GRAT: Group Reputation Aggregation Trust for Unstructured Peer-to-Peer Networks

Abstract

Peer-to-Peer (P2P) reputation aggregation methods are used to evaluate the trustworthiness of participating peers and to combat dishonest and malicious peer behaviors. The reputation aggregation method is to calculate the global reputation score from local score gained from each individual peer in P2P networks. On unstructured P2P networks, each individual peer exchanges own local score for other's local score and calculates the global reputation score. In this paper, we propose reputation aggregation method called GRAT (Group Reputation Aggregation Trust). The proposed method calculates global reputation scores by dividing entire peers into groups. Some peers create one group per a peer that is calculated global reputation score. Each peer exchanges local score among peers that belong to same group. Thus, even if the number of peers increases in the network, it takes shorter time to calculate global reputation score by using GRAT. Simulation results show that GRAT can efficiently exchange local score and accurately calculate global score in unstructured P2P networks.