Affiliations: [a] School of Computer Technology and Engineering, Changchun Institute of Technology, Changchun, 130012, China | [b] School of Science, Changchun Institute of Technology, Changchun, 130012, China
Abstract: In cloud computing environment, a larger number of tasks are executed simultaneously, and therefore task scheduling strategy is a key factor to determine the performance of the system. For the problem of static scheduling regarding related tasks in cloud environment, this paper minimizes the scheduling length and keeps load balance as the main goal. By combining the list scheduling and task duplication algorithms, a new task scheduling algorithm is proposed. This algorithm is composed of three steps of operations. Firstly, a task scheduling queue is constructed by computing the priority value for each task; secondly, in order to reduce communication latency between tasks, the parent tasks of the current task are duplicated selectively by taking advantage of the timeslots of the current virtual machine rationally; lastly, each task is assigned to the virtual machine that made the task have the earliest execution time and keep load balanced of system. The experimental results display the influence of the number of tasks on the performance of the algorithm when CCR is different and the new task scheduling algorithm balances the loads among virtual machines in the cloud computing system, and improves the resource utilization effectively.