QoS Cloud Workflow Scheduling Algorithm Based On Critical Task Duplication

Cloud computing is a new type of computing model.It is combined with the business service model and supported by the academic and industrial circles.At the same time,cloud computing technology can provide high utilization of computing resources and the overall performance of computing,making more workflow applications implemented in the cloud environment,such as scientific computing,multi-layer Web and large-scale data processing applications.The difference in cloud computing is that it pays for the use of service resources,that is,users can choose the appropriate network services according to their actual needs,and it is more convenient and fast,and the cost of management is low for users because they do not need too much interaction with the service providers.The new features of cloud computing make the scheduling of workflow applications in the cloud environment an important research topic.Because the task resource mapping is a NP difficult problem,we need to allocate resources according to the different QoS requirements of the user,and the performance of the cloud environment is unstable,and the data storage and inter task communication optimization need to be considered.Most of the current cloud workflow scheduling algorithms are transplanted in grid scheduling,which can not meet the current scheduling needs.The main work of this article includes the following four aspects:Firstly,it studies the different scheduling strategies under the cloud workflow,introduces the current mainstream scheduling algorithms in detail,analyzes the low complexity list scheduling algorithm and the scheduling algorithm based on duplication,and studies the existing problems of each algorithm and the application defects in the cloud workflow.This paper also explains the scheduling process of workflow in detail,and makes mathematical modeling and explanation for workflow,cloud resources,QoS preferences of users and task scheduling.Secondly,combined with task duplication scheduling algorithm and user's QoS requirement under cloud workflow,a scheduling algorithm named BSTD for critical task duplication is proposed.The algorithm is based on the user's QoS requirement preference,which can reduce the task communication delay by scheduling the forward drive to the processor's idle time period of the processor to duplicate the critical task.This static scheduling algorithm can effectively improve the scheduling efficiency and meet user's different needs.Thirdly,because of the instability of cloud environment,this paper proposes a dynamic scheduling strategy in the execution of cloud workflow,which optimizes the scheduling performance when the task delay exists.Lastly,the workflow is simulated and dispatcher through the cloud computing simulation platform CloudSim.First import static BSTD algorithm and dynamic strategy.Then,the multi group of DAG instances are generated according to the different communication computing ratio,and the cloud platform is extended to complete the static scheduling of the workflow and the dynamic scheduling with tasks delay,comparison of other related scheduling algorithms,experimental results show that the algorithm proposed in this paper performs better.