Research On Incentive Mechanism For Search Technology Of P2P Networks

As the rapid development of P2P network, various applications based on P2P network emerge. These applications have been applied widely to file sharing, distributed computing, the border service and so on. According to ipoque, a German research institution, P2P has completely dominated the Internet. About 50%~90% of all throughput is produced by P2P applications. As the scale of network grows, how to find nodes and locate resources quickly and effectively becomes the one of the key issues to solve.Gnutella, which is a typical P2P application, adopts Flooding which searches resources randomly. Nodes in the network don't consider other node's attribute, only duplicate the query message and forward to all their neighbors simply with great blindness. With the number of node increases, the redundancy information also grows substantially, which will increase the network load significantly, and even lead entire network to be useless.Up to now, the goal of intensive mechanism for P2P network is encouraging peers to share their resources, such as eMule and Maze. But this policy stands users'side, that is to say, nodes with higher score have higher priority to download resources. The efficiency of searching and performance of network don't improve virtually.Faced with inefficiency of searching resources in unstructured network, we propose an integration of search and incentive strategy (ISI), which not only exploits incentive in downloading but also in searching. The ISI includes two parts, one is neighbor selection strategy, and the other is score updating strategy. Neighbor selection strategy optimizes neighbor list and use new way to forward query message. Score updating strategy mainly encourage node to share resources. ISI can address the inefficiency of current searching algorithms, and can better locate resources and promote network sharing. Finally, to verify the effectiveness of proposed strategy, we deploy ISI to NS2 platform and realize simulation. Simulation results demonstrate that ISI not only improve search efficiency, but also can improve P2P system performance on download success rate, download time for P2P users, bandwidth utilization and throughput. Therefore, ISI can further improve search efficiency of unstructured P2P network, and promote resource sharing.