Multi-objective Optimization Policy For Energy-efficiency Management In Data Centers

With the development of big data and cloud computing technology with popularization and development in the world,cloud data center infrastructure and related supporting facilities in the number of high-speed growth.Data center with a lot of computationally intensive and data intensive operation need to be done quickly and efficiently to ensure the normal operation of the data center and meet the requirements of users for cloud service quality of service.This would require the synergy between the servers.Cooperation between mass servers will produce large amounts of energy consumption and data center for energy utilization will be improved.This makes the cloud data center operating costs on the basis of increasing cause huge waste.So the cloud data center energy consumption problem should be solved now.At present,the cloud data center energy consumption problem got wide attention of scholars.The solution of the main strategy is divided into two aspects of hardware and software strategy of energy conservation.In the aspect of software,the virtualization technology has been proved to be an effective way to solve the problem of energy consumption of cloud data center and it is the key for this paper.Virtual machine consolidation strategy is one of the most important strategies of energy-saving aspects of current software.It includes the host selection,VM selection and distribution of the virtual machine.This paper mainly focuses on the selection and distribution of the virtual machine process.Live virtual machine consolidation is the effective method to improve the level of green data center energy efficiency.At present,the green data center energy consumption evaluation model based on CPU usage as the main influence factors.However,due to the intensive processing of GPU generated huge energy consumption,evaluation model is not suitable for the data intensive computing.In this article,we based on the utilization rate of CPU and GPU proposed a new evaluation model of energy efficiency and the strategies of the two kinds of real-time dynamic migration of virtual machine.A strategy applied to virtual machine choice,another is applied to the virtual machine allocation.Some researchers have put forward respectively strategy based on the virtual machine or virtual machine allocation policy to its own solution.The virtual machine selection and distribution of the virtual machine,however,these two strategies become together will get a more efficient real-time dynamic migration of virtual machine strategy.Based on this,a fast real-time VM integration strategy based on artificial colony algorithm was put forward.DataABC by adopting the idea of the artificial colony algorithm,and it is a rapid virtual machine migration to get global optimization of virtual machine migration strategy.Traditional virtual machine allocation has the problem of the process of speed is difficult to meet the characteristics of data intensive operation and have the phenomenon such as easy to fall into local optimum and choice over and over again.Therefore,this article introduces artificial colony algorithm in virtual machine allocation process,combined with the gradient descent algorithm and simulated annealing thought,proposed a fast and parallel ABC based energy-efficiency live VM allocation policy in data centers(FP-ABC).In order to meet the needs of data intensive operation for response speed,this article introduces the gradient descent algorithm,and at the same time introduces the simulated annealing algorithm to reducing the energy consumption,improve the efficiency of resource use and reduce the operation cost of the data center.In the process of energy saving problem in the center of the research data,the researchers put forward a variety of energy-saving strategy,for example,the on/off strategy,the virtual machine consolidation strategy,DVFS strategy,etc.But,each strategy has its implementation conditions and their own characteristics,so a variety of energy-saving strategy integration will achieve a better data center energy saving.Realize the sustainable development of the data center.