Rearch And Implementation Of The Taskscheduling Algorithm Based On IVCE Platform

As a new computing mode, cloud computing become a business and hotspot in the research at home and abroad. Cloud computing system has a very complex structure, the task scheduling is one of the technologies in cloud computing. The study of task scheduling is very helpful to handle user service requests and improve the quality of service. In addition, the cloud on processing information with greater security flaws, how to provide a more secure and more trustful cloud computing environment has become a key technology in the study of cloud computing.Inorder to solve these problem, this paper proposed a task scheduling algorithm with task clustering in the base of the practical operation of iVCE platform. Also, this paper joined in the task scheduling system of communication message authentication mechanism to improve the security of the platform. The innovations in this paper as listed below:(1)Improved the DBSCAN algorithm. The dissertation analysis the deficiency of the traditional DBSCAN clustering algorithm is that there will be a poor result if the parameter selection for Eps is not good. And because the algorithm takes the global unified value of Eps, there will be a poor result if the data set is uneven. For these deficiency, before clustering the data, the improved algorithm use the particle swarm algorithm to obtain appropriate parameter value of Eps to improve the clustering effect. And use the traditional DBSCAN algorithm and the improved DBSCAN algorithm to cluster the standard data set about Iris. The accuracy of the former clustering is 54%, and the average accuracy of the latter is 83.27%.It can be seen that the improved algorithm significantly improves the accuracy of clustering.(2) Applied the improved clustering algorithm to the scheduling system.The dissertation use the clustering algorithm to iVCE platform and cluster all tasks to clustering of resources consumption. After that all the tasks will be clustered to task of computing, task of bandwidth and task of disk. Then the dissertation schedule the clustered tasks, such as the dissertation schedule the computing task to the computing resource. And compared with the Min-min scheduling algorithm, the scheduling algorithm based on the task clustering has a more prominent performance about the success rate of task execution, resource utilization, and reduce the task execution time.(3) In view of the existing cloud computing task scheduling in the unsafe factors that exist in the system, such as an attacker posing as a legitimate user to send a lot of garbage tasks to the virtual resources, this dissertation added the add a message authentication mechanism during the user layer submitting the task scheduling and task scheduling layer distribute the task to the resource. In this project, the dissertation use the HMAC - SHA1 algorithm. And the dissertation use the Mersenne twister algorithm to generate pseudo random sequence, then the key in the HMAC-SHA1 use the generate pseudo random. This scheme ensure the security of the task scheduling system and improve the quality of service about the iVCE platform.Finally, a scheduling system is proposed and the dissertation designed and implemented the proposed scheduling system and the dissertation introduce the function and the concrete implementation of each module about the task scheduling system. And the dissertation test functionality and performance of the system, it can be seen from the test result, the proposed scheduling algorithm has certain advantages compared with the traditional Min - min algorithm in these aspects, such as the success rate of task execution, resource utilization, the task execution time and safety.