> I don’t understand what is difficult with WebRTC videos. It is difficult to scale a pure p2p model to multiple participants. Each participant would need to c

> I don’t understand what is difficult with WebRTC videos. It is difficult to scale a pure p2p model to multiple participants. Each participant would need to create multiple WebRTC streams, one for each participant. So if some call has 10 participants, each one of them would need to broadcast 9 streams and receive 9 streams with the assumption that each one of them have their microphone and cameras turned on. This would consume a lot of CPU and bandwidth. There are alternate models where you send your stream to a central server, and that server distributes the stream to other participants. This reduces the CPU and bandwidth required per participant, but requires a central server. This looks like the model google follows. So I don't think keeping such a service unauthenticated would be a good idea.