Effect of VM Selection Heuristics on Energy Consumption and SLAs During VM Migrations in Cloud Data Centers

Abstract

The virtual machine (VM) provisioning in cloud computing offers a good possibility for energy and cost saving, given the dynamic nature of the cloud environment. However, the commitment of giving the best possible quality of service to end users often leads to the requirement in dealing with the energy and performance tradeoff. In this work, we have proposed and evaluated three different virtual machine selection policies (MedMT, MaxUT and HP) to achieve a better performance as compared with the existing state of art algorithms. The proposed policies are evaluated through simulation on large-scale workload data conducted over a period of 7 days in a series of experiments. The results clearly indicate how the virtual machine selection algorithms can improve upon the energy consumption by data centers as well as the overall reduction in service level agreements (SLAs), thus reducing the cost significantly.