A Metaheuristic Approach on Virtual Machine Allocation Using Backtracking and Simulated Annealing in Cloud Computing

Abstract

Cloud computing is the foundation of all Internet-based applications and services rendered throughout the globe. Cloud serves as a driving force behind all computer infrastructure managed worldwide. Infrastructure as a service is one of the 3 basic services offered in the cloud. Proper management of infrastructure is a must, and hence, more algorithms and approaches borrowed from nature, mathematics, genetics, and others are experimented with to effectively use the cloud resources, such as Physical Machines (PM) and Virtual Machines (VM) inside them. In cloud business, the jobs are ultimately submitted to VMs, and it is called VM Allocation, which is composed of two main subtasks: VM Placement for jobs and VM selection for migration if needed for load balancing. In the research work, two metaheuristic approaches, namely, Backtracking for VM Placement and Simulated Annealing for VM Selection, are proposed. The experimental studies show there is a commendable improvement in the parameters of consideration Time and Energy consumed by the algorithms, such as a 10.43% decrease in the number of hosts shutdown, an 8.44% decrease in energy utilization, 23.42% and 12.65%, respectively, decrease in mean and standard deviation of execution time. The approach is termed as metaheuristic as it employs higher level problem-solving techniques.