Research On Heuristic Scientific Workflow Scheduling Based On Service-Oriented Architecture In Cloud Computing

In many scientific fields such as astrophysics,earthquake science,gravitational physics,and bioinformatics,researchers often simulate and analyze real-world activities in the form of scientific workflows.Cloud computing relies on its elastic computing power and convenient access.It provides a practical and efficient deployment environment for scientific workflows.There are still many problems to be solved in scientific workflow task scheduling under cloud computing environment: 1)Different from general independent task scheduling,scientific workflow task scheduling needs to ensure that the dependencies between task nodes are not destroyed.2)Cloud service users have different resource requirements.The design of cloud data center tasks and resource allocation strategies must take into account service-oriented issues.3)Since cloud computing concentrates a large amount of data and computing power in the cloud,while designing a scheduling algorithm that can improve user service quality,it is also necessary to achieve efficient resource utilization and load balancing.How to efficiently implement the scientific workflow task scheduling strategy in the cloud computing environment has become one of the research hotspots in recent years.Aiming at the above problems,the main research work done in this paper is as follows:First of all,according to the design requirements of cloud computing platform under the guidance of service-oriented architecture technology,the structural characteristics of cloud computing platform and SOA technical details are analyzed in detail,and the possibility of the fusion design of the two is explained,and finally the cloud computing based on SOA is constructed.The architecture model provides theoretical guidance for the construction of enterprise informationization and the improvement of user Qo S.Secondly,in view of the heterogeneous characteristics of cloud service users’ demand for resources and the consideration of the link carrying capacity of cloud data centers,under the guidance of the concept of SOA-based cloud computing architecture model,an SOA cloud data center model is constructed.In this paper,a cloud task scheduling strategy based on link awareness is proposed,which realizes task scheduling based on service type and dynamic allocation of link resources.The Cloud Sim cloud computing simulation platform is extended to support network link simulation,and the algorithm strategy is simulated on the extended platform.The experimental results show that the algorithm of the proposed strategy can be used in a short communication time and at a low communication overhead cost.,reduce system energy consumption and improve resource utilization.Finally,from the independent task scheduling research in the cloud computing environment to the scientific workflow task scheduling research,aiming at the problems of low resource utilization and single service type in the existing scientific workflow task scheduling,a cloud workflow based on two stages is proposed.Task scheduling strategy.According to the characteristics of the cloud workflow structure level,the task scheduling process is divided into the task mapping stage and the resource allocation stage.Based on these two stages,the corresponding initialization task execution sequence generation algorithm and resource allocation algorithm are designed respectively.The Cloud Sim simulation platform is also extended to support the scheduling simulation of workflow tasks with dependencies.The comparative experimental results show that the two-stage cloud workflow task scheduling strategy can effectively optimize the cost and time span of scientific workflow scheduling,and Achieve high system resource utilization.