一种满足可靠性和能效的云工作流调度方法

为了同步解决云工作流调度时的失效和高能耗问题,提出一种基于可靠性和能效的工作流调度算法.算法为了在截止时间的QoS约束下最大化系统可靠性并最小化调度能耗,将工作流调度过程划分为四个阶段:计算任务优先级、工作流任务聚簇、截止时间子分配和任务调度.算法在满足执行次序的情况下对任务进行拓扑排序,并以通信代价最小为目标对任务进...

A CLOUD WORKFLOW SCHEDULING METHOD MEETING RELIABILITY AND ENERGY EFFICIENT

In order to solve the problem of failure and high energy consumption of cloud workflow scheduling synchronously,a workflow scheduling algorithm based on reliability and energy efficiency is put forward in this paper.In order to maximize the system reliability and minimize the scheduling energy consumption under meeting deadline QoS constraint,our algorithm divided the workflow scheduling process into four stages:calculating task priority,workflow tasks clustering,deadline sub-distribution and tasks scheduling.It sorted the tasks topologically in the case of meet the execution order,and clustered tasks with minimizing communication cost;it divided the deadline among all tasks,and scheduled all tasks after clustered with the right frequency/voltage level,which could minimize the system power consumption while ensuring the reliability of tasks scheduling.Through the integrated simulation test of randomly task graph and Gaussian elimination task graph,the results show that our algorithm performs better than the baseline algorithms on reducing the total execution energy consumption and improving the workflow scheduling reliability.