Research On Edge Computing Service Migration Strategy Of LEO Constellation Broadband Network

Low orbit constellation broadband network has the characteristics of large spatial and temporal scale,time-varying topology and strong heterogeneity of resources,which are significantly different from the traditional ground network.The application of edge computing technology to low orbit constellation broadband network can significantly reduce the task processing delay and bandwidth consumption,and it has been widely used and deployed in the traditional ground network.In addition,the high dynamic changes of broadband network topology and user traffic in low-orbit constellations will increase network latency and packet loss rate,leading to a serious decrease in network quality of service(Qo S)and user quality of experience(Qo E).In recent years,in order to reduce the negative impact caused by high network dynamics,the industry began to pay attention to the introduction of service migration technology in ground network into the edge computing architecture of low orbit broadband network,so as to further optimize and improve the performance of the edge computing architecture of low orbit broadband network.In this thesis,an edge computing architecture is designed for a new generation of low orbit constellation broadband network.In this thesis,geosynchronous orbit satellite cluster is used as the control layer of edge computing architecture,and low orbit satellite cluster is used as the service implementation layer in the architecture,because of the variety and heterogeneity of edge nodes in the new generation low orbit broadband network.Due to the limited computing,storage and network resources on the LEO satellites,lightweight virtualization technology and virtual resource technology are used in this thesis to improve the resource utilization rate of the LEO satellites.Further,the broadband network of the low orbit constellation is dynamic,and the inter-satellite link will be temporarily interrupted,which will lead to the offline of the low orbit satellite.Therefore,when the communication between the LEO satellite and the control center is interrupted,the LEO satellite is offline,and the data information sent by the control center cannot be synchronized in time,resulting in the failure of the LEO satellite when performing the task.Therefore,this thesis designs the LEO satellite edge control module,which makes the satellite nodes work independently and ensures the continuity of services required by users.Based on the above-mentioned LEO constellation broadband network edge computing architecture,this thesis comprehensively considers the dependency between the high dynamism of low orbit constellation broadband network and the performance requirements of business applications,and puts forward an innovative service migration strategy based on fine-grained service division and joint optimization.The service migration decision can be made according to the change of upstream and downstream traffic between LEO satellite nodes.The strategy first uses the maximum flow algorithm to find all the nodes on the optimal bandwidth path and takes them as the pre-selected migration node set.Then,the strategy considers the benefits of service migration delay and transmission energy consumption,and further reduces the matching range of migration nodes,and finally selects the best service migration node by combining the balance of resource utilization between nodes and services.Based on the open-source network simulation architecture NS-3,this thesis developed a simulation system for the edge computing architecture of broadband network in low-orbit constellation,and tested and evaluated the performance of the service migration strategy proposed in this thesis.The simulation results show that the strategy can effectively reduce the task end-to-end delay by 11% and the migration energy consumption by up to 58%.It can significantly improve the throughput and packet loss rate of LEO constellation broadband network in congestion scenarios,optimize its bandwidth usage,and effectively improve the network service quality and user experience quality.