The Research Of P2P File Anti-Pollution Based On Hierarchical Reputation Model

Currently, P2P file sharing has been widely used on the Internet, however, it has producedmany negative effects. P2P file sharing system can provide users with convenient downloadservice, but the file sharing system can not guarantee the reliability of the downloadedresources. The P2P network itself has the characteristics of openness and anonymity, so anynodes can add into the P2P file sharing system. when malicious node joins the system, it willcause great harm to P2P file sharing system, and will cause users often download the pollutionfiles. Owing to losing the confidence in the system, the system will be paralyzed by losingusers. So it is necessary to take certain measures to combat file pollution in the P2P filesharing system. This thesis make a deep research on file pollution in P2P file sharing system,and design a reputation model based on hierarchical reputation model. Main work completedas follows:First of all, the thesis researches the several antipollution credit model proposed by thecurrent scholars, and analyses the characteristics of these types of credit models. Based oncomparing the advantages and disadvantage of several popular reputation model on fileantipollution, the paper proposes to establish a reputation model, which is more reasonableand more credible, called hierarchical reputation model. This reputation model is mainly usedin P2P file sharing system.Secondly, this paper proposed the hierarchical concepts and definitions, and thecalculation of the reputation of the nodes. The hierarchical reputation model is stratifiedaccording to the value of the node reputation, and is grouped by the contents of the nodesshare in each layer. When calculating and updating the credit value of the nodes, hierarchicalreputation model mainly consider the relationship between the reputation of the individualnodes, the reputation of groups, the nodes’ global reputation. In the hierarchical reputationmodel, mainly considers the relationship between resources and the nodes. By collecting theuser’s commentary after download and using the updated reputation, users can estimate thebehavioral characteristics of nodes, to guide download resources.Finally, The results of simulation show that the global reputation of hierarchicalreputation model can iterative convergence, and the reputation evaluation from the interactivenode have a high reliability. This model has a better effect on identification files and nodes，can guide users to choose the nodes to download resource which have high reputation, and hierarchical reputation model can recognize and prevent file pollution effectively in P2P filesharing systems.