Research On Emulation Technologies For Edge Computing Architecture Of LEO Satellite Constellation Broadband Network

In recent years,LEO satellite constellation broadband networks have gradually become a research hotspot in the industry due to relatively low total latency and wide network coverage.At the same time,as the characteristics of fast changing network topology,high satellite link propagation delay and large number of users in LEO satellite constellation networks,by designing and building edge computing architecture on LEO satellite constellation networks can reduce the service delay of high real-time application tasks,save satellite network bandwidth and protect user data privacy.However,since satellite networks are expensive,it is impossible to use real satellite networks for experimentation and testing in most cases.Therefore,it is of great theoretical and practical importance to study emulation techniques for the edge computing architecture of LEO satellite constellation broadband networks.In order to meet the emulation requirements of edge computing architecture of LEO satellite constellation broadband networks,this thesis studies emulation technologies based on the resource virtualization platform,focuses on the model construction of edge computing architecture of LEO satellite constellation broadband networks,designs and implements the emulation system for edge computing architecture of LEO satellite constellation broadband networks.The specific research includes the three aspects: firstly,the thesis models the networks and users at the bottom layer of the LEO satellite constellation,as well as the edge computing tasks at its upper layer.Among them,the network model mainly includes node model,link model and constellation model;the user model mainly includes distribution model and behavior model;the task model mainly includes transmission model and duration model.Secondly,this thesis designs and implements an emulation system for edge computing architecture of LEO satellite constellation broadband networks based on virtual resources.The network emulation layer and the edge computing layer in the system are responsible for the emulation of the underlying LEO satellite constellation networks,users and the upper edge computing architecture respectively.The functional test results show that the emulation system designed in the thesis has the capability of emulating network management,network characteristics and network services.The performance evaluation results show that the emulation system designed in this thesis has a 4.4% lower CPU utilisation,12% lower memory consumption and 44.3% lower eimulation time consumption than the traditional LEO satellite constellation network emulation system for the same network size.Finally,a resource scheduling strategy based on node connectivity is proposed to address the problems of resource fragmentation and emulation network performance in the emulation system.The strategy places network nodes with long-time connectivity in the same emulation computing node to improve emulation network performance,and reduces the number of emulation computing nodes used to reduce resource fragmentation.The experimental results show that,compared with the traditional Open Stack scheduling strategy and the first adaptation strategy,the resource scheduling strategy proposed in this thesis improves system CPU utilisation by 9% and 31%,memory utilisation by 7% and13%,and bandwidth utilisation by 2.3% and 2.6%.In summary,the emulation technologies for edge computing architecture of LEO satellite constellation broadband networks studied in this thesis can meet the requirements of high fidelity and high flexibility emulation of the edge computing architecture of LEO satellite constellation broadband networks,and this thesis has the value of theoretical research and practical application.