P2P VoD Content Distribution Strategy Based on Gradient Model

Abstract

The characteristics of P2P VoD (Video on Demand) system, such as low cost, great scalability and high quality of service, make it become the research hotspots and development trends of current network applications, while content distribution policy is one of its key technologies. To solve the problems of data scheduling strategy priority and node performance discrepancy in content distribution, the influences on scheduling policies of network performance information, the degree of content scarcity and playback emergency were studied. Based on the investigation, the gradient model based content distribution strategy is proposed. Distributed gradient model based strategy, sender startup strategy and receiver startup strategy were introduced to make each peer merely interact with parts of peers to reduce the collisions caused by the load balance. The experimental results show that gradient model based content distribution strategy improves the delivery speed and reduces the network load.