I'm new to Kubernetes and amazed by the potential, but I'm having quite a hard time to set my mind and think a solution to this:
I'm trying to set a WebRTC MCU (Multipoint Conferencing Unit), all peers connect to a single server, server combine videos/audios and send the output to peers. (Not discussing if I should use MCU or not, I just need to for reasons out of this question).
Of course this is heavy loads the CPU of the servers. So it's most likely to cause scaling.
The question is, I need to have all the peers (which can be all around the world) to be connected to the same pod because the video is processed in one place. How do I make sure of that? Is that even possible to achieve?
Right now I'm starting a common small server for each "room" and assigning that server address to it. Which works, but I'm wondering if I can do better.
