[Robustness] Minutes and telecon on May 9 11AM EST USA -- final r eminder
The next robustness teleconference will be held on Tuesday, May 9 at 11AM EST USA.
The phone number to call is:
(613)688-2795 Passcode: 2721# You must enter the # sign after the passcode.
Agenda:
1. We will discuss the tradeoffs of three different approaches to robustness (discussed in minutes to appear by Monday - 5/8) 2. Next steps for Japan meeting
Minutes of H.323 Robustness Teleconference - May 4, 2000 Prepared by Maureen Stillman, Nokia
Next Teleconference: May 9, 2000 11AM EST USA
Present on the call: Tim Chen, Mahesh Bhan & Archana Nehru - Trillium Tim Sipsey - Aspect Terry Anderson - Lucent Randy Stewart - Motorola Maureen Stillman - Nokia
Agenda: 1. Plans for Osaka 2. Continue discussion of technical options
Discussion:
There was a discussion of what messages can be lost at the H.323 (H.225/H.245) layer.
There are two ways to approach the problem of lost messages. In Q.931 all "application" messages are ACKed. In H.323 (call control and call setup) some end-to-end messages are ACKed and others are not. One way of solving this problem is to define end-to-end ACKs for all H.323 messages.
To make H.323 backwards compatible with earlier versions that would not support this function, we would introduce a flag that would indicate that this component supports end-to-end ACKs. Another flag could indicate that the component supported robustness mechanisms. In this manner, we could signal support for end-to-end ACKs but not other robustness mechanisms or support for both. The end-to-end ACKs would be mandatory for any component that supports robustness mechanisms.
Trillium agreed to write up a submission on end-to-end ACKs for the Osaka meeting.
Given this discussion we identified three different ways to proceed for the Osaka meeting:
1. Status Query mechanism without DDP
The status query mechanism is the robustness procedure submitted to the Geneva meeting. This mechanism would be enhanced by support for the end-to-end ACKs. The gatekeepers would be required to recover state from its upstream or downstream neighbors.
2. DDP with end-to-end ACK
See Osaka submission APC1772 "On H.323 Robustness Architecture Using DDP and SCTP". The alternate Gatekeeper mechanism would be replaced by DDP. DDP would serve two purposes, one to transparently select alternates and the second to use a bulletin board or other mechanism to keep state consistent. Bulletin board would not be the required mechanism as any other mechanism used to checkpoint would be sufficient.
3. end-to-end ACK without DDP
The end-to-end ACK mechanism would be supported. We would make use of the alternate Gatekeeper mechanism. Any other mechanism for pooling and checkpointing would be proprietary.
At the next teleconference, we will discuss these three alternatives.
-- maureen
Maureen Stillman Member of Scientific Staff Voice: (607)273-0724 x62 IP Voice Networks Fax: (607)275-3610 127 W. State Street Mobile: (607)227-2933 Ithaca, NY 14850 e-mail: maureen.stillman@nokia.com www.nokia.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ For help on this mail list, send "HELP ITU-SG16" in a message to listserv@mailbag.intel.com
participants (1)
-
Maureen Stillman