Game theoretical bandwidth request allocation strategy in P2P streaming systems

Due to the merits of lower bandwidth consumption at streaming server and higher scalability, P2P streaming systems have been widely developed and deployed. However, the heterogeneity of bandwidth resource and playback position at peers may easily lead to load unbalancing problem, especially in the era of emerging booming if mobile Internet applications. This may severely deteriorate video playback quality at peers. In this paper we study bandwidth request allocation strategy, aiming at balancing the traffic load at peers and thus improving peers' playback quality in P2P streaming networks. We formulate a non-cooperative game model to analysis the bandwidth resource competition between multiple requesting peers and service peers, through searching the Nash Equilibrium of this game, the optimal bandwidth requesting strategy can be obtained. Then an distributed algorithm is proposed, called Game based Bandwidth Request Allocation strategy (GBRA). We conduct simulation experiments to validate the effectiveness of GBRA, and numerical results show that the proposed strategy can significantly improve the load unbalancing problem in P2P streaming system and decrease the latency of streaming data retrieval at peers in P2P streaming networks, compared with the classical bandwidth request allocation strategies: proportional strategy and greedy strategy.