BGET: A Blockchain-Based Grouping-EigenTrust Reputation Management Approach for P2P Networks

Trust, as an effective way to reduce complexity and risks of systems, is experiencing various challenges in distinguishing reliable partners of distribution environment like Peer to Peer (P2P) networks. As one of the most known and successful reputation management systems, the EigenTrust reputation management system has been widely used. However, this kind of system uses centralized reputation calculation strategy, which relies heavily on the mechanisms of global ranking and pre-trusted peers. It causes high-reputation peers to center around pre-trusted peers and uncontrolled spread of inauthentic downloads, as a consequence, other low-reputation peers will be marginalized despite potentially they could be honest. To deal with these problems, we put forward a blockchain-based grouping-EigenTrust (BGET) reputation management approach. BGET uses grouping-ET algorithm to manage reputations of different peers, which utilizes uniform grouping strategy and intragroup random walk strategy to divide peers into different groups to guarantee the uniform distribution of high reputation peers and limit the spread of inauthentic downloads. Moreover, BGET provides reliable verification services based on blockchain, which can improve the credibility and quality of transactions. Through simulations, we proved that BGET has good extensibility and robustness. BGET can effectively maintain higher success rate of tasks even there are many malicious peers in the network.