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(a)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(a)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
