Study On Modular Data Center Energy-saving Management Based On Virtualization Architecture

In recent years, with the rapid development of application of enterprise informationization and mobile internet, the number and size of the data center continues growing, which increasing density and complexity of data center, as a result, data center energy cost accounts for a growing proportion of the total cost. According to IDC, energy cost of data center caused by large scale infrastructure and server increases at an average annual rate of 20%, and is associated with 50% of capital expenditure and about 60% of operating costs[1]. However, due to the low utilization efficiency of overall data center resources, the idle server or excessive cooling consume majority of energy, thus energy saving of data center has been deserved great attention. With the "software defined data center" concept was put forward, the data center virtualization will become a new development direction of the data center.This paper research on the construction of green data center virtualization, by means of virtualization turn physical resource into a set logical resources to serve for data center applications, and improve utilization of physical resources through the dynamic scheduling management of physical resources, so it accomplish saving energy costs of data center. Specific studies as follows:(1) Research on energy optimization of virtualization of data center. First of all, identifying key units of energy consumption and dissipation equipment in data center through analysis of factors of the data center energy consumption factors, and establishing energy consumption model of whole data center and key energy consumption units. Then, by the way of physical resource virtualization management, it establishes virtual management model of main energy consuming equipment of data center, to provide important theoretical basis of dynamic energy management of data center.(2) Research on data center’s computing resource virtualization energy saving management. Firstly, according to the virtual server migration theory, it establishes model of server virtualization energy saving scheduling process. Then provides preparations of computing resources dynamic scheduling through the choice of strategy of design of virtual machine migration occasion, migration object and target host. At last, based on construction of dynamic calculating resource migration entity, it designs computing resources dynamic migration method and migration scheduling algorithm to manage the data center’s computing resource energy saving.(3) Research on dynamic ring resource energy saving management in the foundations of cooperative scheduling. According to the characteristics of data center, it establishes essential model of data center’s dynamic ring resource collaborative manage, and put forward the collaborative management strategy of three main sub data center’s systems: computing resources, cooling resource and power distribution. Through the deep study of distribution of interpolation of data center’s temperature field and evaluating method of temperature field hot point, it can effectively identify the distribution of requirement of the data center ring resource, and do precision cooling management to data center in the foundation of demand supply principal. At last based on such above collaborative energy saving management, it designs architecture, collaborative scheduling rules and scheduling algorithm of data center’s collaborative scheduling.(4) Evaluation on data center energy efficiency management. Firstly, based on the studying of data fusion theory, provide a theoretical basis for analysis of data center’s energy consumption data fusion. Then according to features of data center’s energy consumption data source, it put forward method of data center’s energy consumption data fusion. Based on deep study of a variety of energy efficiency evaluation method, it designs evaluation method and comprehensive model of data center energy management, which can provide reliable instruction to data center’s scheduling energy optimization.(5) Analysis of data center’s energy management application. Based on above research, combing with actual work, it integrally designs data center energy saving software system, deploy it to the application of actual project, and analysis the effect of data center’s energy optimization management.