Research On Electric Power Cloud Computing Resource Scheduling System

Cloud computing can assign the computing task to the different computing centers or LAN composed of a very large number of computers. This will speed up the deployment of the applications, and reduce the expense on purchasing the equipment, electric power and the cost of maintenance. Besides different kinds of application systems can get the computing power, storage space and information service according to the needs. Nowadays the electric power systems become a complicated information system that have a considerable amount of data. How to use the cloud computing technology to integrate the exiting software and hardware resources of the electric power information system, improve the on-line data analysis and real-time monitoring capacity, so as to ensure the intelligent scheduling of the electric power grid in a long time of the future is the emphasis of this research.The virtual machine is the main resource of the cloud computing environment. By using the virtualization technology, we can build a virtual resource pool. After that the virtual machines can be dynamically deployed for user to use. For those reasons, this paper makes a study of the electric power cloud computing resource scheduling system, and the research focus on the electric power cloud resource’s scheduling of the virtual machines. Firstly I made an analysis of the resource scheduling’s research situation in the present cloud computing environment, and then made a summing up of the scheduling algorithm strategy, also analyzed the Map Reduce scheduling model and according to that I gave out a kind of virtual machine resource’s scheduling algorithm based on the Ant colony algorithm. At last I did a simulation experiment with the simulation software "CloudSim", the result indicated that, to ensure users’ service, the virtual machine resource’s scheduling algorithm based on the Ant colony algorithm was more efficient to find a appropriate virtual machine resource for tasks’ execution requirement than the Polling algorithm. At the same time I designed a kind of system architecture of the virtual machine, implemented the electric power cloud virtual machine resource’s scheduling strategy as the paper said above.