A Hybrid Approach for Clustering and Selecting of Cloud Services Based on User Preferences Evaluation

Abstract

With the increasing use of cloud computing, it is very important for the Cloud users to analyze and compare performance of the Cloud services. Since Cloud services selection problem contains several conflicting criteria, it is considered as a multi-criteria decision making (MCDM) problem. On another side, one of the most popular unsupervised data mining methods is Clustering which is used for grouping set of objects. The contribution of this paper is to propose an approach based on clustering, Pareto Optimal and MCDM methods. Our approach allows users to specify the quality requirements of the cloud services they want to use. It consists of three steps: in the first step, we use the clustering, more precisely the artificial neural network, to minimize the very large number of cloud services on the Net. In the second step, we apply Pareto Optimal algorithm to select non-dominated services. Finally, in the third step, we use the weights provided by the user to select the most appropriate cloud service for these requirements. To demonstrate the effectiveness of the proposed approach, a case study is presented.