A Multi-objective Optimization Data Scheduling Algorithm for P2P Video Streaming

Abstract

In P2P video streaming, each peer requests its wanted streaming data from others and responses others’ requests by its data scheduling algorithm. Recent years, some data scheduling algorithms are proposed either to optimize the perceived video quality, or to optimize the network throughput. However, optimizing the perceived video quality may lead to low utilization of the senders’ upload capacity. On the other hand, optimizing the network throughput may lead to the degrading perceived quality, for some emergent data may not be transmitted in time. In this paper, to improve the two objectives simultaneously, we formulate the data scheduling problem as a multi-objective model. In the formulation, we not only consider the segment quality and emergency which affect the perceived video quality, but also consider the rarity of the segments, which influences the network throughput. Then, we propose a distributed data scheduling algorithm to solve the multi-objective problem in polynomial time. Through simulations, we show the proposed algorithm outperforms other conventional algorithms in perceived video quality and utilization of peers’ upload capacity.