Traffic Distribution and Self-Similarity of P2P Application

P2P file sharing application traffic has become a major component of Internet traffic, and to understand the characteristics and impact of P2P traffic has become an important issue. After analyzing the network traffic of Maze system, we have found that most part of traffic content are taken up by small part of flows in Maze application, but this extent is lower than the proportion of WWW (World Wide Web) applications and the ratio of the whole link. We have also confirmed the view that the rare part of hosts transfer most part of traffic, but the extent of which is lower than that studied before. Although the self-similarity for flow analysis shows that the arrival time and the arrival rate of flows all have significant self-similarity, self-similar factor has not reached a high level.