Abstract:

In order to improve the resource utilization of cloud computing systems and optimize the system performance,by taking into account the users' QoS demand,a workflow cloud computing system is established by integrating cloud computing with workflow,and a two-stage resource scheduling model is constructed for the cloud computing workflow system.In the first stage,by considering the time and cost constraints of QoS,the dependencies among the tasks in the workflow,and the processing of the intermediate data from each task,a modified particle swarm optimization algorithm ( MPSO) is proposed,and the Pareto is used to obtain an optimal solution so as to improve scheduling efficiency.In the second stage,by considering the resource allocation on hosts,a scheduling strategy with load-aware is proposed to perform resource scheduling according to system loads,so as to improve the resource utilization of the system.Experimental results show that,in the resource scheduling process of the workflow cloud computing system,the modified MPSO algorithm is superior to the earliest heterogeneous finish-time algorithm and single-objective optimization genetic algorithm in terms of execution speed,resource utilization and users' satisfaction with QoS,and that,the proposed scheduling strategy with load-aware can schedule more efficiently according to the system loads,and the task execution efficiency and the resource utilizationare are thus improved.

Key words: cloud computing, workflow, particle swarm optimization, load-aware