M. Mazilu, A. Valentine, I. Wakeman, G. Parisis, An Experimental Study of Congestion Control in LEO Satellite Networks, in Proc. of IFIP Networking, 2026. Low Earth Orbit (LEO) satellite networks introduce unique congestion control (CC) challenges due to frequent handovers, rapidly changing round-trip times (RTTs), and non-congestive loss. Developing effective transport for these environments requires a clear, comparable understanding of the performance profiles of both state-of-the-art Internet CC and emerging LEO-tailored designs, yet producing such profiles is challenging because LEO dynamics make experiments difficult to control, reproduce, and interpret. To address this, we conduct an emulation-driven evaluation of CC schemes in LEO networks, combining realistic orbital dynamics via the LeoEM framework with targeted Mininet micro-benchmarks that isolate specific conditions. We evaluate representative algorithms from three classes: loss-based Cubic, model-based BBRv3, and learning-based approaches Vivace and Astraea, alongside LEO-tailored SaTCP and LeoCC, across diverse scenarios. Results show that (1) handover-aware schemes can reclaim bandwidth but often at the cost of increased latency; (2) BBRv3 sustains high throughput with modest delay penalties, yet adapts slowly to abrupt RTT shifts; (3) learning-based schemes underperform under dynamic conditions, despite strong resistance to non-congestive loss; and (4) fairness degrades markedly under RTT asymmetry and multiple bottlenecks, particularly for human-designed CC schemes. These findings highlight limitations of today’s CC approaches and help inform the design of LEO-specific schemes.