[OpenSER-Users] RTP-Proxy

Jerome Martin jmartin at longphone.fr
Thu Jan 10 18:34:44 CET 2008


On Thu, 2008-01-10 at 18:06 +0100, Iñaki Baz Castillo wrote:
> CDR should always be managed by a system knowing the state of the RTP.

I don't agree.
This is a myth if you ask me. The only proper way to garantee that an
UAC is still alive is to make sure it is responsive to signaling. What
you care about in that context is signaling (BYE _is_ signaling).

RTP detection can help, but I consider this an ugly workaround when used
alone. What happens when someone hit the "secret" key on his phone with
blank compressions enabled and go get a coffee cup ? What happens for
SIP sessions without RTP ? Or one-way RTP only ?

Also, when your installations gets serious, you have MANY situations
where you want to optimize RTP path. Sometimes you want to avoid RTP
proxying completely.

To summarize, I'd say that if you rely on RTP detection for billing,
then you have the following limitations :
- unreliable problem detection
- stuck with RTP proxying for ALL calls
- problems with VAD
- problems in corner cases with re INVITES changing the RTP stream
extremities

What do you think ?


BTW, I work for a telco, so YMMV. In a setup with only one or two
million minutes per month, well I supose this might be a different
story.

Regards,
-- 
Jérôme Martin | LongPhone
Responsable Architecture Réseau
122, rue la Boetie | 75008 Paris
Tel :  +33 (0)1 56 26 28 44
Fax : +33 (0)1 56 26 28 45
Mail : jmartin at longphone.fr
Web : www.longphone.com <http://www.longphone.com>





More information about the sr-users mailing list