Research On Routing Algorithm Of Double-layer Satellite Networks Based On Global Topology

With the rapid development of intelligent optimization algorithms and software-defined technologies based on global topology,the feature of large-area coverage of satellite networks and the strong stability and invulnerability have greatly compensated for the inadequacies of terrestrial communications networks.SDN is used to separate the control layer from the data forwarding layer.Since it has the characteristics of global optimization,centralized control,and flexible configuration,and the intelligent optimization algorithms use the global topology information to find the optimal solution and combine them organically.We focus on the research on the routing algorithm of global topology based double-layer satellite network..Firstly,the background and significance of the research of this paper are analyzed,and the research progresses at home and abroad for satellite communication network,satellite network routing algorithm and SDN are combed.In order to meet good coverage performance and link performance,GEO/LEO double-layer satellite network with double-backbone is designed,and SDN switches and SDN controllers are added,which greatly improves network resource utilization through centralized control.For different scenarios,the classical routing algorithms like Dijkstra,Floyd,Prim,Kruskal,and random walk search strategy are compared and analyzed.Secondly,by the reason of the heuristic intelligent optimization algorithm can use the global topology information to find the optimal solution,the genetic algorithm and ant algorithm are used to perform the matlab simulation analysis for the classical shortest path problem.Afterwards,for the multicast routing problem with multi-objective Qo S constraints,combining the data of inter-satellite distance,bandwidth,and delay that provided by the satellite network topology.Also the bandwidth constraints and the time delay constraints are set,and simulations are conducted based on the ant colony algorithm.The number of ants in each iteration of the algorithm,the pheromone importance factor,the path length importance factor,the pheromone evaporation coefficient,and the pheromone increase intensity coefficient are explored to fit the algorithm convergence speed and avoid falling into local optimum.The two factors are superior,and the ant colony algorithm is improved by modifying the pheromone update rules.A better path finding strategy is achieved.The convergence time of the algorithm is lower,and global optimization is performed to avoid the algorithm falling into a local optimum.Finally,the network topology was built and simulated under the Mininet of Ubuntu system,in order to solve the problem of queue blocking and link separation path,a new routing algorithm based on network traffic is designed: k maximum and minimum link separation path algorithm.The algorithm is roughly divided into two parts.The first part finds a set of paths from the source node to the destination node through the improved Dijkstra algorithm,and the second part uses the greedy algorithm to choose the link separation path with the smallest bottleneck bandwidth.Compared with the SPF shortest path algorithm and the First-k-max algorithm,the proposed algorithm has better throughput than the other two algorithms in the minimum and average bottleneck bandwidth.