Data Scheduling Strategy in P2P VoD System Based on Genetic Algorithm

In P2P (Peer-to-Peer) VoD (Video-on-Demand) System, data scheduling strategy is critical to make full use of node resource in P2P network, and it can help optimize user experience as well as system throughput. But for a normal peer how to efficiently schedule media data still remains a challenging task. The kind of problem is NP-hard problem due to the dynamic characteristics of P2P network. This paper presents a novel data scheduling strategy for resolving this problem, called GA scheduling, the major characteristic of which is to adopt genetic algorithm for optimization. Four steps (selection/ crossover/ mutation/ evaluation) are applied iteratively to optimize scheduling effect until stopping condition is met. Several experiments are designed at last on the P2P VoD system. The results demonstrate the effectiveness of the data scheduling strategy proposed in this paper.