Research And Implementation Of The Distributed Task Scheduling Algorithm In Cloud Environment

Cloud computing is a new and significant change in IT,driving the development of new industries and value chains.Cloud computing platform unified manage of computing,storage and network resources,and provide them to users in the manner of network services.It achieves the sharing of resources and improves the utilization of resources.However,due to the large scale of cloud resources,heterogeneous distribution and dynamic changes and other characteristics,how to reasonable and effective schedule task,and achieve a reasonable allocation of resources is facing various difficulties.In view of the problem of cross-regional heterogeneity and efficiency of resource nodes in workflow task scheduling in cloud computing environment,and the scheduling problem of data processing task,this thesis makes the following research:(1)Proposed a workflow scheduling model based on artificial immune algorithm in cloud environment.The regional distribution of resources in the cloud environment has a great impact on the communication between the task nodes in the workflow task scheduling process.In this thesis,the completion time of the task nodes and the transit delay of the inter-nodes are defined as the constraint function,Using the advantages of the artificial immune scheduling algorithm,the completion time and cross-node delay as constraints to generate the fitness function of resources and tasks,and use gene recombination method to improve the antibody mutation process of the artificial immune algorithm.A workflow scheduling model based on artificial immune algorithm is proposed to obtain the scheduling scheme between workflow task nodes and resource nodes.And the experiment proves that the algorithm can improve the task scheduling efficiency under the condition of resource region distribution.(2)Proposed a task scheduling strategy based on data distribution in cloud environment.During the scheduling process of data processing task,it is necessary to use a reasonable scheduling strategy to fragment the data and carry out the resource acquisition,and allocate the data fragments to the requisitioned resource nodes.This thesis presents a task scheduling strategy based on data fragmentation.The strategy can be used to establish the optimal scheduling mathematical model of the data fragmentation according to the the data granularity and the performance difference of the available resource nodes.The ideal segmentation ratios of the data fragmentations can be computed using the data partitioning model,then,actual task sheduling method can be made using the actual cuttable granularity and the ideal segmentation ratios of the data fragmentations.The experimental results show that the strategy can effectively reduce the completion time of data processing task.(3)Based on the infrastructure management platform,the data service cloud platform is designed and developed.The scheduling management module of the platform is implemented using the task scheduling algorithms proposed in this thesis,and made a detailed design description of the scheduling management module,which consisits of calculation center,expropriation center,data center and requisition machine service.Through the platform test,it presents the effects of the scheduling algorithms proposed in this thesis.