Energy Saving Methods On SDN-Based Datacenter

With the high-speed development of IT industry, the demand for large cloud services has been growing, and the calculation, processing, storage of a large amount of data have become an irreversible trend.The larger scale datacenter in order to meet these requirements appears constantly, while energy consumption goes up exponentially. There is no doubt that huge energy consumption will incure additional economic costs, reduce the data center network performance, and bring the problem of environmental protection. As a result, the energy conservation of datacenter has gained increasing interest recently.Software-Defined Network(SDN) decouples the control plane and the data plane of the network device. Due to features of centralized control, flexibility, and programmability, SDN has been widely used in many occasions. In the research of energy saving method of SDN- based datacenter; the paper made full use of the characteristics of SDN to control network traffic centrally, monitored resources utilization of datacenter in real time and optimized virtual machine scheduling method.Currently energy-saving methods of SDN-based datacenter are classified into network traffic aggregation and virtual machine consolidation. The former focus on reducing the network energy consumption and the latter saves energy by closing idle server. However, the network energy consumption and energy consumption of the server are both important parts of the datacenter energy consumption. Therefore, considering both factors can achieve an optimal performance for energy saving.This paper considerd the energy saving method of datacenter in the time dimension, including the initial creation of a virtual machine, live migration phase, and network routing phase. First of all, the paper analyzed the energy consumption model of datacenter. On this basis, the paper designed the whole framework of energy saving method and put forward the Routing Aware Energy Best Fit algorithm(RAEBF). RAEBF can reduce energy consumption of the server through virtual machine consolidation while making use of exclusive routing algorithm to schedule network flows between the virtual machines, so as to optimize the network energy consumption. The live migration of virtual machine can ensure QoS of datacenter, and can improve the utilization of resources, while avoiding the energy consumption of low resource utilization server. Finally, by using the CloudSimSDN simulation platform, RAEBF algorithm has been demonstrated the advantages of reducing energy consumption, and the simulation experiment has proved the benefit of the virtual machine live migration of effectively guaranting high performance of sever and optimal utilization of resources.