Analyzing peer-to-peer traffic's impact on large scale networks

Currently, peer-to-peer file sharing systems are playing a dominant role in the content distribution over Internet. Therefore understanding the impact of peer-to-peer traffic on large scale networks is significant and instrumental for the design of new systems. In this paper, we focus on Maze, one of most popular P2P systems on CERNET. We perform a systematic characterization of Maze's [1] traffic impact on CERNET. We investigate the traffic volume and bandwidth on different spatial levels aggregation. According to our log-based analysis, we claim that current P2P systems have much room to improve in reducing backbone network consumption. Locality-aware content delivering mechanism can reduce the traffic on backbone network effectively. Moreover, a system with less free-rider[3] will further reduce the traffic consumption. Thus the designers of P2P system should pay more attention on incentive mechanism to reduce free-rider.