Research On Flexible Task Scheduling Method Of Edge Computing And Cloud Collaboration

With the rapid development and popularization of 5G and the Internet of Things,the data generated by terminals at the edge of the network is an urgent problem to be solved at present.The emergence of edge computing,which can process data near the edge of the network where data is generated and reduce data transmission delay.However,it only has the local advantage of the edge layer,and it has shortcomings in the global task scheduling and dynamic resource configuration,which may lead to the problems of low resource utilization rate,long task processing delay and unbalanced system load,so as to lead to affect the service quality of users.Therefore,this thesis studies terminal task scheduling based on edge computing and cloud collaborative environment,mainly including the following contents:(1)Research on coarse-grained task offloading method of edge computing and cloud collaboration.Task offloading is one of the important contents of edge computing,and it is also a hot topic of current research.Most of the existing studies do not mention the collaboration of cloud centers or the fact that cloud centers do not participate in scheduling.In order to solve the shortcoming of global scheduling in edge computing,this thesis introduces the cloud center into the binary architecture of edge computing,and proposes a “Cloud-Edge-Terminal”three-layer cooperative offloading model in which the cloud center participates in scheduling.The model dispatches tasks from the cloud center and returns the generated offloading strategy to the terminal and edge layer devices.On the basis of this model a coarse-grained task offload method have been proposed,this method is fully aware of the task priority,according to each server resource usage,load using artificial bee colony algorithm to optimize task offload strategy,implement the task elastic offloading,compared with random offloading methods,effectively reduce the task processing time delay.(2)Research on fine-grained task partitioning and scheduling methods of edge computing and cloud collaboration.At present,most of the task scheduling takes coarse-grained tasks as the scheduling object,and the tasks generated by the terminal are regarded as a whole,which is not allowed to be divided.The tasks are processed completely at the terminal or offloaded to the server at the edge layer.If the tasks are divided into finer granularity during task scheduling,the processing time can be shortened by increasing the degree of parallelism.This thesis explores the elastic scheduling method of fine-grained tasks.Based on the three-layer collaboration model of “Cloud-Edge-Terminal”,a fine-grained task elastic scheduling method is proposed to divide the tasks into more fine-grained tasks.In this method,a heuristic algorithm based on the fusion of genetic algorithm and simulated annealing algorithm is used to improve the task scheduling in edge computing and cloud collaboration,and the total task processing time is reduced compared with the original method.By designing and implementing a series of simulation experiments,the correctness and effectiveness of the improved fine-grained task scheduling method are verified.The above methods have been used in the regional epidemic prevention and control grid cloud platform for trial.Under the “Cloud-Edge-Terminal” collaborative framework,an "intelligent gatekeeper" system is being designed and developed to quickly and accurately screen the health status of the entrant.