The Fragmentation And Scheduling Method For Cloud Workflow

The development of cloud computing has injected new vitality for enterprises and government informationization and the Internet industry, but it encountered some challenges at the same time. For example:under the premise of guaranteeing the high service quality, how to shorten the execution time of multiple instance workflow. The emergence of cloud workflow provides a good solution for this problem. At the same time, the cloud workflow scheduling problem is core problems in cloud workflow.Cloud workflow scheduling is based on the traditional workflow scheduling, but it also exists difference with the traditional workflow scheduling. Thus, it brings some new problems to be solved, such as:how to reduce the communication time distribution in different execution resources among tasks? How to allocate resources rationally so that all of the workflow instance execution time is the shortest?In view of the above problems, the main work carried out the following:(1) In order to reduce the communication time between tasks in the workflow instance, put forward clustering based adaptable fragmentation method for decentralized workflow. And it is compared with other relevant fragmentation methods through the simulation experiment. The results show that the proposed method in terms of response time and throughput have certain advantages.(2) For all the execution time of workflow instances to be shortest as the goal of cloud workflow scheduling problem, reasonable cloud workflow scheduling model is established. And then, in terms of priority, based on the improved Bottom Rank (BR) values of the sorting method, put forward the Deadline-based Bottom Rank (DDBR). In turn, batch scheduling model is used in this paper. Finally, combining the results of workflow fragmentation, fragmentation based genetic algorithm (FBGA) is proposed through improving genetic algorithm. The method considered the results of workflow fragmentation when the genetic algorithm is initialized and mutated. Thus, it narrows the scope of the solution space, and is helpful to achieve a better solution. And it is compared with other related workflow scheduling methods through the simulation experiment. The results show that the proposed method in terms of all workflow instance completion time and optimization efficiency has a good promotion in some cases.In this paper, the proposed workflow fragmentation method and cloud workflow scheduling method make up the defect of the existing research, enrich and improve the theory and method of workflow management, provides theoretical guidance for the cloud workflow, but also has good application prospect.