Research On Virtual Machine Scheduling Optimization For Energy Efficiency In Cloud Datacenter

As computer theory and technology fast promoting, marvelous services areprovided with the help of various embedded, mobile devices, high performance serversand the Internet. Cloud datacenter played an important role, but it also brought seriousenergy problems. Under the auspices of the global concept of sustainable green energysaving, research on energy-efficient algorithm for cloud datacenter becomes a trend.With the increasing energy expense in current data centers, many emergingenergy-saving mechanisms have been proposed to reduce the energy consumption. Mostof these methods assume data center is in a homogeneous environment. However, mostof current data centers are heterogeneous as different types of servers are purchased atdifferent time in reality.This paper analyzed the energy consumption characteristics of heterogeneouscloud datacenters, constructed the energy consumption model of virtual machine, andconsolidated virtual machines on the overloaded and underloaded servers respectively.From the perspective of virtualization, we schedule the virtual machines by introductingthe parameters “Performance/Power” of servers. That s the proposed energy-efficientmethod PVMAP, which is aimed to make the server with high energy efficiency fullyutilized with high priority in the dynamic VM consolidation. Results obtained fromsimulation experiments demonstrate that the PVMAP method is better than existedmethods in the scalability and stability.In addition, we realized PVMAP within the xGreen system which is anenergy-saving management platform based on OpenStack. Its performance isverified by the experiments using benchmark data.