Virtual machine placement of cloud computing for energy reservation

Abstract

© 2019 Int. J. of GEOMATE. Cloud computing has been widely deployed. The virtual machines (VMs) are created in servers upon the requests from users and they are deleted from the servers after the rental time expires. This is called dynamic workload condition. VMs should be consolidated into the servers to reduce the number of servers running VMs. Servers that do not have any VMs should be changed to sleep mode to reduce energy consumption. Therefore, VM scheduling which selects servers to run VMs has to find servers to place VMs and has to migrate servers under this dynamic workload condition. However, migration also consumes energy, so the number of migrations should be limited to save energy. In this paper, a VM scheduling method called Energy-Aware Scheduling Updating (ESU) which reduces total energy consumption in the data center is purposed. It chooses servers to create VMs. In addition, it updates the locations of VMs when changes occur while it limits the number of migrations to reduce energy consumption. The performance is of ESU is evaluated by computer simulation. The results show that ESU has a better performance considering energy consumption among the protocols used in comparison.