Research on the implicit feedback mechanism in peer-to-peer trust models

Abstract

Most existing feedback or recommendation-based peer-to-peer trust models are implemented based on credible and sufficient feedbacks. However, there exist too many malicious nodes which provide false feedbacks, and what is more they do not even provide feedbacks. Also, since indirect trust is usually used to calculate a node's trust in most feedbackbased trust models, the high communication complexity is required for getting feedbacks in the network. Aiming at solving these problems, this paper presents an implicit feedback mechanism in peer-to-peer trust models. Any node owns its proxy nodes, and the node's trust and feedbacks are both calculated by its chief proxy node implicitly and automatically without human intervention, thus reducing the impact of false or insufficient feedbacks on the trust calculations. Meanwhile, all the feedbacks and evaluations on a node are managed by the node's proxy nodes implicitly, rather by itself, reducing both the possibility that they are falsified by the malicious nodes and the communication complexity. The simulation results show that our implicit feedback mechanism could both improve the performance of peer-to-peer trust models and reduce the communication complexity.