SpaceNet: A Resource-Efficient Simulation Testbed for LEO Mega-Constellation Networks

The rapid expansion of low Earth orbit (LEO) mega-constellations, such as SpaceX’s Starlink and Amazon’s Kuiper, aim to transform Internet access. These systems face substantial challenges due to the dynamic nature of LEO satellites, which result in frequent updates to network topologies. To support research into these challenges, we introduce SpaceNet, a resource-efficient testbed designed to emulate LEO mega-constellation networks and enable future experimental studies. Leveraging the Mininet framework, SpaceNet can efficiently simulate thousands of satellites, creating realistic network scenarios for testing and analysis. Our two-phase design enables fast, efficient simulation and emulated real-time, reducing hardware resource requirements while maintaining fidelity. We evaluate SpaceNet’s performance under an isolated ISL-only topology scenario, highlighting differences compared to real-world Starlink data. Additionally, we showcase SpaceNet’s resource profile by examining the impact of constellation size and routing interval on CPU utilization, memory usage, and real-time active network execution. SpaceNet enhances research into non-terrestrial networks, offering a scalable, open-source solution to address many research areas yet to be defined.