Comparison of Data Replication Strategies for Peer-to-Peer Video Streaming

Abstract

Abstract-Current peer-to-peer (P2P) file-sharing systems are mostly optimized for file availability. This paper investigates P2P architecture for video streaming in general, and the performance impact of data redundancy schemes in particular. In particular, this work show that maximizing file availability is not the best strategy for video streaming as another constraint -peers' streaming bandwidth, comes into play. To address this limitation, a request-rate minimization policy is developed and evaluated using simulation. The resultant optimized replication strategy is then compared to data redundancy scheme based on erasure-correction coding. Simulation results show that with sufficient peer storage and a low erasure coding overhead, erasure-correction coding can achieve substantially better streaming performance than replication-based strategies.

Exploring the optimal replication strategy in P2P-VoD systems: Characterization and evaluation

“…Poon et al [12] reformulated the replication problem in terms of minimizing server's workload. Cheng et al [2] proposed and evaluated a heuristic algorithm of lazy replication.…”

Distributed Video Streaming over DHT P2P Overlays

“…The goals of a media streaming strategy include minimizing jitter and latency, while maximizing bandwidth utilization and visual quality. It has been shown that maximizing file availability through techniques such as replication is often not the best strategy when constraints such as bandwidth come into play [2]. Thus, popular methodologies for media streaming are usually achieved through direct server-client relationships, or else, by leverage the strengths of P2P transfers in the form of multicast trees.…”