Geographical Statistics and Characteristics of P2P Query Strings.
Adam Shaked Gish, Yuval Shavitt, Tomer Tankel · 2007
Abstract — P2P file-sharing applications are quickly being adopted by a wider and more mainstream audience. There is much to be learned from keyword searches users perform in order to retrieve content from these networks. This paper presents a large-scale measurement study of search terms in the modern Gnutella network. We developed a highly parallelized architecture capable of capturing an unprecedented amount of geographical-identified queries. We applied this architecture to generate daily logs of search queries. We collected over 25 such daily Gnutella logs, with more than 15 millions unique queries in each, over a three month period. We analyzed both static snapshots of the Gnutella networks, as well as the dynamics of the network over time. In particular, we look at the geographic location and the trends of searches to better understand the dynamics. I.