Hi! I would like to develop public audio conference (some sort of personal radio). Now I am looking for the best solution to do this. I think, that there will be one, two, three, maybe 5 people who has right to talk. Others can only hear them. And, the leading man can give this possibility to talk or deny it. I see some ways, all of them use media servers:
- One stream for every talking man: everybody who has right to speak in this radio, publishes his NetStream. Other people play that stream.
- Everybody who has right to speak publishes his stream and all these published streams are combined on media server. So, everybody plays that combined stream and listens to all of them.
What way is better (from the view of scaling)? Maybe, you suggest other ways?
Also (if you have experience), what is maximum number of connections in such applications? Is it poosible to make a cluster of these servers? And which servers' performarmance is better?