Re: [Ietf-behave] [AVT] RTCP SSRC collision and draft-ietf-behave-nat-multicast
Mark Baugher skrev:
Magnus,
On Oct 4, 2006, at 7:21 AM, Magnus Westerlund wrote:
Mark Baugher skrev:
Magnus,
I am unclear why we need to address this topic at all. I don't
understand why we need to support a napt for multicast. I'd like to
understand this basic requirement.
Mark
The task of the BEHAVE WG is to try to move to a world where NAPTs
behave in a more deterministic behavior. In that scenario multicast
can't be excluded, especially today when we start seeing more and more
deployment (at least on local scale) of multicast. For example to
support IPTV.
This is a chartered and milestoned work of the BEHAVE WG.
Maybe it's obvious, but there is no re-writing of the destination
address to a unicast or multicast administratively-scoped address. The
rewriting occurs only on the source address of an outgoing multicast
packet. This was not immediately obvious to me.
Exactly, the issue is that for a RTCP receiver outside of the NAT the
source address tuple of any RTCP packet received may change due to
bindings timing out between each transmission from the NATed sender.
Cheers
Magnus Westerlund
Multimedia Technologies, Ericsson Research EAB/TVA/A
----------------------------------------------------------------------
Ericsson AB | Phone +46 8 4048287
Torshamsgatan 23 | Fax +46 8 7575550
S-164 80 Stockholm, Sweden | mailto: magnus.westerlund@xxxxxxxxxxxx
