云计算平台中非定期任务并行调度仿真

由于云计算平台的动态不确定性和非定期任务调度本身的复杂性,使得非定期任务调度过程中的耗时长和负载不均等问题很难得到有效解决。针对上述问题,提出一种非定期任务并行调度方法,并应用到云计算中。通过多方面考虑云平台客户非定期任务的截止时间底线、调度估算等并行调度约束条件和各种可用资源的性能参数,对非定期任务调度的多目标约束条件进行量化建模。基于建模生成的隶属度函数将非定期任务多目标约束的调度优化问题转变成单一目标约束问题,采用模拟退火算法对该问题进行求解,最终实现对非定期任务的并行调度。分析实验结果可知,与传统方法相比,所提方法能够有效减少非定期任务的传输时间,并且能够均衡节点负载,表明所提方法具有有效性。

Cloud Computing Platform Non-Recurring Task Parallel Scheduling Simulation

Due to the dynamic uncertainty of cloud computing platform and the complexity of the non-periodic task scheduling,it is difficult to solve the problems about long time and uneven load in task scheduling process.Therefore,a parallel scheduling method for non-periodic task was proposed and applied to the cloud computing.The multi-objective constraints in non-periodic task scheduling process were quantitatively modeled by considering the deadline baseline,scheduling estimation and other parallel scheduling constraints of non-periodic tasks on cloud platform and the performance parameters of available resources.Based on the membership function generated by the model,the scheduling optimization problem of multi-objective constraints of non-periodic tasks was transformed into a single objective constraint problem.Simulated annealing algorithm was used to solve this problem.Finally,the parallel scheduling of non-periodic tasks was achieved.Experimental results show that compared with the traditional method,the proposed method can effectively reduce the transmission time of non-periodic task.Meanwhile,this method can balance node loads,so that the proposed method is effective.