A Novel Data Replication and Placement Scheme for Time-Shifted TV Cluster

As a new multimedia streaming service, time-shifted TV introduces new challenges in content distribution. A number of replication algorithms were proposed to alleviate the bandwidth bottleneck, but these algorithms are not suitable for time-shifted TV due to not understanding television user access behavior. In this article, we presented a new popularity model based on analyzing of television access pattern. By using the mode, we proposed a classification-based replication algorithm and a novel replicas placement algorithm, named as least load proportion first placement (LLPFP). The algorithms focus on the time-varying nature of time-shifted TV, and use previous stored information to reduce the cost of redeploying. Simulation reveals that the proposed algorithms are cost-effective on improving the performance of time-shift TV servers in terms of total outgoing bandwidth requirement.