An Intelligent Resource Selection Framework for Cloud Computing Environments

Abstract

Cloud computing provides computing resources as utility services. It offers clients with several usage scenarios. These scenarios fulfill the services requested by the clients with a lower cost as compared with deploying them locally. This paper introduces an intelligent resource selection framework that provides the client with the best set of resources in terms of computing cost and the fulfillment of client's requests. It is assumed that the requested resources can be provided to the client from multiple data centers. Additionally, cloud resources are allocated in units determined according to client's request. The client submits his resource requirements in terms of units of computing power, storage, networking, software ... etc. Then, the proposed framework finds the best set of resources that satisfies the client's request. Moreover, it takes into consideration the inter-dependencies among resources that lead to mutually allocating certain resources together. Additionally, it considers the overhead incurred in acquiring and using the requested resources. This problem is formulated as a constrained optimization problem and has been tackled in this paper using genetic algorithms. Simulation studies show that the proposed framework provides a near-optimal solution for the studied cases.