Hello folks, Attached are the minutes of the 7/27/98 audio call. Our next call is Wednesday 7/29/98 at 11:00-1:00PM EST. Please let me know if you wish to have a number for the call. Thanks, George Kajos gkajos@videoserver.com VideoServer, Inc. 63 Third Avenue Burlington, MA. 01803 781-505-2193 (tel) 781-505-2101 (fax)
SQ16 Q14 Multimedia Management Information Base Audio Call 7/22/98 Minutes
Agenda 1. Review of Zvi Mizrahy*s email regarding gateway MIB proposal 2. Reiew of Orit Levin*s document discussing H.323 MIBs 3. Discussion of MIB hierarchy picture from Joon Maeng 4. IETF BOF
Participants: Mike Thatcher, Cisco Zvi Mizrahy, Prina Vortman, IBM Haifa Orit Levin, RadVision Bill Strahm, Mark Baugher, Intel Nicole Gallant, Nortel Don Milotte, James Pansarasa, Picturetel Joon Maeng, VTEL Jackson Wynn, White Pine George Kajos, Irina Suconick, Don Dewar, VideoServer
H.323/H.320 Gateway We reviewed Zvi*s mail message. The points are repeated and annotated below: 1) The Gateway MIB will include the following subtrees : h323gwSystemDescription (new) h323gwCapabilities (new) h323gwConfiguration h323gwStatistics h323gwCalls h323gwConnections h323gwNotifications h323gwCommands
We also suggest to add an empty subtree for private vendor information.
There was general agreement to the hierarchy, except in the h323gwCommands subtree. Here there was some discussion about what is appropriate for standardization and what should be vendor specific. The approach seemed to be to limit the tree to general commands relevant to a gateway application, but people probably need to see more details. It was pointed out that the MIB could contain an OID which could point to a vendor specific MIB as part of the standard MIB.
2) The structure of the MIB should be flexible as possible to allow future extensions, for example , plugging in other types of endpoint protocols. The concept of "call connections" suggested in the proposal supports this.
There was agreement to this principle, after a review of the relationship between calls (end to end) and connections (endpoint to gateway) in terms of scope.
3) The "IP" connection type should be changed to "H323" connection type. Some of the variables defined by this connection might overlap with other MIBs such as the RTP MIB. This is still opened for discussion.
There was agreement to this principle
4) The MIB should contain an "H320" connection type in addition to the "H323" and the "PSTN" connections. A point to point call is represented by an H323 connection + either an H320 or PSTN connection. In addition, each call has also some common information . We have also suggested a 'sub connection' entity that allow to distinguish between audio, video or fax information for each connection type. IBM Haifa will appreciate contributions for the H320 connection type.
There was agreement to this principle. VTEL and VideoServer agreed to help with the H.320 representation.
5) The "storage connection" should be taken out, possibly to a private vendor information.
6) The Hardware group will be taken out to a private MIB, as this represents a vendor type of information.
7) The issue of the location of the gateway MIB has been agreed. It will be placed under the H323media MIB.
There was agreement to points 5-7.
8) The issue of representing conference (multipoint) calls in the gateway MIB is still under debate. IBM is convinced that multiway calls should be able to be described in the gateway MIB despite the fact that the MCU is responsible for these types of calls.
This is still an open issue. While most agree with this view, the general architecture of a gateway needs to be finalized. This was discussed in much greater detail in the minutes from the 7/08/98 call.
9) The gateway MIB will use conventions or 'enumerated types' that should be added to the textual conventions MIB proposal ( TBD).
There was agreement to this. Zvi should send Mike Thatcher proposed changes.
H.323 MIBs We reviewed the H.323 MIBs by the following the outline presented in Orit Levin*s document. From section 2.1 H225-MIB, we agreed to change the indices of the h225ConnectionsTable to h225SrcCallSignallingAddressTag H323TAddressTag, h225SrcCallSignallingAddress Taddress, h225CallId H323GlobalIdentifier
and agreed to add clarifying text that in order to support H.323v1 instances, h225CallId should be filled with default value ZERO of an appropriate type and length.
From section 2.2 RAS-MIB we agreed to change the indices of the
RegistrationTable to: RegistrationSrcCallSignallingAddressTag H323TAddressTag, RegistrationSrcCallSignallingAddress Taddress,
From section 2.3 RAS-MIB we agreed to change the indices of the
RegistrationAliasTable to:
RegistrationSrcCallSignallingAddressTag H323TAddressTag, RegistrationSrcCallSignallingAddress Taddress, registrationAliasIndex Integer32
From section 2.4 RAS-MIB we agreed to change the indices of the
AdmissionTable to: admissionCallSignalingAddressTagH323 H323AddressTag, admissionSrcCallSignallingAddress Taddress, admissionDstCallSignallingAddress Taddress, admissionCallId H323GlobalIdentifier, admissionAnswerCall Boolean and greed to add clarifying text that in order to support H.323v1 instances, h225CallId should be filled with default value ZERO of an appropriate type and length. Orit volunteered to supply a description of the uses of the callSignalling addresses in the various tables; for instance when its used as a listen port and when it*s a dynamic port.
The discussion around section 2.5 H.245-MIB centered on combining the various tables into a single table and how to refer them back to the H.323 call. There was general agreement to the idea of using a single H245CallIndex. Further discussion was focused on how to represent the H245 capabilities. Orit agreed to try to draft an example of how she would represent the capabilities.
Joon Maeng*s Hierarchy picure Joon sent a hierarchy diagram to the mailing list. In principle, people tended to agree with the top level view presented in the picture for how to represent a gateway and MCS, but the meeting was near closing and we agreed to take up again the following week. IETF Our request for a BOF at the Chicago, IETF meeting (8/24-8/28) has been accepted by the IETF Operations and Maintenance Area Directors Bert Wijnen and Harald Avelstrad. Next Meeting: Wednesday, July 29, 1998 11:00 - 1:00 PM EST. Please send mail gkajos@videoserver.com for a bridge port number.
Current Action Items: 1. George Kajos - Publish meeting minutes (this report) and schedule next conference. 2. George Kajos - Confer with ISO and IETF network management experts regarding placement of H.Multimedia MIB. 3. Group - Admission table indices - See updated RasV2Mib.rtf for CallID change and text in report 4. Group - Admission and CallSignalling Tables index correlation. - closed 7/22/98 5. VideoServer - Generate MC and MP proposal - See h323MCmib.rtf on Picturetel site. 6. George Kajos - Generate BOF request to O&M Area Director - Sent on 7/13/98; accepted 7/27/98. 7. George Kajos/ Mark Baugher - Set up a mail reflector which can be used for IETF input and access - George does not think VideoServer will be able to do this and Mark Baugher said he would look into whether Intel can take this on. - Intel created HMIBS on 7/17/98 8. Contributors - place any updated the MIB proposals on the Picturetel reflector. The current status is: Avc-site\9806_Can\APC-1380 - overview Avc-site\9806_Can\APC-1387 - textual conventions Avc-site\9806_Can\APC-1388 - RTP Avc-site\9806_Can\APC-1389 - H.323 Call Signaling Avc-site\Incoming\RasV2Mib.rtf - H.323 RAS Avc-site\9806_Can\APC-1391 - H.323 Terminal Avc-site\9806_Can\APC-1392 - H.323 Gatekeeper Avc-site\Incoming\APC1393 - Gateway Avc-site\Incoming\h323MCmib.rtf - H.323 MC Avc-site\9806_Can\APC-1396 - H.320 Terminal Avc-site\9806_Can\APC-1397 - H.320 Call Signaling Avc-site\9806_Can\APC-1398 - H.221 Avc-site\9806_Can\APC-1399 - H.320 Multipoint Control Unit Avc-site\9806_Can\APC-1400 - H.245
The working proposal is that any updates will be placed in incoming by identifying name and Mr. Okubo and the mailing list will be informed. 9. Group - Gateway - we have to decide the architectural model for a gateway, what are the standard pieces, should multiway connections be modeled, etc.? 10. Group - Review all MIB proposals - ongoing 11. Orit Levin - send MIB critique to the mailing lists - done 7/21/98 12. Zvi Mizrahy - send update to the gateway proposal to the mailing list - done 7/21/98 13. Group - Review MC MIB