Research and Implementation of Resource Scheduling Method Based on Openstack
|Keywords||cloud computing resource scheduling virtual machine migration cluster powerefficiency OpenStack|
With the rise of cloud computing, many service providers will mainly focus on the construction of data center, and many of the companies and the colleges are also establishing public clouds and private clouds. Tell the number of data centers is growing. With the emergence of this trend, an important problem emerged, that is the energy consumption of the data center server problem. Because of the low utilization rate of resources, many servers in the data center is in low load condition, cause the idle resources. In order to more rational use of hardware resources in data center, many scholars of data center scheduling problem carried on the thorough research, to save energy consumption of data center server clusters, and improve the efficiency of data center server clusters.According to above problem, this paper applied more extensive OpenStack cloud service build platform technology for data center resources scheduling problem of research. This paper firstly analyses the existing resource allocation way of OpenStack. And then online migration of OpenStack were studied, analysis the feasibility of dynamic resource scheduling under OpenStack. Combining previous dynamic resource scheduling method is proposed with the specific characteristics of OpenStack technology for dynamic resource scheduling method to carry on the design. By analyzing the load information of new aviation management system abstract the specific test task combined with CloudSim for simulating the performance of the algorithm. Scheduling the quality of the algorithm is analyzed.Resource scheduling method based on OpenStack is proposed in this paper using virtual machine migration between server hardware resources redistribution, thus achieve the goal of resource scheduling. Algorithm is mainly solved in the process of resource scheduling of virtual machine migration trigger problem determination and virtual machine migration process involves the target selection problem. In the process of solving concrete adopt the method of threshold combining prediction algorithm implementation delay trigger migration, to solve the instantaneous changes caused by unnecessary load migration. Through reasonable set the trigger condition and idle server shut down strategy to improve work state of the server utilization, reduce the work so as to realize energy consumption of the decline in the number of the server achieve energy efficiency improvement in data center server clusters.