Cloud Service Provider Selection using Non-Dominated Sorting Genetic Algorithm

Abstract

Cloud Computing is a specialized form of computing paradigm which maintains the underlying resources like processors, storage, and memory, completely abstracted from the consumer. It includes access to infrastructure/application services on demand and as well on a subscription basis. The drastic increase in the number of Cloud Service Providers (CSPs) offering these services provides more options for the customers to choose from. The evaluation and selection of cloud service is a multi-criteria problem, which includes many intangible factors that are difficult to measure. The cloud user may have to make the right selection so that they do not incur undue losses during the tenure in which they are locked. Trust-based decision-making models ease this selection process and help to carry out transactions in such open and dynamic environments. The proposed work have developed a trust estimation framework for infrastructure-based cloud service selection using Fuzzy Multi-Objective Optimization methods. The developed trust estimation framework uses Service Measurement Index parameters which are obtained from the estimates of their defining attributes. This trust estimation framework provides a quantitative estimate of trust for various service providers and helps in rating them based on their infrastructure.