Research On Static Task Scheduling Mechanism In Cloud Computing Environment

Cloud computing applies the distributed computing,parallel computing and virtualization technology to integrate a large number of heterogeneous computing resources,which forms a huge virtual resource pool to provide personalized service for users.In a cloud computing system,it is always going to handle massive data and application tasks.How to allocate the tasks to different processors in a reasonable way and reduce the computation time so as to reduce the scheduling cost,while ensuring the whole system in a balanced load and keeping a high resource utilization rate has been a major problem in the field of cloud computing.In this paper,the static task scheduling in cloud computing is researched.Aimed at minimizing the total completion time,independent task scheduling and associated task scheduling mechanism are studied.The main work of this paper is as follows:(1)According to the dependencies of tasks,the task scheduling is divided into independent task scheduling and associated task scheduling.As for independent task scheduling,the corresponding mathematical model is established based on the inherent attributes of the tasks.For associated task scheduling,a DAG(Directed Acyclic Graph)model is established according to the order constraint relationship between tasks and resources,which lays the foundation for subsequent research on scheduling mechanism.(2)An independent task scheduling strategy based on improved PSO(Particle Swarm Optimization)algorithm is proposed.Owing to the influence of the optimal location and the global optimal position in the particle optimization process,the traditional PSO algorithm is easy to fall into the local optimal solution.In order to avoid the defect,the initialization process of PSO algorithm is improved by using LBMM algorithm to pre-schedule the tasks and taking the result as the global optimal solution of particles.Further,the fitness function of PSO algorithm is redesigned according to the actual demand and the crossover and mutation operation of GA(genetic algorithm)are integrated into PSO algorithm to expand the search space of particles.Finally,the proposed algorithm and other scheduling algorithms are analyzed and compared by experiments.(3)An associated task scheduling mechanism based on list scheduling is proposed.The communication relationship between tasks and resources is directed by DAG.In this paper,the hierarchy of task nodes and the difference of processor processing ability based on traditional list scheduling are considered at the same time.Firstly,the uplink weight of each task is redefined and calculated from the export task so as to determine the scheduling priority order of each task.On this basis,the insertion strategy is applied into the allocation of the tasks to corresponding resources.Finally,the effectiveness of the improved algorithm is verified by experiments.(4)A task scheduling system in cloud computing environment is designed and implemented.In this system,actual examples are used to verify the feasibility of two algorithms proposed in this paper.