Multipoint Cascading Issue!

Gaurav Makkar gmakkar at HSS.HNS.COM
Mon May 17 14:57:00 EDT 1999


     Please refer to H.323 recommendation, clause 8.4.5, "There shall only be
one master MC in a conference. A slave MC shall only be cascaded to a master MC.
Slave MCs shall not be cascaded to other slave MCs. This allows only dumb-bell
or star cascaded configurations.".

     My query is: why the above restriction? If one can have a multi-level tree
structure in a T.120 conference, why should the above restriction be applied to
an H.323 conference? Also, using the above, a conference (using a single MC) can
have only 256 participants (terminal label is 8-bits). Cascading would be
required to increase the size of the conference - again limited to 65536
participants (that ofcourse is large-enough for a planned conference, if not
adopt H.332) and that too in a "wierd" star configuration. Point is why have the
Master MCU and its network subnet be overloaded with information from all slave
MCs - why not permit a configuration in which an MC can act as a slave to one MC
and a master to another. That way a tree configuration for the conference could
be built.

     Am I missing something?? I would be greatful if someone could give some
direction on this account.

     Thanks!

Regards,

Gaurav Makkar
Senior Technical Leader,
Hughes Software Systems



More information about the sg16-avd mailing list