A Scalable Method for Taking Detailed and Accurate Geo* Snapshots of Large P2P Networks

Peer-to-peer file-sharing systems offer high transfer speed at remarkably low data ownership cost. In many cases, the transfer speed can be increased if the users possess timely and accurate information over the location of other users. Furthermore, robustness and fault tolerance, issues of intense research investigation, are inextricably linked with the location and activity of the network users. However, the sheer volume of available information makes it hardly useful in the absence of a fast and accurate processing method. We present a new method for taking geographical and geopolitical snapshots of large Peer-to-Peer networks. We show that this method has a mapping performance an order of magnitude faster than that required for taking snapshots each minute and that the network snapshots it produces are accurate. We validate the method by processing a large data set for the currently most used Peer-toPeer file-sharing system, BitTorrent. Finally, our results give evidence that BitTorrent content presents non-trivial locality features and show that exploiting these features could greatly benefit the world of P2P systems, and in particular BitTorrent.