Research Of Task Scheduling Algorithm In The Cloud Environment

With the progress of academia and industry, Cloud computing is moving from theory to practice. Some hot issues are discussed and researched widely on the researching process of Cloud computing, and task scheduling strategy is one of them. Task scheduling strategy in Cloud environments focuses on how to build a mapping relationship between tasks and resources. Furthermore, migration of virtual machine technology in Cloud environment can provide a reliable guarantee for task execution, one of the key points of migration is the memory pre-copy strategy.Based on theses issues mentioned above, the work of this paper can be mainly included from two aspects:1) When ant colony algorithm is used to schedule tasks in Cloud environments, the repeated searches of ants may lead to low searching efficiently.In order to solve this problem, the Task Scheduling Strategy base on Cooperative Ants (TSSCA) is proposed in this paper. In this algorithm, the concepts of Brother Ants and RLP (Rest-Life-Pheromone) are proposed to help ants communicate with each other. On the basis of ants’cooperation, the proposed strategy can reduce the invalid searches, expend searching range of ants, and find better execution nodes for tasks. Simulation results show that this algorithm can improve the searching efficiency and accuracy of ants when they are used to schedule tasks in Cloud environments.2) Memory migration is a core issue in the process of virtual machine migration, and the pre-copy memory migration strategy is the most common strategy of memory migration. In order to reduce the transfer time and data amount of transmission during the migration process of virtual machine, a memory pre-copy strategy of virtual machine based on a kind of association rules algorithm-Apriori is proposed in this paper,which calculates associated memory papers according to the Apriori algorithm before memory migration of virtual machine. When a dirty memory paper is to be sent on next round of iteration, it does not be translated if its associated memory papers are still working, which means this paper will probably get dirty. Through this strategy the duplicated transmission of memory papers and total migration time of virtual machine can be reduced efficiently, and thus the impacton the quality of service causing by the migration of virtual machine can be also reduced.