Trust Evaluation and Prioritization of Cloud Service Providers Using MADM Methods

Abstract

With the increasing market for cloud computing and the variety of Cloud Service Providers (CSPs) becoming available, making an appropriate choice of a CSP turns out to be challenging for the consumer. In order to cater to the user’s specific needs of quality, the services provided need to be user-centric. Trust plays an important role in the selection of cloud providers by users. This paper proposes CSP Trust evaluation, Prioritization and Recommendation System (CTPRS) for evaluating the trust of CSPs based on Quality of Service (QoS) parameters, and accordingly ranking and recommending them. MADM methods, namely ARAS, DEA, and GRA are used to evaluate trust by considering the user's preferences across multiple QoS parameters of CSPs. Comparative analysis of all three MADM methods is done using rank conformance, rank correlation, sensitivity, and computational complexity analysis. Experimental results demonstrate that GRA is superior to the other two methods. Thus, GRA is found to be the most suitable method and is therefore used for evaluating trust in CTPRS.