An effective scheme for performance improvement of P2P live streaming systems

To solve the problems of long start-up latency, low playback continuity and data distribution rate in P2P live streaming system, this paper proposes two optimization strategies. One is an adaptive peer selection algorithm based on peers' real-time service ability. In the algorithm, a simple but effective method is proposed to calculate the peers' ability. And source peer adjusts the amount of requested data and the priority of destination peers according to the destination peers' abilities and the limit of out-degree. Thus a load balance and efficiently system can be constructed. The other one is a push-pull combination data distribution mechanism. This mechanism adaptively chooses push or pull model to distribute data fast and efficiently. The results show that the strategies have a significant effect on reducing start-up latency and control message overhead and improving the playback continuity and data distribution rate. © 2014 ACADEMY PUBLISHER.