A Model to Seize the Instantaneous Performance of P2P Streaming Platforms: Simulative and Experimental Validation

In recent years P2P technology has gained great popularity not only in file sharing applications but also in the field of video distribution. This work proposes a simple model to assess the performance a P2P system designed for video streaming can achieve, measured in terms of average video delivery rate. The model allows to compute system efficiency in a simple but accurate way through the partition of peers in two distinct populations: bad peers, that are not collaborative, and good peers, that contribute to share their video contents with others. As a meaningful example, it is employed to investigate the behavior of a real P2P prototype subject to high peer dynamics: its effectiveness is proved via experimental results, and sheds a new light on the way a streaming service overlay can be centrally and timely monitored.