An Adapted Multi-way Tree with Holder-chain for Large Scale P2P VoD System

Abstract

Abstr act. Peer-to-Peer Video-on-demand (P2P VoD) systems have attracted great interest in research community. Many hybrid methods have been proposed to provide high-quality video-on-demand services to support VCR-like operations, which can scale to large population of users. In this paper, we propose a hybrid overlay DONet-AMT, which combines DONet and an adapted multi-way tree. DONet is applied to distribute the video segments efficiently. The adapted multiway tree consists of two layers: a complete multi-way tree and holder-chain, which can support efficient VCR-like operations. Our scheme can ensure that every video segment has k (typically k≥4) replicas at least when the number of peers is large enough, and thus guarantees high data availability. Finally, the performance analysis shows that DONet-AMT can provide better VoD services than DONet-VoD. With the prevalence of broadband Internet access, VoD services over the Internet have become increasingly popular. The traditional client-server architecture can hardly scale to a large population of users due to the limited service capacity of the servers. Recent years, P2P networks have been shown to be a promising approach to provide high scalability VoD streaming service [1][2][3][4][5][6][7][8].In the P2P VoD systems, supporting VCR-like operations such as jump forward/backward is a great challenge. The VCR-like operations result in frequent seeking, and thus can impose great impacts on the viewing quality. We can define seeking latency as the interval between the time when a segment is requested and the time when the segment is ready for playback. To achieve a high viewing quality, the seeking latency must be as short as possible. To reduce the seeking latency,