Data Scheduling Strategy in P2P VoD System Based on Adaptive Genetic Algorithm

Abstract

In Peer-to-Peer (P2P) Video-on-Demand (VoD) systems, data scheduling strategy is critical to make the most use of nodes in P2P network, and it also helps optimizing the time consuming of data scheduling and startup delay as well as the utilization of neighbors’ uplink bandwidth. But it is a challenging task to design an efficiently scheduling strategy. On one hand, the scheduling strategy should balance the need of segments between one requesting node and the whole system; On the other hand, the problem of scheduling segment transmission is a NP-hard problem due to the dynamic characteristics of P2P network. To resolve those problems we propose a scheduling strategy based on adaptive genetic algorithm (AGA). The scheduling strategy is based on segments’ emergency degree, supply and demand ratio and regional position to select the needed segments to schedule in a scheduling cycle. Then through the AGA to select the suitable supply nodes to transmit segments. Experiment results show that our proposed algorithm outperforms the RANDOM and Hill-climbing in terms of reducing the time consuming of each data scheduling, startup delay, and making the most use of the network uplink bandwidth.