A measurement study of neighbor selection mechanism in a P2P live streaming system

Live streaming video has become one of the most popular P2P applications with its high-quality performance and playing effect. Since the real-time requirement, live streaming systems have to adopt an appropriate neighbor selection mechanism to realize effective content distribution. This paper takes PPLive which is a classic P2P IPTV system as the object to measure and analysis. We pay our attention to the quantitative variety, the ISP and autonomous system locality characteristic and the response time distribution of the neighbors. According to the research result, we find PPLive employs a simple neighbor selection mechanism we call it “the fast response, the best”. In such selection policy, peer prefers to select the peers with low response time as neighbors, and most of the data is downloaded from the lower-response neighbors.