Topology Re-formation Algorithms for Ubiquitous P2P Networks Based on Response Statistics

Recently, peer-to-peer (P2P) networks are becoming popular as communication forms for ubiquitous environments. Flooding-based P2P networks such as Gnutella have a problem that they generate a huge amount of network traffic. Moreover, their recall ratio is generally low because their topologies are constructed regardless of preference of each peer. In this paper, we propose a topology re-formation algorithm for flooding-based P2P networks. In the proposed algorithm, each peer observes responses that passed through itself and re-forms the topology so that peers with a content matching the query can be located near the query issuer. By simulation experiments, we show that peers with similar preferences get collected in the P2P network, and the recall ratio improves greatly