Design And Implementation Of Cloud Workflow Task Scheduling System

With the rapid development of cloud computing technology,cloud task scheduling problem has become a hot research topic in related fields.Most of the traditional task scheduling algorithms are based on the minimal completion time,which ignores the problem of quality of service.Cloud computing is different,it is a commercial application,and one of the most important business applications is customer satisfaction.Service quality has a direct impact on the customer's experience of cloud computing platform,cloud computing service quality is the key to the sustainable development of cloud computing.So for cloud computing task scheduling problem,not only to the optimal span or the minimum completion time as the ultimate goal,but also should focus on how to improve the service quality of the system.Aiming at the resource allocation problem of task scheduling in cloud computing environment,this paper makes a deep research on the cloud workflow scheduling problem.The research work in this paper is as follows:(1)The key technology of the scientific workflow,combined with cloud computing background,carries on the demand analysis to the cloud workflow task scheduling system,according to the demand of design this dispatch system,divides the system into cloud entrance,instance model definition module,management module,log module,cloud workflow engine and Hadoop six function module,and for each module of detailed design,points out the design process and functions of every module,lay the foundation for the algorithm research and system implementation.(2)Analysising the cloud workflow task scheduling model,mapping the cloud workflow task will be attributed to NP complete problem,establishing the workflow directed acyclic graph(DAG)model,specified model of various constraints and optimization target.Also for cloud computing quality of service(QoS)special requirements,improving and optimizing it from the cost of their time and resources,proposing REMin-Min cloud task scheduling algorithm based on Min-Min algorithm,and gives the programming REMin-Min algorithm implementation in Cloudsim.Under the same experimental conditions,the algorithm is compared with the existing similar algorithm OPT-Min-Min algorithm and Min-Max algorithm.By comparing the simulation results of the above three algorithms,it is Intuitive to obtain the conclusion that the proposed method is superior to similar algorithms on the quality of service.(3)Achieving an oozie workflow system based on hadoop platform,and the algorithm is implemented as the core scheduler for the workflow task scheduling in the system.Through on the system platform specific workflow instance running and monitoring under the actual operation of the system with good performance,effectively prove the proposed algorithm REMin-Min not only not only has a good effect in simulation,but also can be used in the actual system.