Topology Analysis of Gnutella by Large Scale Mining

In this paper, we present a characterization of P2P network topologies, focusing on degree distribution and visualization of Gnutella network based on the data collected by our crawler. A new peer classifying method which classifies peers into asked and unasked groups is proposed in this paper. The distribution of asked peers´ degree and replied peers´ degree is identified. It shows that the distribution of asked degree is accurate, stable, well shaped, and the distribution has never been reported before. Understanding why Gnutella network has these features and figure out how to use these structures to modeling the Gnutella network is our future work.