A Measurement Study of a Peer-to-Peer Video-on-Demand System.

TLDR

P2P video‑on‑demand services attract significant interest, yet most research relies on simulation and neglects user experience, leaving the real‑world effectiveness of P2P VoD largely unknown. The study analyzes real‑world performance of a deployed P2P VoD system on CERNET 1 by examining two months of operational logs. We collected and analyzed these logs to evaluate system behavior, user experience, and resource requirements. The results indicate that channel popularity and a moderate number of concurrent users yield satisfactory experience, but adequate peer bandwidth and server provisioning are critical; a simple prefetching algorithm improves random seeks, and overall the system can deliver cost‑effective VoD with acceptable quality, though scaling trades off with user experience.

Abstract

Despite strong interest in P2P video-on-demand (VoD) services, existing studies are mostly based on simulation and focus on areas such as overlay topology. Little is known about the effectiveness of P2P in VoD systems and the end user experience. In this paper we present a comprehensive study of these issues using the two-month logs from a deployed experimental P2P VoD system over CERNET 1 . Our key findings are: (1) the key factor is the popularity of channels and a moderate number of concurrent users can derive satisfactory user experience. However, good network bandwidth at peers and adequate server provisioning are critical. (2) a simple prefetching algorithm can be effective to improve random seeks. Overall, we believe that it is feasible to provide a costeffective P2P VoD service with acceptable user experience, and there is a fundamental tradeoff between good experience and system scalability.