Attaining reliability and energy efficiency in cloud data centers through workload profiling and SLA-Aware VM assignment

Abstract

To attain the desired reliability and energy efficiency of a cloud data center, trade-offs should be carried out between system performance and power consumption. Although the deployment of energy-efficient hardware is a crucial step, getting rid of underutilized servers is a far more effective approach. Achieving this in a cloud data center would require monitoring resource utilization as well as managing the quality of service and the profitability of a cloud infrastructure. In this study, we propose a resource provisioning technique which involves clustering of similar VMs in a cloud datacenter based on their workload profile. We then combine it with an SLA-aware approach for VM assignment in order to improve the efficiency of virtual resource provisioning and increase the quality of service. The proposed approach was compared with other techniques and as shown by the results, performance and energy efficiency is improved.