EVMPCSA: Efficient VM Packing mechanism in cloud computing using Chaotic Social Spider Algorithm

Abstract

VM packing is one of prodigious challenge in cloud computing as it depends on the incoming workload onto virtual resources. Workload in the cloud is highly dynamic and it is difficult to predict the upcoming tasks and place them properly onto an appropriate VM is a challenging task. While scheduling workload onto VMs it is important to pack them in an appropriate physical machine because there is a chance of wastage of virtual resources and which leads to high energy consumption. Therefore, it is necessary to pack VMs into an appropriate Physical machines based on utilization of CPU. Many of authors proposed various consolidation techniques to assess parameters makespan, energy consumption, Througput but still there is a research gap and we can minimize energy consumption based on utilization of CPU in proposed approach i.e. EVMPCSA. Consolidation of VMs and chososing of VM for migration onto a Physical host is based on utilization of CPU. Chaotic Social Spider algorithm is used as a methodology for VM packing mechanism. EVMPCSA uses cpu utilization as constraint and used Chaotic Social Spider algorithm as methodology in this work to solve VM packing problem. It is simulated on Cloudsim and evaluated aganist existing algorithms named as PSO, CS and ACO. When it is compared with PSO, CS and ACO makespan is greatly minimizes by 30.25, 23.5% and 17.31% respectively and energy consumption is minimized for PSO, CS and ACO by 27.6%, 24.78% and 10.09% respectively.