Research On Minimum Critical Path Of The Optimization Scheduling Alorithm For Cloud Workflow

With the development of information science technology, Workflow has been produced for the growing complexity of research, work and production activities. Cloud Computing technology has become maturer, but some problems, such as performance, security, reliability problems, have emerged. However, Cloud Workflow task scheduling, as one of the key technical problems to be solved, runs through the whole process from the bottom of the Cloud Computing resources allocation to the upper service delivery. Cloud Computing has the feature of user-centric and service-oriented commercial in nature, which allows users to require reasonable scheduling strategy choices on the cost, time, efficiency, safety and other problems. The original user scheduling algorithm,which is in Cloud Computing environment with its own characteristics, has not fully optimized when combining various factors users considered. Cloud Workflow scheduling algorithm also has problems in the efficiency of expense and time. This paper will combine the user’s expectation, that is to achieve optimal scheduling strategy, and optimize method research of the existing Cloud Workflow minimum critical path scheduling algorithm to attain the win-win goal of achieving customer satisfaction and cloud resources reasonable scheduling. The existing minimum critical path algorithm does not consider the problem of margin distribution time, but schedule the minimum resource to meet the time expense in the cloud resources according to the time difference between the latest and the earliest node task is completed. Although you can select the resources which expense least within a specific deadline, the overall expense does not reach optimal state. Great deals of researches have been done in this thesis about the optimization method of the minimum critical path algorithm.The paper first of all introduces the workflow scheduling and the related definition of cloud workflow as the theoretical foundation to popularize the meaning of the critical path algorithm and also make a detailed analysis of the critical path of DAG scheduling model, accordingly, an in-depth study of its algorithm would be easily conducted. On this basis, in the paper, it is clear to see that the minimum critical path scheduling algorithm has been optimized and a workflow scheduling algorithm based on critical path has also emerged at the same moment. Undoubtedly, the significant procedure of algorithm lies on its critical path for the tasked DAG model, which is well available to complete the task time margin, namely the latest finish time of LF and the earliest finish time of EF. Then, according to the established rules, the remaining time would be fully allocated to each sub task, so as to extend the execution window period of each of them. On choosing period, we will take available resources that possess the longest time and their execution time are still in the window period. Given the algorithm description and specific examples, the optimized model would achieve a minimum critical path optimization algorithm to ensure that the task to implement lower cost of destination within the deadline required by the user.Finally, optimization is inspected to verify the actual effect of the algorithm optimization in this paper. The simulation software, Cloud Sim, achieves new scheduling algorithm after the minimum critical path algorithm optimization under the Cloud Workflow, with the purpose of the optimization of the expense within the deadline. Then compare the actual results.