P2P 環境中的群體語音通訊系統

Abstract

由於網路的普及以及頻寬的增加，越來越多人利用P2P的方式使用多媒體通訊，而語音通訊更是其中之一，但是在大多數系統以及使用者的使用習慣中，一對一的語音通訊仍然為主要的通訊方式；然而，隨著遊戲產業的發達，越來越多玩家在網路上彼此在遊戲中聊天，因此，我們希望能夠發展出一套讓多人通訊的系統，不會因為NAT的阻隔而無法使用網路語音通訊，並且發展出高容錯性的網路語音通訊系統，讓不論是語音品質或是使用的便利性都能更上一層樓，不只充分利用網際網路，也可以利用本系統拉近現代人的生活。

基於多人語音通訊系統的特性，由於節點進出頻率會較一對一語音通訊頻繁，因此在維持語音品質上的困難度更為提升，但是多人語音通訊系統也同樣具有多條傳輸路徑的性質，本論文利用這個性質結合HyperCube來達到語音容錯的目標，且讓在NAT後的使用者可以透過其他的節點加入多人群組並發送以及接收語音。最後將本系統和線上遊戲系統整合，讓正在玩遊戲的玩家可以利用本系統語音聊天。

Since Internet is growing, more and more people communicate with voice via P2P network. Since the game industry is growing, more and more online game player chat in the game, we hope to develop a group voice communication system without NAT restricted and with high fault tolerance mechanism. The better voice quality and convenience of group voice communication will make people closer.

Due to the nature of group voice communication system, its node transaction frequency is higher than those in one-to-one voice communication. Accordingly, the difficulty of assure its quality is greatly raised. Nevertheless, group voice communication system also has the characteristic of using multi-path transmission. This paper combines this characteristic with Hypercube to increase fault tolerance, and enable users behind NAT to communicate and form groups through different nodes. Finally, this system is integrated with online games in order to let players talk through this system.