Folks, Below are the minutes from the SNMP audio call held on 5/13. We will convene to discuss the action items on 5/20 from 11:00 - 1:00 PM EST. Please let me know by end of day Tuesday 5/19 if you plan to call into the bridge and need a phone number. Thanks, George Kajos VideoServer 63 Third Avenue Burlington, MA 01803 tel: (781) 505-2193 fax: (781) 505-2101
--------------------------------------------------------------------------
SQ16 Q14 H.Multimedia MIB Audio Call 5/13 Minutes
Agenda 1. Introductions - identify company, area of interest, likelihood of a submission to the June ITU-T SG16 Q.11-14 Rapporteur's meeting in Cannes, France 2. Discussion of the April SG16 Q.11-14 Rapporteur's meeting in Japan and agreements identified in the Q.14 trip report (TD16, sec 6.2.3) 3. Discussion of the enclosed hierarchy document 4. Proposed changes to APC-1347 and APC-1349; discussion of all the APC contributions 5. Organization of the H.Multimedia MIB document 6. IETF relationship 7. Planning for future meetings and division of labor
1. Attendees: Mike Thatcher, Irene Kuffel, Dan Klenke, Cisco Mark Baugher, B.Strahm Intel Joon Maeng, Vtel David Walker, Wade ?, Mitel Orit Levin, Radvision Chris Gindel, PakNetX Don Dewar, George Kajos, David Reich, Irina Suconick, Santo Wiryaman, VideoServer
2. Temporary document 16, sec 6.2.3
The Q.14 trip report was read out loud and discussed. TD-26 is a liaison to SG4 to determine if and how to add the H.Multimedia MIBs under the CCITT tree. It was questioned whether any published MIBs are under trees other than {iso (1) org (3) dod (6) internet (1)} and, if so, where they are published. The security issue was restated without discussion. It was pointed out that the MIBs developed in this work need the ability to be proxied for use by TMN (telecommunications management network).
3. Hierarchy document
The updated hierarchy was discussed. While it was generally agreed to, doubt was expressed as to whether there would be any contributions to the media codecs and T.120 areas before September 1998.
4. How to reduce overlap with Gatekeeper and RAS MIBs
Mike Thatcher discussed that he split out the H.323 textual conventions into a separate document so all other modules can import it. Irina Suconick and Mike Thatcher discussed their updates to APCs 1347 and APC 1349. Principle decisions to reduce the overlap included 1) the gatekeeper MIB would rely on the registration and admissions table in RAS MIB. Most Gatekeeper tables except the Zone and Segment moved to RAS MIB The Zone table indexed by unsigned integer instead of by name to keep OID smaller. Gatekeepers will have information on their own zone(s) and zones of other gatekeepers that a gatekeeper knows about. There was a great deal of discussion about the indexing scheme for the admission table. One concern was that 5 indices were too large to fit in as a 128 byte OID. Another was that without the Revision 2 CallID, the indices do not provide uniqueness. The use of CallID was debated because it is unavailable to version 1 gatekeepers. Additional input was that there should be index correlation between the Admission Table and the CallSignalling MIB Call Table. However, there is no requirement that a direct call model gatekeeper support H.225 Q.931 making this somewhat difficult. On the topic of correlation between the Zone Table and Registration table, Mike thought that an indirect correlation existed. Others disagreed. Another concern was that most tables need to have read/write access. The SNMP V2 rowStatus technique was pointed out as a mechanism to allow administrative static entries into registration and other tables.
4a MIB counter32 discussion
A more general comment was that there are too many counters. George asked that she specify which counters need to be removed. The response was that all global counters should be removed as it is no longer standard practice. George strongly disagreed and pointed out that among the reasons for the MIB effort is fault isolation and performance measurement. Most of the SNMP V2 RFC MIBs have *counter32* for this reason. Resolution of this issue will be done by discussing each counter and justify its existence or removal by group consensus.
5. Organization of the documents
George wants H.Multimedia management document to describe goals, conventions, relation to other work and make the MIBs individual annexes. The SG4 liaison requests input on this issue. The structure of the H.Multimedia MIB document will be a topic at the June SG16 meeting. In particular it was thought that a proposal should be generated and presented to Mr. Bigi.
6. IETF relationship
The group expressed a preference for conducting the work in the ITU rather than the IETF. However, the overall study group agreed to have the work examined by the IETF to determine if there was any interest. This will still be the plan of record. If reaffirmed in the June ITU-T SG16 Q.11-Q14 meeting, then we will ask for IETF input during the August IETF meeting. It was reaffirmed that the work would use the SMI from SNMP V2.
7. Next Meeting - Audio conference Wednesday, May 20, at 11:00-1:00 PM Eastern Standard Time. Send email to gkajos@videoserver.com for inclusion.
Action Items 1. George Kajos - Publish meeting minutes (this report) and schedule next conference. 2. George Kajos - Confer with ISO network management experts regarding placement of H.Multimedia MIB. Check with Mr. Thom on status of liaison to SG4 discussing topic. 3. Group - Research us of Global counters as they relate to usefulness in fault isolation and performance monitoring in network management applications. 4. Group - Try to achieve resolution on Admission table indices. 5. Group - Resolve correlation of indices between Admission and CallSignalling Tables. 6. Group - Resolve correlation of indices between Registration and Zone Tables. 7. Michael Thatcher - Send revised APC-1349 and textual conventions MIB to mailing list. 8. Irina Suconick - update APC-1346, APC-1347, and APC-1348 to use textual conventions described in document sent to mailing list by Michael Thatcher. 9. Joon Maeng - Reflect meeting minutes to Q.11 mailing list and request contributions for H.320 MIBs.
Minor correction on Georges minutes
- How to reduce overlap with Gatekeeper and RAS MIBs
Mike Thatcher discussed that he split out the H.323 textual conventions into a separate document so all other modules can import it. Irina Suconick and Mike Thatcher discussed their updates to APCs 1347 and APC 1349. Principle decisions to reduce the overlap included 1) the gatekeeper MIB would rely on the registration and admissions table in RAS MIB. Most Gatekeeper tables except the Zone and Segment moved to RAS MIB The Zone table indexed by unsigned integer instead of by name to keep OID smaller. Gatekeepers will have information on their own zone(s) and zones of other gatekeepers that a gatekeeper knows about.
A GatekeeperIdentifier has a max length of 128 and is too big to be used as an index.
participants (2)
-
George Kajos
-
Mike Thatcher