A study for task time performance dynamic prediction model in cloud resource scheduling

The unique of resource utilization method in cloud computing determines that there are strict requirements for the resource management and scheduling, as virtualization technology is widely applied in cloud computing, it extremely enhances the dynamic of the system and improves the efficiency of resource utilization. However, it also brings the resource scheduling with great challenges. Making an accurate prediction of task workload will improve the efficiency of cloud resource scheduling, and this paper regards this as the starting point, makes the task completion time as a metric of task workload, proposes a dynamic prediction model for task workload, studies the statistical property of task completion time in different load conditions, and the simulation experiment results show that the prediction model is correct and feasible.