Chasing: An Efficient Streaming Mechanism for Scalable and Resilient Video-on-Demand Service over Peer-to-Peer Networks

Provisioning scalable and resilient Video-on-Demand (VoD) service is both challenging and interesting. Recently, peer-to-peer (P2P) networks are introduced to address the scalability of VoD service over Internet. Most of existing work follows the line of cache-and-relay (CR) scheme to accommodate the asynchronous characteristic of requests from a community of end users. Aiming to take full advantages of bandwidth capacities at each node and pre-recorded feature of requested video files at streaming server, we improve traditional CR approach by efficiently exploiting surplus bandwidth and proactively prefetching media contents from either the server or other peers. Our proposed basic chasing and advanced chasing mechanism not only achieve significant reduction of workload on streaming server, which could translate into better scalability, but also help the streaming session to adapt to volatile network fluctuation. Our extensive experiments have demonstrated encouraging results with respect to increased system performance.