On Monday 09 August 2004 17:32, Brian Rosen wrote: > The conferencing mechanisms being developed in SIPPING/XCOM > explicitly include end system mixing. There is a central > FOCUS, but not a central mixer. This means that one element > (which could be one of the endpoints) has dialogs with all > participants, and sends out the SDP with all the streams. > > For three way calls, a UA with an existing dialog could > REFER it's existing partner to the conference URI and INVITE > the 3rd party to the URI (assuming it doesn't need to > REFER itself). The dialogs would have either multiple > unicast, a single multicast, or multiple multicast > descriptions in it. > > Each UA would get two streams, either two streams in > one multicast, or two streams in unicast or multicast > and mix them. > > Brian
Hi Brian, Thank you for the information, can you please send me a pointer to a document describing such situation? I'm interested especially in packet traces like in draft-ietf-sipping-service-examples-07... I'm unable to figure out the required SDP data from the availavle RFC's and draft. When you say that : > Each UA would get two streams, either two streams in > one multicast, or two streams in unicast or multicast > and mix them. > You mean that the re-Invites sent by the focus to the endpoint will contains 2 'c' �lines? I don't understand how, for example, the conference focus is able to determine UDP port numbers other endpoints will use to communicate between them... Thanks Vadim _______________________________________________ Sip-implementors mailing list [EMAIL PROTECTED] http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
