Defending Peer To Peer Network System Against Pollution Using A Reputation Model

P2P file system is one of today’s Internet, one of the most important applications, such as KaZaa, and BitTorrent. Because no infrastructure, P2P systems can quickly create, share a highly efficient platform for distributing documents and files. P2P file-sharing networks the traditional C/S mode in the dissemination of the data have more advantages, including robustness, scalability, and data transmission. However, as P2P networks such as the openness and anonymity of the essential characteristics of the user when the lack of responsibility in the shared files, easy to attack by malicious nodes in the network to spread malicious files. In the now popular P2P network, you can see a lot of anonymous attacks of malicious nodes. For example, the virus dissemination of malicious nodes in the network, resulting in VBS. Gnutella worm epidemic. Recent research has discovered a new malicious behavior, called the contamination, malicious nodes in the network into a large number of false documents to try to prevent users from downloading files to a topic. In the Kazaa network, more than80%of the audio file is contaminated, which seriously reduced the availability of the system.Previous work has focused on two aspects. One is through the establishment of different flow models to analyze the spread of pollution in the file system; the other is to establish the credibility of the file system to curb the spread of contamination. However, these methods are rarely practical implementation, and evaluation from the performance point of view there are many limitations. Therefore, the purpose of this study was to design an effective way to identify the malicious nodes, stop the spread of polluted files malicious nodes to improve the availability of P2P file sharing system.In this paper, we propose a reputation system to help reduce pollution in the number of files, and to avoid proliferation in P2P systems. First, we calculate the credibility of the node, to help us get the correct index information. Secondly, we screened the database node, malicious node removed from the system. Finally, we calculate the credibility of the file object; it uses a new voting system, weighted evaluation value of nodes in the network, and the introduction of good incentives and system dynamics. This ensures an honest node to provide a credible evaluation of the file and correct. The system combines the reputation of the node and the credibility of the file object; you can prevent malicious nodes active, and reduce the spread of the file system contamination, while resistance to the attack. Simulation results show that the credibility of the system can effectively fight against pollution attacks, but Also has good robustness.