Faculty of Computer Engineering, K. N. Toosi University of Technology, Tehran, Iran
Abstract
Distributed Hash Tables (DHTs) are used in the peer-to-peer networks for scalable and distributed localization of resources. However, malicious peers may affect the lookups by manipulating the routing information. Without any central monitoring nodes, pollution will be a real concern in these networks. In this paper, using the peer feedbacks about each other, a reputation system is formed. The proposed consensus algorithm assigns reputation scores to the peers of the system. By using consensus algorithm, the peers can aggregate on the reputation score of the desired peer and make a trust decision about it. The success rate of searches in the proposed consensus algorithm is 15% higher than the ReDS algorithm which is one of the best reputation systems based on DHTs that is designed against pollution in the peer to peer systems. In the following, the weighted consensus algorithm has been proposed. In the weighted consensus algorithm, the reputation scores suggested by other peers are weighted by their own reputation scores. In the weighted consensus algorithm, the success rate of the searches is approximately 40% increased while the memory usage is increased about 14% compared to the consensus algorithm.
)