Optimization And Research On Task Scheduling Algorithm In Cloud Computing

With the rapid development of grid computing, pervasive computing and computer-communication technology, more and more people wanted to put the resources, software and application together in integrated way, and provide them to the user in the form of services, so cloud computing emerged as the times require. In order to achieve the integration of resources and services, a more general and service oriented architecture is needed. At the same time as the cloud computing environment is heterogeneous, distributed, autonomous and service diversity, higher requirements has been put forward on the cloud platform scheduling mechanism, so more and more attention has been put on the study on the cloud architecture and scheduling mechanism.Firstly, this paper analyzed in-depth and researched the basic theory of cloud computing, cloud architecture and cloud task scheduling mechanism, and summarized the problems existing in the current cloud architecture and scheduling mechanism. Aiming at these problems and combining with the characteristics of cloud computing, this paper mainly do the research and improvement of the following two aspects:(1) An improved MapReduce model is proposed for the large complex tasks request in the cloud computing environment. Before Map is started, the improved model firstly transforms the large complex tasks request into a DAG map, and then transforms it into a minimum spanning tree, simplifies the execution process of complex task, reduces the execution time of task, effectively improves the execution efficiency of task. The improved model enhances the performance of Hadoop architecture, enlarges the application field of Hadoop architecture. The Hadoop architecture can be used not only in the interactive application, but also can be used in the calculation of science.(2) In view of the existing shortcomings in task scheduling using the basic ant colony algorithm in cloud computing environment, the paper improves the basic ant colony algorithm, proposes BBPA algorithm based on the brothers of ants and remaining pheromone of life, then simulates the BBPA algorithm on the CloudSim platform. The simulation results show that, the improved ant colony algorithm has better task searching efficiency and executing efficiency.Finally, this paper summarized the innovative aspects of the scheduling model and scheduling algorithm, and prospected the research direction of the future development of cloud computing and task scheduling.