A Credible and Lightweight Multidimensional Trust Evaluation Mechanism for Service-Oriented IoT Edge Computing Environment

Abstract

To meet the increasing need of a reliable and service-oriented IoT edge environment, a credible and lightweight trust evaluation mechanism is expected urgently. Due to the task openness, device heterogeneity and resource dynamics of the IoT edge computing, the lack of trust is inherent, which hides the public from excellent computing services. Therefore, in this paper, a service-driven collaboration mechanism based on multidimensional trust evaluation of IoT devices is proposed. Due to the design of the multidimensional trust evaluation mechanism, the service's quality affected by low-capability devices can be improved and the performance against malicious attacks could be well improved as well, which means this trust evaluation scheme is robust. With the double-filtering design based on K-means clustering algorithm in this mechanism, the feedback with low similarity of current task context and the feedback from malicious devices could be filtered effectively, which makes the trust evaluation mechanism more credible. In addition, owing to the application of low-overhead algorithms, our evaluation mechanism is lightweight, which could have an outstanding efficiency in large-scale IoT edge computing. According to the experiment results under the simulation environment, the credibility and robustness of the IoT edge environment could be obviously improved with the application of our trust evaluation mechanism.