Data Dependencies Based Workflow Scheduling In Cloud Environment

Cloud environment provides rich computing resources that can be flexibly used on demand for the efficient execution of workflow and attracts more and more users.For multiple workflows with data dependence and timing dependence,how to effectively reduce the amount of data transfer and reduce the overall completion time of workflows is an important problem faced by workflow scheduling in the cloud environment.In order to solve the above problems,this paper proposes a workflow scheduling method based on data dependency.Firstly,workflow is modeled in the form of process algebra and the formal definition of workflow is simplified.Secondly,the estimated starting and ending time of each task is calculated according to the execution time of each task in the workflow and the time-dependent relation.Then the data transmission amount of each path in the workflow is calculated according to the data dependence relation between each task.Based on the principle of selecting the maximum data amount of each path,the workflow is divided into main task set and non-main task set.Where,the tasks in the main task set are uniformly scheduled to a master server;the tasks in the non-master task set will be scheduled to the dynamically updated idle time of each server according to the estimated starting time and estimated ending time,so as to realize the comprehensive optimization of data transmission amount and completion time of multi-workflow.In this paper,the feasibility of this scheduling method is demonstrated through a concrete example,and a simulation experiment is designed to demonstrate the optimization effect of the workflow scheduling method based on data dependency on the total amount of data transmission and completion time by comparing with other scheduling algorithms,which verifies the effectiveness of this method.