Research On The Deployment Of Multi-SDN Controllers Of Partitioned Hierarchical In Space-Air-Ground Integrated Network

The Space-Air-Ground Integrated Network(SAGIN)is a remarkable network framework for realizing the ”6G Smart connect everything”.With the emergence of the 6G space-ground integrated network,network management faces many problems.For example,the traditional satellite network management mode solidifies the network functions for a superintensive network,and the network heterogeneity makes network management more difficult.At the same time,Software Defined Network(SDN)technology has successfully applied to the ground Network,which has the characteristics of centralized management,control/forwarding separation,and Network programmability.To solve the above problems,many researchers applied SDN technology to SAGIN.In the SDN-enabled SAGIN,the controller is the core device of the network.Its deployment strategy will directly affect the performance of the entire network.Different from the controller deployment on the ground,the controller deployment on the satellite faces the following problems: the dynamic change of satellite network topology? the largescale network nodes? the number of aircraft in the sky changes,and the distribution of users on the ground is uneven.These factors make controller deployment extremely challenging.Thus,it is of great significance to study how to optimize the deployment of multiple controllers in the SDN-enabled SAGIN.Given the above problems,this thesis classified the SDN-enabled SAGIN model into the based single-constellation partitioned hierarchical SDN-enabled SAGIN model and the based hybrid-constellation partitioned hierarchical SDN-enabled SAGIN model.Then,for these two scenarios,a single-constellation-based partitioned hierarchical multi-controller deployment strategy and a hybrid-constellation-based partitioned hierarchical multi-controller deployment strategy are proposed.The specific work is as follows:(1)For the based single-constellation SDN partitioned and hierarchical multi-controller deployment strategy,the hierarchical controller deployment structure is adopted in this thesis,and the first-level and second-level controllers are deployed on the ground network.Firstly,the delay model of the network,the load model of the SDN controller,and a loss value as a measure of whether the network delay and controller load are optimal are defined.Then,using the distribution relationship between the SDN controller and the switch node as the solution space,and taking the loss value as the optimization goal,a multi-controller deployment strategy based on the simulated annealing algorithm(SAA)is used to search for the optimal solution space.Lastly,considering the network topology changes dynamically and the SDN controller imbalance,a switch migration strategy oriented toward load balancing is proposed.The paper aimed to determine the controller deployment plan through the above two points,balance the controller load,and then improve the network performance.The experiment uses Satellite Tool Kit(STK)to simulate the Satellite scene and collect data at fixed time intervals.The algorithm strategy is implemented through Python,and the validity of the algorithm is verified from the aspects of network propagation delay and controller load through data simulation.(2)For the based hybrid-constellation SDN partitioned and hierarchical multi-controller deployment strategy,this thesis adopted a hierarchical controller deployment structure,deploying the first-level controller on the ground network and the second-level controller on the satellite.Firstly,this thesis constructed a hierarchical propagation delay model and a controller load model for the hybrid-constellation-based partitioned and hierarchical SAGIN.In addition,the network mathematical model defined a loss function model by combining the delay model and the load model.Then,this thesis proposed a multi-controller deployment strategy based on k-means and genetic algorithm(GA)to optimize the propagation delay and controller load of the network.Finally,STK is used to simulate the operation scenario of the satellite constellation,MATLAB is used to collect experimental data at fixed time intervals,and Python is used to implement the algorithm strategy.The experiment verifies the algorithm feasibility by analyzing propagation delay and controller load performance and gives a reasonable multi-controller deployment strategy for different delay and controller load requirements.