Explaining BitTorrent traffic self-similarity

Peer-to-peer applications have become killer network applications. Understanding the nature of network traffic is critical in order to properly design and implement peer-to-peer network. Recently BitTorrent which is one of primary peer-to-peer applications has become one of most important information share tools on Internet. In this paper we examine the mechanisms that give rise to self-similar BitTorrent network traffic. We present an evidence for traffic self-similarity, and show that the self-similarity in such traffic can be explained based on the heavy-tailed distributions of BitTorrent transmission times and quiet times.