Efficient Parallel Simulation Algorithms Research For System-Level Of Mega Constellation Communication Networks

Satellite Internet,as an important component of the integrated spaceground information network,is a hot research hotspot nowadays.Many scholars have undertaken research in the areas of constellation networking design,network protocol design,and communication performance assessment,and their main research tool is software simulation.Traditional stand-alone network simulation simulators based on OPNET or NS3 are constrained in the simulation efficiency of mega satellite networks with many nodes and wide coverage because of the limitations of computer hardware conditions and software performance.The problem of long simulation time and high memory consumption seriously affects the simulation needs to be solved.Parallel discrete-event based simulation can provide efficient simulation efficiency,but its simulation efficiency is affected by topology partitioning,and there is an urgent need for reasonable resource evaluation to ensure load balancing in order to reduce simulation time and simulation consumption.Based on this,the research of this paper is shown as follows.1.The simulation of mega satellite constellations will lead to the problems of time-consuming simulation and high memory consumption.Therefore,this paper adopts parallel discrete-event simulation technology to simulate mega satellite constellations based on the loose correlation between different areas of the global satellite network,and proposes a parallel simulation architecture in which satellite constellation simulation,network protocol simulation and parallel simulation are jointly coordinated to achieve the goal of improving simulation efficiency while meeting the communication requirements of mega constellations.2.The network simulation realism,the computational resources required for simulation and the simulation runtime are the influencing factors for the performance of parallel network simulation.Therefore,this paper conducts computational resource modeling and proposes a CPU and memory-based resource prediction algorithm,and obtains a set of accurate CPU clock cycles and memory computational resource models related to the number of users and satellites through a system-level simulation system of satellite protocol stack built by OPNET platform and multivariate fitting prediction algorithm.The simulation verifies that the fitting accuracy of the proposed algorithm is controlled at about 95%,which ensures the accuracy and reliability of the resource evaluation3.Topology partitioning algorithms are studied,in which the merit of topology partitioning in parallel simulation environment determines the resource balance of each sub-topology and has a decisive role in the time efficiency of the overall simulation,with poor performance of traditional uniform partitioning and high computational complexity such as greedy algorithm.Based on the above problems,this paper proposes a topology partitioning algorithm based on the characteristics of Walker constellation configuration,which is optimized in terms of load balancing and computational complexity with the resource evaluation fitting result as the optimization target.The optimal partitioning of parallel network topology and the scheduling of load balancing of massive simulation events computation are realized.The simulation results show that the proposed algorithm has better load balancing performance and lower computational complexity compared with uniform partitioning algorithm and greedy algorithm.4.Finally,this paper builds a parallel simulation platform based on the combination of STK,OPNET and Proxmox VE,which is responsible for satellite constellation simulation system,central control node and satellite constellation protocol stack simulation cluster system,and unites CPU and memory-based resource prediction algorithm and load balancing-based intelligent topology partitioning algorithm to realize a parallel simulation system for mega constellation communication network.The simulation results show that the parallel simulation efficiency of the algorithm proposed in this paper is close to the theoretical value,and the performance is improved by 18.6 times in the scenario of 900 satellites with 1500 terminals.