I am investigating the use of the Jitsi Meet (JVB, Jicofo, Jigasi) as a way to conference SIP endpoints.
To test the feasibility of this solution I have made a simple POC where a simple XMPP client creates the Jitsi conference and dials out via Jigasi to the conferees. The SIP endpoints receive the calls and can answer, after which they stream audio RTP.
One of the key issues with our SIP infrastructure is that we use an SBC Relay for guarantied NAT traversal (no ICE enabled) and our SBC will not forward RTP between clients until it has seen RTP from both legs of the call.
So in my simple test the clients are not receiving any RTP from the bridge because our SBC will not forward their RTP until it sees RTP from JVB.
My suspicion is that JVB won’t send any conferenced RTP until it has received RTP from at least one of the conferees.
So, assuming this is the case, is it possible for the JVB to send RTP (say comfort noise in place of the missing mixed audio) unit it receives RTP from the endpoints?
Is there are config setting that can enable this behaviour?
Thanks in advance for any help with this.