Research On Resource Scheduling Algorithm Based On Virtual Resources

With the emergence and rapid development of cloud computing,"cloud computing" is gradually from theory to practice, the "cloud" is being used as a new resource by more and more users, and cloud computing has become the focus study in the academia and industry. The key of cloud computing is to use the mature virtualization technology to make the traditional infrastructure into a common pool of computing resources, and make resource pool resources for unified management scheduling. So the cloud resource scheduling problem becomes the issue not be ignored. While cloud providers have provided different types of virtual resources to meet the different needs of users. The current mainstream resource scheduling algorithm is considered from the system energy-efficient and the quality of request service, and they do not take into account differences between virtual resources provided by cloud providers.In this paper, the author surveyed the key technologies of cloud computing, especially on the basis of virtualization technology, researched and summarized the current commonly used cloud computing resource scheduling algorithms, and based on the previous algorithm, considering the differences between virtual resources, proposed the resource allocation algorithm based on the heterogeneous virtual resource RA-HR; while ensuring low system resource usage and fairness between virtual resources, the author proposed a two-tier resource scheduling framework DTRAC. RH-HR firstly divided the task submitted by the user into different combinations in accordance with the at least the principle of free memory from a global perspective of mission packages considerations, then according to the difference between cloud resources, allocated resources on the corresponding combination. DTRAC has two layers, including a request scheduling layer and local resource allocation. Request scheduling layer dispatched the application requests to the minimum server resource utilization using the feedback server resource usage back; resource scheduling layer first clustered virtual resources on the server, and then allocated resources for the request from each class according to the proportion of demand, to ensure the fairness between the allocation of resources among virtual machines on the same server.In this paper, the algorithms were simulated by Cloudsim tools. In the simulation RA-HR, paper has simulation results from full virtual resources and inadequate virtual resources:When the cloud provider provided different types of virtual resources, RA-HR is better than traditional algorithm in the use of virtual resources. RA-HR algorithm can have better performance in saving the virtual resources in the premise of meeting the needs of users; while DTRAC has the simulation results: DTRAC is superior to the traditional algorithm in system resource usage, the fairness of the distribution of resources. DTRAC can effectively reduce the resource utilization of system resources, and can have a good effect on the fairness in allocation of virtual resources.