Gnutella (2001)

Studing unstructured peer-to-peer networks (for my undergraduate thesis project). Unstructured peer-to-peer networks such as Gnutella are attractive for certain applications because they require no precise control over network topology or data placement. However, the query algorithm used in Gnutella does not scale. Each query generates a large amount of traffic and large systems quickly become overwhelmed by the query-induced load. In this project we studied, through simulation, whether it is possible to achieve less traffic by keeping the classical routing algorithm of Gnutella and changing the network topology. Additionally, a portion of this project explored, through simulation, several alternatives to the query algorithm of Gnutella. We proposed a routing algorithm, which does not send messages to all the neighbors of a node but to some of them according to the number of hops that the messages have traveled.