Date Added: Jan 2010
The importance of distributed information sharing systems has been confirmed by the appearance and popularity of P2P applications such as Gnutella and Napster. One problem that the users face with such applications is that due to lack of a centralized server to keep track of what data are being stored in what peer, there is no clue about which peer contains what information. This makes it necessary for a peer to broadcast his query to all his peers. Likewise, his peers propagate the query to their peers and so on. Such a way of working faces two major problems. A lot of resource is wasted by peers during the handling irrelevant query as every query is broadcasted to every peer in the network. This also leads to increased traffic across the network. With an aim to offer some solution for these problems, this paper presents a new strategy for information retrieval over the Peer-to-Peer (P2P) network. The paper proposes a method to cluster all the peers that share similar properties together. This will help in organizing data inside the P2P network in a fashion similar to Yellow Pages. This will check query messages flooding and will also save resources used for handling irrelevant queries. Further, it also presents a new intelligent query routing strategy called the Firework Query Model to ensure efficient use of clustered P2P network.