Research On Tasks Scheduling In Cloud Computing Environment

Cloud computing is composed of a series of dynamically scalable resources, which are provided to cloud computing users through virtualization technology. Users can visit cloud computing, and rent resources by demand, so that the client’s burden can be reduced and users could enjoy the strong computing power of cloud computing. Now, many enterprises set up cloud computing servers. Cloud computing service providers should have their own schemes of resource allocation and task scheduling, but there is not any unified dispatching specification and standard. Therefore, it’s important to do some research on task scheduling and resource allocation in the cloud computing environment. The paper makes a comprehensive study on characteristics of cloud computing and grid computing, and main research contains the following three parts:(1) A scheduling algorithm that adapts to the changing tasks in the cloud computing is proposed. On the condition that mass user tasks are scheduled with limited resources, for the reason of huge user groups in the cloud computing, there is the leap of situations that users remove or delete the task for personal factors after committing. All these matter would lead to waste of cloud resource and impact on cloud computing service provider benefits. Existing scheduling algorithm can not solve the matter well. To address the problem, the paper proposed a new scheduling algorithm. Based on the DAG, on the time of removing task, the algorithm updates the DAG according to its dependence on removed task, and then uses a heuristic algorithm to schedule. The result in CloudSim shows that the algorithm in this paper avoids the scheduling execution of revoked tasks and can improve use ratio of cloud computing resource, while the algorithm is better than Min-Min and Max-Min in executive time span.(2) In consideration of user tasks preference, virtual machine resources characteristics and customer satisfaction, a cloud computing scheduling algorithm based on evolutionary game model is proposed to allocate the resource to the task in changing cloud resources environment, and the scheduling algorithm is able to ensure the quality of distinguish user tasks. Cloud computing tasks compete for resources by means of kinds of preferences, and virtual machine resources compete for tasks according to their quality assessments (such as calculation type, storage type, bandwidth type), all of which can form a mixed game. The scheduling algorithm can improve scores of virtual machine resources and their populations according to task scheduling information and users’feedback score. Finally, the equilibrium of the game is the virtual resources’stable structure, that is to say, each node and population can display their actual performance which can achieve equilibrium. The simulation experiment result indicates that the algorithm in this paper is available and efficient. The algorithm can accord to the types of tasks assign different characteristics of the virtual machine resources, then improve scores of virtual machine resources by users’feedback score to reach the virtual resources’ evolutionary stable structure, and get equilibrium. In the meanwhile the algorithm guarantees that users of different kinds can have a better QoS.The paper has made some achievements in cloud computing task scheduling and proposed the scheduling algorithm which can adapt to new feature of cloud computing better. This paper provides some ideas concerning the analysis and solution for the problem of cloud computing tasks scheduling. It also has some significance for the research on cloud computing tasks scheduling.