Re: [SR-Users] CallControl and MediaProxy

wrote:

Hello, I am not using mediaproxy at all (but nathelper/rtpproxy), neither the call control module, but making an option (module parameter or function parameter) for call control to bind to another module like media proxy, should not be big deal if it is all you are looking for -- I can look over it and send a patch if you are going to help testing it. I cannot do it these days, though, being out of the office. Cheers, Daniel On 2/23/12 8:59 PM, Reda Aouad wrote: First, I am posting about the wrong behavior of CallControl (or most probably Kamailio modules) which leaves no option. I should be the only one deciding about how to handle timeouts. If I decide to take some risk, no module should oblige me to do otherwise. Mediaproxy detects ONLY RTP timeouts from BOTH parties, because linux conntrack rules it uses are bi-directional. If a single party stops sending RTP for whatever reason (connection lost, codec with silence detection used, ....), mediaproxy doesn't care and doesn't act upon it. This is a feature, and a wanted one, to mainly support voice-detecting codecs. Think also about conferences for example, in which only a single person talks for a long time while others are silent and don't send RTP. Single-side RTP timeout because of a real problem (loosing network connection for example) should be handled with other methods, such as SIP session timers. MY POINT IS : I don't see it practical to handle RTP flows for EVERY call to handle the least probable scenario: an RTP timeout from both (or all) parties. If I understood well, mediaproxy updates the CDR when it detects an RTP timeout from both parties. CallControl can look in the CDR to debit the correct balance, instead of attaching itself to the dialog module to detect dialog termination. This is an extract from the call_control module : Even when mediaproxy is unable to end the dialog because it was not started with engage_media_proxy(), the callcantrol application is still able to detect calls that did timeout sending media, by looking in the radius accounting records for entries recorded by mediaproxy for calls that did timeout. These calls will also be ended gracefully by the callcontrol application itself. Unless there is something I miss.. I also opened a bug about the issue because call_control doesn't have the same behavior with OpenSips. It doesn't force mediaproxy. Reda On Thu, Feb 23, 2012 at 20:00, Jeff Brower <jbrower(a)signalogic.com>wrote;wrote: _______________________________________________ SIP Express Router (SER) and Kamailio (OpenSER) - sr-users mailing listsr-users@lists.sip-router.orghttp://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-users -- Daniel-Constantin Mierla -- http://www.asipto.comhttp://linkedin.com/in/miconda -- http://twitter.com/miconda