A new trust evaluation method based on reliability of customer feedback for cloud computing

Abstract

One of the most important factors for the adoption and success of cloud computing is trust. Adaptive trust evaluation is a fundamental component for secure computing in the cloud environment. Although several methods have been proposed recently for modeling and evaluating trust in distributed computing environments, there is no general accepted trust evaluation model for cloud computing. In this paper a new trust evaluation method has been proposed. In this scheme, each new customer of a cloud service can determine its trust level using the past experiences of previous customers of this cloud service. The proposed method introduces an adaptive method that helps distinguish between malicious and reliable customer feedbacks. The proposed scheme assigns a reliability weight to each customer feedback. Users who have shared any cloud service before and did rate it similarly are more likely to have the same opinion of a new service. Hence, their evaluation of a new service should be given a higher weight. In cases where no customer has shared a common service with us before, then existing customers who have had a feedback closer to the majority (in any service) may seem to be more reliable and hence their opinion is given a higher weight. The balance between the feedback of customers with shared service and those of majority consensus in the final decision is determined dynamically and based on the rate of malicious attacks. This way the method can respond better to changes in the rate and sources of malicious attacks. The proposed method has been compared with an existing trust evaluation method and proved its superiority in minimizing the effect of malicious feedbacks, and having a faster response time.