developed.
-
- 1
-
-
-
-
-
-
- Proxy or Not?
-
- We concluded that the manager need to have access to the management
- information in every client in a direct way. It amounts to the
- following:
-
-
- o A client may have a Management/TCP stack.
- o A client may be supported by a server that acts as ``proxy'' on his
- behalf.
- o The manager need not be aware of which one of the techniques above
- is being used in the workgroup.
-
-
- However, we recognized that in the proxy mode, the server may have two
- types of objects: server resident, and client resident. For the second
- type, a server-client mechanism has to be developed to allow the
- implementation of a proxy.
-
- Initial Documents and Editors
-
- The following documents have been identified as needed. Initial editors
- were also selected.
-
-
- o SMI Extensions: Pranati Kapadia, HP
- o MIB Objects: Jim Greuel, HP
- o NDIS Extensions (not assigned)
- o Transport APIs for NM Information access (not assigned)
- o Client-Server management protocol (Amatzia Ben-Artzi, 3Com)
-
-
- Relationship to the MIB group
-
- A real objective of this MIB group is to work under the ``BIG'' MIB
- group. One implication was that the MIB specification should follow the
- 1066 RFC (specifying all attributes as ``objects'') with an appendix
- that actually describe the containment relationship (Same technique that
- was used in the CMOT RFC to re-state the supported MIB)
-
- A big question mark is SMI. Can we live with the guideline of ``no SMI
- extensions'' ? We shall address it when the first required extension
- shows. We do know, however, that EVENTS (or alerts, or alarms) are a
- big issue, but we where not sure if this was an SMI issue or what.
-
- We also feel very strongly that the recommendation of the previous MIB
- group should be followed: Lan Manager should be assigned a number of
- the MIB (Like TCP, IP, or CMOT) and define it objects under this branch.
- Then, bring them forward to the larger group for discussion and
- approval. ONLY EXPERIMENTAL OBJECTS SHOULD BE PLACED UNDER THE
- EXPERIMENTAL BRANCH.
-
-
- 2
-
-
-
-
-
-
- The whole branch is OPTIONAL, so people who don't implement it do not
- have to worry about conformance. It seems like we would want, for
- simplicity of conformance, at least initially, to say: the branch is
- optional, but if you implement it, it is ALL mandatory. The target is
- roughly 20 - 30 objects initially.
-
- Mailing list
-
- Welcome a new mailing list:
-
- LanManWG@spam.istc.sri.com
-
- As usual, there is also a LanManWG-request.
-
- Timetable
-
- March 17: First draft proposal for MIB objects in the mail.
-
- March 31: Comments are due back to the editor
-
- April 11-14: IETF meeting. We shall meet sometime there to discuss the
- proposed draft (currently planned as a half-day meeting)
-
- Thanks to all the participants for a very effective meeting.
-
- Attendees
-
- aguilar@istc.sri.com SRI
- jimg@hpcnd.hp.com H-P
- ...!ucbvax!mtxinu!excelan!ramesh Excelan
- ...microsoft!henrysa Microsoft
- geo@ub.com Ungerman-Bass
- kapadia@hpda.hp.com H-P
- davep@esd.3com.com 3Com
- jcham@mbunix.mitre.org Mitre
- Hunter@nmfecc.arpa Laurence Livermore Lab
- ...!ucbvax!mtxinu!excelan!pramod Excellan
- jonab@cam.unisys.com Unisys
- kzm@twg.com The Wollongong Group
- amatzia@spd.3com.com 3Com
-
-
-
- 3
-