An Adaptive Threshold-Based Modified Artificial Bee Colony Optimization Technique for Virtual Machine Placement in Cloud Datacenters

The usage of cloud computing service platforms are exponentially growing to provide on-demand services for end-users for using advanced technologies. These platform services are achieved through resource virtualization to maximize the resource usage and minimize energy requirements. Energy consumption is a key factor for designing efficient and manageable cloud data centers. Optimal techniques are used for placing virtual machines in physical machines to reduce the energy consumption ratio of physical hosts. This paper proposes a novel efficient virtual machines placement algorithm for a cloud computing environment. This method exploits a modified artificial bee colony optimization algorithm for identifying under-utilized physical machines based on energy consumption and resource allocation charts. An adaptive threshold method is then proposed to select suitable threshold levels for energy consumption to identify under-utilized physical host machines. A comparative analysis with state of art methods is carried out by using the CloudSim 3.0 simulator. Simulation results show the superiority of our method, able to achieve the highest accuracy values of 97.2% for accuracy and of 97.9% for precision rate, thus confirming the efficacy of our approach for virtual machine placement in cloud environments.