I have been interested in Jitsi Videobridge because it is one of the
main SFU actors on the scene of the WebRTC world. I tried the demo on
https://meet.jit.si/ and I wanted to know what are the advantages
provided by the VideoBridge instead of having a P2P call using WebRTC?
If all you are interested in is a simple one-to-one call, then none.
Jitsi Videobridge becomes useful when you want to do multi-party calls. Doing those without a central element implies that every endpoint needs to generate as many streams as there are participants.
This simply doesn't scale. You may get away with it for three ... maybe four participant calls but that's about it. Beyond that you will simply lack the network bandwidth and CPU to handle this.
With Jitsi Videobridge you only send your media once.
Also, when using SFUs like Jitsi Videobridge they can do smart things for you. They can drop the media from some people or simply parts of it and further optimise your bandwidth. Note that Jitsi Videobridge does not do these things just yet but we are working on them and they should be available within a month or two. It would then be easily possible to participate in 100-participant conference calls.
This can never work with full mesh networks.
Once you receive a packet, it's received. You can't unsend it.
Does this answer your question?
On 01.04.14, 05:33, Regnoult Francois wrote: