Tweaked optimization based quality aware VM selection method for effectual placement strategy

Abstract

Cloud computing has become a standard and promising distributed computing framework for the provision of on-demand computing resources and pay-per-use concepts. Operations of these computing resources result in maximum power consumption, enraged cost and high Co2 emission to the environment. The major difficulties faced when accessing cloud data center are SLA violations, increased time, less utilization of resources, high consumption of power and energy. Hence, considering these difficulties, a novel virtual machine (VM) selection approach is proposed to minimize the constraints while maintaining the SLA. First, based on the assumptions of VMs and physical machines (PMs), the overutilized hosts are detected using a static threshold approach, while underutilized hosts are identified based on the utilized resources. After load detection, the VMs that need to be migrated over other PMs are selected using the tweaked chimp optimization algorithm (TCOA). After selecting VMs without influencing the capacity of other VMs, the placement process is performed over other PMs using a power aware best fit decreasing approach. The proposed approach can greatly improve the QoS by selecting the optimal VMs that need to be migrated. Cloudsim is used as a simulation tool, and the results are compared with existing techniques in terms of migration time, energy consumption, SLA violation per host and so on to prove the superiority. The energy consumption of the proposed model is obtained to be 195.3 kWh, the overall SLA violation rate is attained to be 0.032%, and the migration time for 500 virtual machines is 8.72 s