A Multidimensional Trust Evaluation Mechanism for Improving Network Security in Fog Computing

Nodes in fog computing are easily captured since they are generally resource-constraint and usually work in open and unprotected environments. In this article, a multidimensional trust fusion preference (MTFP) strategy is proposed to detect the malicious node, which mines and utilizes multiple factors that affect trust evaluation. First, the service rating gaps between the self-evaluation and other evaluation in a historical sliding window are calculated and differentially used to alleviate the behavior sparse problem of direct trust evaluation. In addition, a multiperspective weight algorithm is designed to improve the rationality of recommendation trust. Moreover, a reward and punishment factor are designed and incorporated into global trust calculation to enhance the reliability of trust evaluation. Compared with the existing mechanism model method of trust evaluation, the proposed MTFP strategy uses more node information, which can improve the detection precision and recall of malicious nodes at the same time. Finally, the experimental results validate the effectiveness of the proposed trust evaluation method.