Jitsi Videobridge Negotiation

Question

I am trying to integrate Jitsi Videobridge into my existing WebRTC application.

Right now the signaling is relatively straightforward, offer, answer, candidates and we're off and running.

I cannot find straightforward documentation or tutorials on how to integrate a selective forwarding solution like Jitsi, everything seems to speak at too high of a level (see here).

Can someone explain to me concretely how the offer, answer, candidate model changes with Jitsi and how can I can even get a 1-to-1 call established through the video bridge?

My current analysis leads me to believe that I'm supposed to hit Jitsi over the REST interface to get candidates and swap out the host, srvflx, and relay candidates I generate currently with the ones from the video bridge. Is that even remotely on the right track?

Ataxia

What are you using for a signaling server? – Manaus

Custom application server we've written that's Java. We use ActiveMQ (JMS, NMS, Stomp, etc.) to communicate with C#, Objective-C, Java, and Javascript endpoints. Right now we do 1-1 WebRTC through these inputs doing the standard peer to peer signaling via ActiveMQ. – Ataxia

You're asking a good question. I want a description at the same level of abstraction, and independent of implementation language. I have not found it yet. – Fungus

What did you end up doing? – Koo

I never got a great answer to this question. Eventually we started shifting resources into the Janus Gateway which we've had some success with. – Ataxia

Hello There, did you find any documentation for above question? Can you please answer this question: #37550331 – Zurek

Jitsi doesn't seem to be designed for easy integration with third party applications. There are other SFUs that are, such as janus.conf.meetecho.com, github.com/pion/ion-sfu or galene.org (disclosure, I'm the main author of the latter). – Calton