Discussion:
draft-ietf-sipping-nat-scenarios, TURN and RTCP
Frank Shearar
2010-06-18 08:24:25 UTC
Permalink
draft-ietf-sipping-nat-scenarios-13 section 5.2.2.2 TURN Solution shows
a user agent sending two Alloc Reqs to a TURN server for its two media
ports (RTP + RTCP). The TURN server returns (message (3)) STUN-PUB-2 for
the first Alloc Req and (message (7)) STUN-PUB-3 for the second Alloc Req.

Surely STUN-PUB-2 is not necessarily the same as STUN-PUB-3?

I would imagine that in most environments it would be, but if it's not a
MUST, what does one do in the SDP, given that "a=rtcp:$STUN-PUB-3.port"
defines the port that RTCP will use, but not the address?

frank
_______________________________________________
Sipping mailing list https://www.ietf.org/mailman/listinfo/sipping
This list is for NEW development of the application of SIP
Use sip-***@cs.columbia.edu for questions on current sip
Use ***@ietf.org for new developments of core SIP
Simon Perreault
2010-06-18 12:20:18 UTC
Permalink
Post by Frank Shearar
draft-ietf-sipping-nat-scenarios-13 section 5.2.2.2 TURN Solution shows
a user agent sending two Alloc Reqs to a TURN server for its two media
ports (RTP + RTCP). The TURN server returns (message (3)) STUN-PUB-2 for
the first Alloc Req and (message (7)) STUN-PUB-3 for the second Alloc Req.
Surely STUN-PUB-2 is not necessarily the same as STUN-PUB-3?
I would imagine that in most environments it would be, but if it's not a
MUST, what does one do in the SDP, given that "a=rtcp:$STUN-PUB-3.port"
defines the port that RTCP will use, but not the address?
Yes, this is important.

Assuming that this exchange implicitly uses the EVEN-PORT and
RESERVATION-TOKEN attributes, I think it should be made explicit in the
text.

i.e. Message (1) contains an EVEN-PORT attribute with the R flag set to
1. Message (3) contains a RESERVATION-TOKEN attribute, which is echoed
in message (5).

Simon
--
NAT64/DNS64 open-source --> http://ecdysis.viagenie.ca
STUN/TURN server --> http://numb.viagenie.ca
vCard 4.0 --> http://www.vcarddav.org
_______________________________________________
Sipping mailing list https://www.ietf.org/mailman/listinfo/sipping
This list is for NEW development of the application of SIP
Use sip-***@cs.columbia.edu for questions on current sip
Use ***@ietf.org for new developments of core SIP
Frank Shearar
2010-06-21 13:22:50 UTC
Permalink
Post by Frank Shearar
draft-ietf-sipping-nat-scenarios-13 section 5.2.2.2 TURN Solution shows
a user agent sending two Alloc Reqs to a TURN server for its two media
ports (RTP + RTCP). The TURN server returns (message (3)) STUN-PUB-2 for
the first Alloc Req and (message (7)) STUN-PUB-3 for the second Alloc Req.
Surely STUN-PUB-2 is not necessarily the same as STUN-PUB-3?
I would imagine that in most environments it would be, but if it's not a
MUST, what does one do in the SDP, given that "a=rtcp:$STUN-PUB-3.port"
defines the port that RTCP will use, but not the address?
It occurs to me that if $STUN-PUB-2.address != $STUN-PUB-3.address then
the user agent could
* repeatedly ask for TURN bindings until it gets a pair with the same
address or
* elect to multiplex RTP and RTCP over the same port, and abandon the
TURN binding for RTCP.
* ???

frank
_______________________________________________
Sipping mailing list https://www.ietf.org/mailman/listinfo/sipping
This list is for NEW development of the application of SIP
Use sip-***@cs.columbia.edu for questions on current sip
Use ***@ietf.org for new developments of core SIP

Loading...