Research On On-board Collaborative Computing Strategy Based On Edge Computing

With the development of ground mobile communication technology,a series of new information services with extremely low latency requirements have sprung up.Especially,the services such as smart distant fishing and intelligent cleaning of marine debris also contain special geographical location requirements.Therefore,the new information services meet some new challenges to the ground mobile communication network in terms of latency and coverage.According to these problems,the mainstream solutions is an integrated network framework.This framework utilizes the cloud central computing platform as the main execution platform of information services and uses the widely deployed edge computing platform to handle parts of businesses of high compression ratio and latency-sensitive,which satisfy the need for the new information services in terms of latency,capacity and reliability in some degree.However,restricted by geographical conditions,there is a huge gap between the limited coverage prospect of ground mobile communication network and the ubiquitous information service demand in the future.Thus,a new network framework designed for the ever-expanding coverage requirements of new information services is urgently needed.To make up for the shortcomings of coverage in the integrated network framework,the low earth orbit communications satellite network has become a hot and reliable solution depending on the wide coverage,not restricted by geographical environment,the low cost and so on.However,due to the frequent satellite-ground long-distance transmission,the current transparent transmission mode of satellite communication technology cannot meet the low latency demands of the continuous development of information services.Considering the low latency requirements of information services,based on the edge computing framework,the paper proposes an on-board collaborative computing strategy of the information services,using the computation and communication resources of low-orbit communication satellite network in order to meet both the coverage and latency requirements of information services.In addition,based on the theory of on-board collaborative computing strategy,a system simulation platform for information service is designed,hoping to explore ways to reduce the processing latency of information service from both theoretical deduction and simulation experiment.Specifically,the master content of this paper is as follows:This paper describes the domestic and foreign research status from three aspects of the edge computing framework,low earth orbit communication satellite network and the container virtualization technology,and analyzes the main research contents and innovation points of this work,so as to guide the subsequent fusion work of the path-type edge computing scheme and the low earth orbit communications satellite network.Based on the path-type edge computing scheme,a path-type on-board collaborative computing strategy is proposed for the low earth orbit communications satellite network.Under the guidance of the service graph model,the service is scheduled to the satellite network in the transmission path by specifying the mapping of the service directed acyclic graph to the time extended graph of the low earth orbit communication satellite network,so as to avoid the problem of high latency caused by the satellite-ground long distance transmission.In order to realize the collaborative computing assumption in the high dynamic low earth orbit communication satellite network,this paper add the link latency weight parameter to the time extended graph model,which is used for the information service processing latency analysis.The simulation show that when the data is 5MB,the latency performance of onboard collaborative computing is improved by 58.9% compared with that of ground-based cloud computing.On the basis of the theoretical research on the path-type on-board collaborative computing strategy,a system simulation platform for the low earth orbit communication satellite network was built,the service scheduling and execution functions were designed,and the collaborative computing process of low earth orbit communication satellite network was simulated.Specifically,first of all,low earth orbit communication satellite network is simulated based on docker technology.Then,using Spring Boot micro-service development framework,Our realize the pre-planning function of information services,the processing functions of information services and function library for image processing services.Simulation show that when the image linked list data is 1050 MB,the experimental performance of on-board collaborative computing strategy is improved by 47.0% compared with the ground-based cloud center computing strategy.