Energy-efficient Virtual Machinescheduling Algorithm In Cloud Computing

Cloud computing, a new service model emerged in 2007, is changing the information technology industry. Through the integration of physical resources, cloud computing service provider provides users with on-demand calculation ability and storage capacity on a pay-as-you-go basis. With the rise of cloud computing, large-scale data center with up to thousands of physical nodes are built around the world, which the attendant is a large energy consuming. Data center energy consumption, especially electrical energy consumption, not only improves the cloud computing service provider’s operating costs, but also releases a large amount of carbon dioxide which pollutes the environment.Currently, virtualization technology provide datacenter with a brand-new solution for high energy consumption. Using live migration technology, virtual machines can be migrated to a small number of physical nodes without interruption. However, the virtual machine migration process will reduce the service performance, and server overload situation occurs occasionally, which impacts the user experience. It will affect the economic benefits of cloud computing service providers.To solve the above problems, a new energy-efficient scheduling algorithm is proposed in this thesis. The algorithm takes both the energy consumption and the frequent migration of virtual machine problems into account. On one hand, through the virtual machine live migration technology, it can integrate existing resources, and improve resource utilization to reduce data center energy consumption. On the other hand, the proposed algorithm can alleviate the problem of virtual machines’ migration frequently to a considerable extent, which ensures the quality of service to users. The main contributions are as follows:Firstly, This thesis uses genetic algorithm to solve the problem of high energy consumption in data centers. For different data center configuration, the thesis design a generic chromosome structure with region, and will restrict migration of virtual machines between homogeneous physical machines. It will get a more energy-efficient virtual machine-physical machine mapping scheme.Secondly, by comparing the original placement configuration with the new one, a specific migration process will be calculated. In the migration process, it must guarantee a minimum of virtual machine migration. In this thesis, the problem is transformed to a maximum weighted independent set problem, and a greedy algorithm is designed to get the specific virtual machine migration process.Finally, we get VM’s workload data by tracing from more than a thousand VMs, and have simulated in Cloud Sim platform. The experimental output shows that the proposed algorithm significantly reduced the energy consumption while avoiding the frequent VM migration.