Dynamic Scheduling Control Of Cloud Resource Based On DES

With the development and breakthrough of the internet technology, the virtualization technology and the big data technology, the Cloud Computing industry booms well, the model of the cloud service is becoming more and more flexible and convenient. But, due to the cloud environment has the nature of highly dynamic and distributed, it is hard to have a standard and unite cloud resource management scheduling policy at present, so the resource management scheduling policy has received considerable attention in recent years.Considering that the Cloud data center entities have the nature of discrete, in this paper, we innovatively propose the scheduling control framework of cloud resource based on discrete event system(DES) after learning the existing references about the cloud resource scheduling research. First, we abstract the whole cloud resource platform as discrete resources entities. Then, we establish the resources allocation model based on DES, and define the status and events of the resources allocation system, by controlling the occurrence of the satiable events efficiently to change the status of cloud resources. Based on this, the paper continues analyze the dynamic scheduling of cloud resources based on the distributed DES, where the cloud environment be constructed as the distributed DES that is composed of the physical server subsystem, the virtual machine subsystem and the cloud task subsystem. We keep the scheduling that use the resource fragments first, manage the VM efficiently and supervisory the state of tasks’ performance. By using the local controllers to control the occurrence of the integration of the resource fragments events, the task decomposition events and the VM scheduling events, to ensure the resource scheduling in each subsystem which will improve the resource utilization and the quality of cloud service. Meanwhile, we apply the distributed DES control theory to the resource scheduling. When uncontrollable events appear in the sub-system, the local controller will send message to the global controller, and the global will contact other local controller to scheduling the resource. By using the information exchange between the global controller and the local controller, scheduling the resources effectively and keeping the resource load balancing, we ensure the utilization of the cloud resource and the quality of the cloud service, and improve the performance of the entire cloud platform.By the simulation experiment on the CloudSim platform, and the comparison with the Min-Min resource scheduling and the FCFS resource scheduling is further. The result shows that our dynamic scheduling of cloud resources based on the distributed DES may well describe the allocation events on a global perspective and efficiently control the events occur, which has the obvious advantages on improving the utilization of cloud resources and the quality of the cloud service.