Performance Comparison of VM Allocation and Selection Policies in an Integrated Fog-Cloud Environment

Abstract

Virtualization generates the virtual machine (VM) instances of software and hardware built on a single physical machine and helps to utilize the resources and infrastructure completely. VMs must be migrated to avoid over utilization and under utilization of hosts which can be done effectively by VM allocation and selection policies in resource management. Such requirements can be completely satisfied when using the fog enabled cloud systems. This paper proposes a combination of VM selection and allocation algorithms for integrated fog cloud system to improve the performance parameters like energy consumption, network use and cost. The results of the simulation show that the proposed strategies outperform the default VM selection and allocation algorithms in fog - cloud environment. The paper also presents the comparison of the above algorithms on the real-time data sets obtained from Google trace and PlanetLab.