I’m a bit curious about the Jitsi backend architecture. First, I understand that Jitsi deploys a few signalling servers to each region with each signalling server being connected to several SFUs in the same region, and also connected to all other SFUs outside the region. My question is, in any given call (like an international call), do all users still share the same signal server and the signal server just assigns SFUs that are in each user’s region or does every user connect to the signal server within their own region and the signal servers talk to each other to coordinate a call, and each assigns SFUs that are in it’s region?
It is one signaling server per conference and multiple jvbs, have you checked this https://jitsi.org/news/new-video-explains-geographical-cascading/ ?
Thanks! Just watched it. Really helpful. Appreciate it!