Resource Information Model And Traffic Detection For Satellite Network Traffic Control

As an important part of the integrated space-ground network,the satellite network is in the process of rapid expansion and development.More and more satellite networks will be equipped with inter-satellite links,and provide services by space networks.Satellite networks will carry more and more traffic,which may cause problems such as congestion or even collapse.Therefore,traffic management and control are needed to ensure the good communicating and operating state of satellite networks.With the development of technology,the on-board hardware resources are getting more and more plentiful,which makes it possible for satellites to work and control autonomously.Characteristics of high dynamics,high latency,and limited on-board resources of satellite networks are considered.In order to meet the demand for autonomous traffic management and control of the Beidou satellite network,a resource information model for describing satellite network resources,a method of satellite network autonomous traffic detection,and a method of congestion control are studied to realize autonomous traffic management and control of space networks.The main research contents of this article are as follows:Network resource information provides information support for traffic management.In order to describe the satellite network resources related to traffic management and control,resources in the satellite network are analyzed,and the abstracting method of resources and the expression method of time information are studied to design a time-varying graph-based resource information model of satellite networks.The graph database Neo4 j is used to store and manage the satellite network resource information model.Experiments on common query scenarios in resource management are conducted and the performances of Neo4 j and My SQL are compared.The experiment results show that Neo4 j can query resource information much faster than My SQL.The experiment results also show that,through the resource information model and the Neo4 j graph database studied in this paper,the resource information of the satellite network can be queried quickly,such as the current network topology information and link historical traffic information.Accurate detection of traffic information is the basis for traffic management and control.Considering the scenarios of intermittently-connected and length-changing links in satellite networks,a dynamic traffic detection method for satellite networks is studied.A hybrid traffic detection framework for the Beidou satellite network which combines both centralized and distributed framework is designed.It is demonstrated that the hybrid framework can improve the timeliness of traffic detection.Under this hybrid detection framework,dynamic traffic detection based on LSTM traffic prediction is studied,and the detected traffic information is stored in the satellite network resource information model.Experiments are conducted to compare the on-board traffic prediction performances of LSTM and ARMA.The experiment results show that LSTM has better prediction performance and tolerable computing and storage consumption.The experiment results of traffic detection show that the detected traffic sequence can present the real traffic well.In order to solve the congestion problems in satellite networks,a congestion control routing algorithm based on traffic prediction is studied.A satellite network data communication simulation platform is built to simulate the data communication process in satellite networks.The experiment results of congestion control conducted on the simulation platform show that the congestion control routing based on traffic prediction can effectively reduce the network packet loss rate and reduce network congestion.The research work in this paper promotes the development of satellite network autonomous control technology,and provides new ideas and methods for the autonomous traffic management and control of satellite networks.It has important reference significance for the automated and autonomous management of satellite networks,and has a good application prospect in the development trend of autonomous satellite network.