Data Scheduling Based On Active Sharing In P2P Media Streaming System

Nowadays P2P media streaming system has become a hot topic in network application researching, and data scheduling as a key problem in P2P media streaming designing get much attention from both industrial and academic circles. Most of currently existing data scheduling strategies focused on requesting peers while ignoring the load states of serving peers, which lead to load unbalancing among peers in the P2P media streaming systems. Aiming at this problem this thesis proposes a data scheduling algorithm based on active load sharing, load of nodes are classified into three levels. And nodes which at low level actively brocast its load level to other nodes, those nodes which loads at high-level would transform some load to the low-level nodes by data request migration.Considering serving peers’load states, there are lots of selfish peers in the system and these peers would lead free-riding and tragedy of common become seriously in P2P media streaming systems. In order to solve this problem we present a P2P incentive mechanism based on game theory, in which peers’uploads are parts of its utility function. In this way peers are stimulated to upload actively. Finally we propose a collusion-resilience P2P media streaming trust mechanism against peers which want to collusion with others for illegal utility in the incentive mechanism. These two mechanisms are proved to improve the effectiveness of active sharing.Simulation results show that using data scheduling strategies based on active sharing can improve system performance with higher average quality of media playback, and also it can improve the efficiency of upload bandwidth utilization and reduce the server load stress. Furthermore, it can improve the overall resource sharing in the whole system, maximize the system utility, and improve the users’experience. As a result it has significant meaning for promoting the application of P2P media streaming systems using widely.