In peer-to-peer (P2P) on-demand streaming applications, multimedia content is divided into segments and peers can seek any segments for viewing at anytime. Since different segments may be of different popularity, random segment caching would lead to segment popularity-supply mismatch, and hence uneven workload distribution among peers. Some popular segments may be far from peers, leading to inefficient search and streaming. In this thesis, we study optimal segment caching for P2P on-demand streaming....[ Read more ]

In peer-to-peer (P2P) on-demand streaming applications, multimedia content is divided into segments and peers can seek any segments for viewing at anytime. Since different segments may be of different popularity, random segment caching would lead to segment popularity-supply mismatch, and hence uneven workload distribution among peers. Some popular segments may be far from peers, leading to inefficient search and streaming. In this thesis, we study optimal segment caching for P2P on-demand streaming.

We first formulate the segment caching optimization (SCO) problem, and show that it is NP-hard. We then propose a centralized heuristic to solve it, which serves as a benchmark for other algorithms. We propose a distributed caching algorithm termed POPCA (POPularity-based Caching Algorithm), in which each peer adaptively and independently replaces segments to minimize the popularity-supply discrepancy and the segment distance from peers. POPCA also proactively advertises updated segment availability in a scalable manner to provide near-instant segment search. Through simulations and PlanetLab experiments, we show that POPCA achieves near-optimal performance, and lower peer workload and segment distance as compared other schemes. It achieves low overhead and search latency, and high hit rate of finding segments in the presence of peer churn. We also present an analytic model which closely matches with the simulation results. This model is simple, and helps us understanding the dependence of various system parameters and designing cache size to achieve a certain hit rate.