Cluster-based replication for P2P-based video-on-demand service
Chien-Peng Ho, Suh-Yin Lee, Jen-Yu Yu · 2010
This paper discusses how to apply a clustering technique for a video-on-demand (VoD) service to optimize the media-data replication strategies over unstructured peer-to-peer (P2P) overlay networks. In particular, the failure recovery of high popularity video can be accomplished with distributing high quality impact and popular replicas to the regions of low peers density or discontinuous area. The simulation results demonstrate that replicas of the popular and high video-distortion data are more significant and reflect the physical demands than those of uniformly distributed replication scheme for media-resource sharing P2P applications. In addition, the workload of VoD server can be reduced, and the averaged visual quality can be improved in the overall system.