Comparative approach for VM Scheduling using Modified Particle Swarm Optimization and Genetic Algorithm in Cloud Computing

Abstract

The Users can access Cloud services anytime and from any location, depending on their needs. In a cloud platform, data of a vast amount is transferred from the user to the server and vice-versa. Whenever the VM Scheduling takes longer than expected, or the selected VM does not exist in the datacenter may utilize more Energy consumption and SLA (Service Level Agreement) violations with more VM Migrations. Because the VM is the primary element in the Cloud Environment, the VM’s assignment must be done correctly; resources must be utilized effectively, and no violations must occur with less VM Migrations. Two approaches are implemented for the comparison i.e., Modified Particle Swarm optimization (MPSO), and Genetic Algorithm (GA). The MPSO resulted better than GA by 6.0S%, LR-MMT by 32.2%, and GA at 27.81% compared to Local Regression-Minimum Migration Time (LR-MMT) in energy consumption. The MPSO resulted better than GA by 48.39%, LR-MMT by 91.6%, and GA by S3.73% compared to LR-MMT in VM migrations. The MPSO resulted better than GA by 5%, Local RegressionRandom Selection (LR-RS) by 71.21%, and GA resulted in 67.21% compared to Local Regression-Maximum Correlation (LR-MC) in SLA Violation. Therefore, the acquired results indicated that the suggested approach converges to optimal solutions with higher quality than existing algorithms compared to the QoS parameters.