A proposal of a new search method with the dynamic group management function on peer-to-peer network

摘要

In Peer-to-Peer (P2P) network like Gnutella, each node relays the received messages to all the connected nodes. The number of relayed messages it proportional to the number of the nodes on the network. The increasing messages cause network overloading. Jxta attempts to reduce the network loading by forming a group of nodes with similar contents and relaying messages only in that group. However when a node fails to discover the group with the target information, the node will participate in another group with unrelated information, or that the node forms another group itself. These nodes give rise to groups with unrelated information or two or more similar groups within the network. Consequently, the number of search messages that were relayed increases. In this paper, we propose a method that realizes the information sharing between related groups, and reconstructs the optimal group based on the contribution rate of each node in the group. This method reduces the number of relayed messages, anti increases the probability of acquiring the target information. This paper proved the efficiency of this proposed method through experimental simulations.