Virtual Machine Selection and Migration: Challenges and Future Directions

Abstract

Over the past ten years, cloud computing has significantly altered many aspects of human life by providing access to hardware and software resources through the internet. Businesses or individuals can use resources without setting up and maintaining their IT infrastructure. A data center is the core computation unit of any cloud environment, and it consists of hardware-oriented machines termed physical machines (PM) and software-oriented machines that are termed virtual machines (VM). The fundamental method of generating different resources from the given physical infrastructure is virtualization. The use of virtual machines is expanding as people use more smart gadgets that are highly computational and require virtual machines to run efficiently. An enormous amount of energy is required to run a cloud’s services when they are established on a large scale. Resource usage and energy management must be carefully managed in a cloud environment to accomplish virtualization. To do this, it should be necessary to manage the workload by dividing it equally among the physical machines. But due to rapid development and growing user requests, the workload cannot be equally divided between the physical machines. There is a need to apply a virtual machine selection process, which will identify under-utilized and over-utilized PMs based on resource utilization. To minimize energy consumption, there is a need to migrate VMs from over-utilized PMs to other PMs and bring all PMs to a neutral state without compromising the quality of service. This paper presents the challenges and future directions in the VM selection and migration process of cloud computing.