Hello folks, The next SNMP audio call is 7/15/98 (Wednesday) at 11:00 AM EST. If you'd like a number please send email to: gkajos@videoserver.com.
Thanks, George Kajos VideoServer, Inc 63 Third Avenue Burlington, MA 01803 781-505-2193 (phone) 781-505-2101 (fax)
Below are the minutes of the 7/8/98 call:
-------------------------------------------------------------------------- -----------------
SQ16 Q14 Multimedia Management Information Base Audio Call 7/8/98 Minutes
Agenda 1. IETF submission/coordination 2. APC 1393 GW contribution
Participants: Zvi Mizrahy, Pnina Vortman, IBM Haifa Orit Levin, RadVision Bill Strahm, Mark Baugher, Intel Nicole Gallant, Nortel David Walker and Annette Lanteigne, Mitel George Kajos, Irina Suconick, VideoServer
IETF Since the previous meeting on June 22, George sent mail to Bert Wijnen and Harald Avelstrad to request an IETF working group for H.Multimedia MIB. Bert responded and requested the following next steps: 1. write-up an agenda and a goal for the BOF at Chicago 2. find a proposed chair (could be George Kajos) 3. get the current work published as one or more I-Ds (internet-drafts) 4. setup a mailing list that people can use to dicsuss the drafts and issues related to the documents. During the meeting George stated that he would draft an agenda and volunteer to chair the BOF. The draft agenda will be sent to Bert Wijnen and Harald Avelstrad on Monday 7/13. Mark Baugher will check if Intel will can support the mail reflector. There is still too much work to do on the existing MIB proposals to comply with 3 before the BOF. The overall structure of the IETF submissions will have to be discussed at the BOF if one is approved.
A recurring issue, discussed at this and previous meetings, is whether the work is going to be performed at the IETF or the ITU-T. One aspect of the issue is how to serve two masters. The model discussed at both the ITU meetings and in the reply from Bert Wijnen is the details of content are clearly ITU charter work, while SNMP and MIB format and organization are inputs we expect to be likely from the IETF. The current ITU-T plan is to prepare the H.Multimedia MIB document for September determination. This will allow companies to start developing SNMP applications and get experience with the MIB proposals. The hope is that enough experience will be obtained to allow the work to reach decision in April 1999. This is analogous with the methodology used in the IETF where Internet-Drafts are expected to be implemented and reported on before going to RFC status.
Gateway - Zvi & Pnina Two weeks ago Zvi submitted the gateway proposal to the meeting group and George placed in the avc-site\Incoming directory on the Picturetel reflector as APC-1393.zip. In email with Mr. Okubo, he approved of using the incoming directory if he and Mr. Webber of Picturetel are kept informed. This is to prevent misuse of the Picturetel resources. Item 8 below is an updated list of what is on the Picturetel reflector and what has been updated since the Cannes meeting.
The IBM gateway proposal is a very comprehensive work. Its focus is significantly more than an H.323/H.320 gateway. For example, the connections can be point to point and multiway. In another example, among the defined connection types are PSTN and storage connections. This aspect of the proposal made George Kajos try to move discussion away from where to place the MIB in the OID tree and start the discussion by focussing what it is we are trying to define in this work. For example, what is the high level architecture of a gateway? The two models below were introduced as possible examples of what could be used as the basis for modeling a gateway.
Model 1: n by n Gateway
+-----+ H.323 ------+ +-----H.320 or PSTN +-----+
Model 2: generalized Gateway
Storage connections | +--+--+ H.323 ------+ +-----H.320 +--+--+ | PSTN
The first model depicts a point to point gateway with one side being H.323. This model forces one side to be H.323 and then it makes sense to place it under the H.323 OID hierarchy. The second model is a generalized gateway. It makes sense to be able to manage such a device but concern was expressed that this gateway night be beyond the charter of a H.323 version 2 device. Other opinions were that it doesn't make sense to design standard MIBs unless there is a specification to point to. Mark pointed out we need MIB specs that result in independent, interoperable implementations; for this reason we need to stick to the protocol specs as closely as possible. The H.246 document has a fairly complete definition of an H.323/H.320 gateway in Annex A and a start towards a H.323/PSTN gateway in Annex B. This is the model that George felt the group should start with or we could be beyond both our charter and our ability to agree on what the standard pieces are. More general gateway models are being discussed in conjunction with H.323 Version 3 and other IETF work. At the Cannes meeting it was agreed to try to focus on H.323 Version 2 and to meet market place for management of deployed H.323 devices. The proposal was for people to thing about these issues over the next week and we would discuss it in the next call.
Because this charter of the work issue came up, some expressed opinions we might want to start with something clearly defined like a H.323/H.320 and make it extensible as possible to other gateway types. Examples were that other network and terminal types such as H.310 and H.324 should be able to be plugged into the framework. This was generally agreed to.
Another primary architectural issue was that of multipoint and multiway connections. To illustrate the point, the following are some of the ways to create a multiway call. A multiway connection can be supported on some network fabrics such as Frame relay and ATM. They can be created at the network layer using IP multicast. They can be comprised of multiple circuits on a PBX, and they can be created by replicating at the server in MCUs. The focus of the discussion then was how many of these are actually to be modeled as a gateway call? Still the general tone of the discussion was that multiway calls should be able to be described in the MIB.
Other editorial comments were discussed. One was that common conventions should be used across all of the MIBs. This makes for consistency across all of the MIBs. Another was that Pnina mentioned that IBM was willing to put the Hardware and Storage sections in a private MIB.
Other Orit made the point that the work is going too fast. George disagreed and felt that it is imperative to get a document ready for determination in September in order that companies in the group would be able to start developing the MIBs in time for the April decision process. George made the point that all the proposals are on the Picturetel web site in a public place and that very little actual feedback had been received by any of the authors. It is a point that will be discussed off line with the SG16 Q12, 13, 14 rapporteurs.
Next Meeting: Wednesday, July 15, 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 4. Group - Admission and CallSignalling Tables index correlation. 5. VideoServer - Generate MC and MP proposal 6. George Kajos - Generate BOF request to O&M Area Director 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. 8. Contributors - place any updated the MIB proposals on the Picturetel reflector. The current status is: Avc-site\APC-1380 - overview Avc-site\APC-1387 - textual conventions Avc-site\APC-1388 - RTP Avc-site\APC-1389 - H.323 Call Signaling Avc-site\Incoming\RasV2Mib.rtf - H.323 RAS Avc-site\APC-1391 - H.323 Terminal Avc-site\APC-1392 - H.323 Gatekeeper Avc-site\Incoming\APC1393 - Gateway Avc-site\IAPC-1396 - H.320 Terminal Avc-site\IAPC-1397 - H.320 Call Signaling Avc-site\IAPC-1398 - H.221 Avc-site\IAPC-1399 - H.320 Multipoint Control Unit Avc-site\IAPC-1400 - H.245 The working proposal is that any updates be placed in incoming by identifying name and Mr. Okubo and the mailing list 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 11. Group - Determine whether they would be in attendance at Chicago IETF BOF