Self-Organizable P2P Document Search Engine for Knowledge Management

Locating contents is an essential function, but it presents a very diÆcult and challenging problem for large-scale Peer-to-Peer (P2P) systems. Many P2P systems, protocols, architectures, and search strategies are proposed for this problem. In this paper, we focus on the selforganization of a community structure based on user preferences for P2P systems. We propose these methods to improve P2P search performance: 1) Extended Pong, 2) Pong Proxy, 3) QRP with Firework 4) Backward Learning, and 5) Community Self-Organization Algorithm. We evaluate the performance of the self-organized community network through simulations. These results show that the self-organized community network maintains a high query hit rate without over ow. As an example of implementation, we propose the Self-Organizable P2P Search Engine.