Quote:Original post by Ma-we-TeSkype is generally one-to-one. Okay imagine you have 40 people in a voice chat and don't say it doesn't happen since I've done it while playing Planetside :P. Okay so lets say Everyone is talking and such randomly and sometimes they're talking at the same time. In P2P you'd be getting multiple streams of sound and that can be a little intensive for some people's bandwidth. As you might know you can put the sounds together as layers since that's what happens when they come out of your speakers. Now imagine 40 people start talking into their mics or maybe they keep their mic open 24/7 and you can hear them all breathing :\. Now in a C/S with a dedicated server you'd be getting all of these streams and simply merging them together and sending out one stream to all players. This scales perfectly for clients. No matter how many users there are talking and such the clients are getting one stream of sound with all of the voices so the bandwidth is constant. (If no one is talking then the download cost for a client is none. If everyone is talking it's the same bandwidth as if one person was talking).
Telephone exchange is to Client-Server what Walkie-talkie is to Peer-to-Peer, so applications, especially games, should be doing at least voice chat via P2P, they don't? -- Signaling, as in neural networks, kind of like binary search tree for pattern recognition. File sharing can also be one-to-many, many-to-one, and many-to-many. Wikipedia article says Instant Messaging and Online Chat apps, as well as Skype and VoIP all use P2P. Conference type of calls, or similar large-number interaction, like voice chat in multiplayer games, therefore seem ideal candidate for this architecture.
Instead of skype look at TeamSpeak or Ventrillo which are conferencing applications
