Research On Scheduling Methods For Deadline-Constrained Security Workflows In Cloud

With the continuous development of workflow technology and its wide application in different fields,more and more workflows need to be deployed in the cloud computing environment for scheduling execution to obtain powerful computing and storage resource support.Users need to pay corresponding fees to use cloud resources,and the lower the cost means the slower completion time.Therefore,the main concerns of cloud workflow scheduling research are mostly focused on the optimization of cost and time.However,there are certain task execution risks and data transmission risks in the cloud environment,which may affect the normal execution of workflow.Therefore,security is also one of the important factors to consider in cloud workflow scheduling.In the light of the above-mentioned problems,this thesis presents a scheduling method for secure cloud workflow with deadline constraints.The main work is as follows.(1)Cloud workflow scheduling method based on dynamic time weight is proposed for cloud workflow considering security risk optimization under deadline constraints.The core idea is that the server with the smallest weighted sum of time and risk is selected to execute the task.The feasibility and availability of this method are tested by a simple workflow instance.Under various experimental parameters,it is indicted that this method can diminish the security risk of workflow in contrast with other methods.(2)A cloud workflow scheduling method based on Q-Learning is proposed for the cloud workflow considering both security risk and cost optimization under deadline constraints.A reinforcement learning model for multi-objective cloud workflow scheduling problem is constructed.Then the idea of Q-Learning is used to optimize security risk and cost.Compared with other algorithms under diverse experimental parameters,it is found that the proposed method can diminish the security risk and cost.When the risk weight is 1,that is,only security risk is considered,compared with this method,the scheduling method based on dynamic time weight can obtain better results.(3)A secure cloud workflow scheduling prototype system with deadline constraints is designed and implemented.The mainly functions are workflow generating and scheduling.The workflow generating can generate workflow randomly according to the parameters.The scheduling executes the scheduling algorithm and displays the scheduling results of the proposed method with a chart.