Analysis And Optimization Of Cloud Computing Task Scheduling Algorithm

Cloud computing systems need to run billions jobs every day,each job will occupy a certain of cloud system resources.In face of such daunting challenges,cloud computing system virtualizes multiple virtual devices based on underlying physical resources with its powerful virtualization technology.Based on these virtual devices,cloud computing system can carry the mass of job tasks at all times.In order to make these job tasks executed efficiently in order to meet user requirements for QoS,task scheduling algorithm of cloud computing has a very important significance.The methods about studying on cloud computing task scheduling strategies are as follows:(1)analying the cloud system structure;(2)task scheduling algorithm is an NP-complete problem,a number of experts have researched for solving NP problems,based on previous research,thinking about how to further innovation,making more efficient solving.Based on the above points of view,the paper analyzes the cloud computing scheduling in detail,achieving a theoretical innovation and has conducted experiments.The innovation of this paper is as follows:(1)Spark and Hadoop are the current mainstream technology of parallel computing in the cloud,the paper analysises the scheduling techniques on them,including FIFO scheduler,capacity scheduler,fair scheduling and so on.(2)This paper presents a model of cloud computing system TSS,as a research background cloud model of task scheduling strategy,the model covers the upward queue of the task,the process of virtualizing underlying physical resources as well as scheduling procedures,etc.(3)On the basis of the analysis of ant colony algorithm,we propose chaos ant colony algorithm based on ant colony optimization algorithm according to the defect of falling into local optimum solution easily.Experiments by chaos ant colony algorithm show that it has a lower probability to fall into local optimum,but still this is the case.In order to overcome this drawback,this paper integrates genetic algorithm and chaos ant colony algorithm innovatively,which show it runs more efficiency.