Optimizing Resource Allocation in Cloud Computing

Abstract

Cloud computing is an essential technology, which provides the services such as newlineInfrastructure, Platform and Software. Providing these services to the multiple users of cloud by newlineefficiently allocating the available resources and balancing the load is challenging. In this thesis, newlinecloud computing resources like memory and processing are considered and the challenges with newlinescheduling these resources are reviewed. A strategy which models the resource allocation by newlineidentifying resource requirements, constraints, and their objectives, has been designed. newlineDeveloping a dynamic strategy is a major challenge for the cloud service providers. This strategy newlinemust ensure the maximum resource utilization of cloud resources and improvement in the system newlineperformance in case of changing user demands also. The proposed strategy integrates concepts newlineand principles of resource allocation mechanism and power awareness. Each entity has specific newlinerequirements in terms of scheduling and performance preferences. The performance preferences newlinein the cloud computing environment are analyzed and a model architecture that provides a newlinescheduling solution and performance concerns in the cloud is proposed. Defragmentation newlinetechnique constitutes as one of the methods for load balancing thus promoting the better newlineutilization of the resources and no need of extra resources for handling the requests. In data newlinecenter, virtual machine allocates the resources to user based on their need. In order to save newlineenergy and cost while the operation is executed and meeting demands, an effective scheduling newlinetechnique better than already surveyed strategies has been proposed. The scheduled load newlinebalancing among the available virtual machines results in saving cost and energy evidently. newlineFinally, as a demonstration of the applicability of the approach, an experimental study is newlineimplemented using CloudSim tool and proposed mechanisms like defragmentation, genetic newlineapproach and task consolidation are providing better results than existing strategies. newline