Re: [AVT] WGLC comments on draft-ietf-avt-ports-for-ucast-mcast-rtp-11: Part 2 Unicast RTP session Termination

[Magnus Westerlund <[e-mail]>]

Ali C. Begen (abegen) skrev 2011-01-19 16:22:
> OK, let's not describe a specific method, and the following text:
> 
> A specification using the Token-based authentication described in this document is REQUIRED to discuss the methods and rules for terminating/ending the unicast session.
> 
> I propose to add the text above to the 5th parag. of section 3 as follows:
> 
>    Upon successful validation and once the unicast session is
>    established, all the RTP and RTCP rules specified in [RFC3550] and
>    other relevant specifications also apply in this session until it is
>    terminated.  >>>
> 

Well based on this, I wouldn't come to the conclusion that one can stop
sending RTCP as server at all unless there has been some explicit
signalling for termination. this is Probably fine for the general
introduction. And it do leave the question of termination wide open.

So, I don't see other than that you have to discuss the termination in
the draft.

"The Unicast RTP session created using the port mapping method can be
explicitly terminated by procedures specified by the using application
or extension. Any unicast RTP session created using this method may be
terminated by a timeout procedure. If either client or Retransmission
Server has not seen any RTCP from any SSRC used by a peer for 5 regular
transmission intervals, the client or server SHOULD consider the RTP
session as terminated, and in that case not send any RTP or RTCP traffic
in the future. The client will have to establish a new unicast session
if required.

If no procedure for explicit termination exist, the client MAY stop
sending RTCP to the Retransmission Server to accomplish session
termination. The Server SHALL NOT stop sending RTCP until the unicast
RTP session is terminated. If Token Verification is signaled to be
allowed also on in the unicast RTP session, i.e. the RTCP messages sent
from c2 to P4, the client SHOULD prefer to terminate by sending RTCP BYE
on all SSRCs it has used in this unicast RTP session."

I think this works fine, and is way better than having no specification
at all.

Cheers

Magnus Westerlund

----------------------------------------------------------------------
Multimedia Technologies, Ericsson Research EAB/TVM
----------------------------------------------------------------------
Ericsson AB                | Phone  [phone]
Färögatan 6                | Mobile [phone]
SE-164 80 Stockholm, Sweden| mailto: [e-mail]
----------------------------------------------------------------------