Energy optimized VM placement in cloud environment
Amandeep Kaur, Mala Kalra · 2016
Large energy hungry datacenters are increasing at an alarming rate, making energy management as one of the key design constraints for the cloud datacenters. Energy consumption in cloud datacenters can be reduced by employing VM consolidation technique which is the process of packing VMs on the least number of servers and VM consolidation uses VM migration technique for this purpose. But unnecessary migrations may lead to performance degradation, service level agreement violations, service down time. Efficiency of system that employs virtualization greatly depends on the technology used to allocate VMs to the appropriate hosts. This gives significant importance to VM Placement Problem (VMPP). The aim of VM placement problem is to put the idle server to low-power consuming mode and to reduce the number of overall migrations in a system. In this paper, an algorithm based on MBFD has been proposed for VM placement. The goal is to reduce the number of active servers and to obtain a stable host for each VM such that the number of unnecessary migrations and total power consumption can be reduced. The performance of proposed algorithm is compared to that of an existing MBFD placement algorithm. The result shows that the proposed algorithm saves more energy than the method compared to.