A Trust and Reputation Model Considering Overall Peer Consulting Distribution

In new widespread peer-to-peer (P2P) systems, peers are exposed to great risk due to frequent trading with unfamiliar peers. Therefore, trust and reputation mechanisms become important issues. For computational efficiency, this paper focuses on localized information trust/reputation mechanisms. Previous studies have not paid much attention to the overall distribution of peer interactions. Based on the scale-free feature of real-world networks, we introduce a power-law distribution of the number of neighbors in P2P trust and reputation systems. To rigorously distinguish the effects of the overall considerations introduced herein, we compare the model proposed in this paper with models in previous studies under the same set of parameters. Simulation results show that the proposed model can discern a small difference between real quality of service (QoS) and other peers' feedback while distinguishing the malicious peers, even when the exaggeration coefficient is high. When one or a group of peers change their QoS, the model exhibits a quick reaction to this change. This response is demonstrated by a rapid decrease in reliability when the QoS change is downward and a slow increase when the change is upward. A slow reaction to the upward QoS change may exclude those peers who frequently change their QoS and encourage consistent reliable service providers.