Evaluating Learning Congestion control Schemes for LEO Constellations

2025-10-29

AI Summary

This work addresses the congestion control challenges in low-Earth-orbit (LEO) satellite networks—namely, frequent handovers, time-varying round-trip times (RTTs), and non-congestion-related packet loss—by proposing LeoEM-Mininet, the first integrated evaluation framework combining high-fidelity orbital dynamics simulation (LeoEM) with lightweight network emulation (Mininet). It systematically benchmarks three major congestion control (CC) paradigms: learning-based (e.g., Vivace, Sage, Astraea), model-based (BBRv3), and loss-based algorithms, while incorporating active queue management (AQM) mechanisms. Results show that BBRv3 achieves the best throughput–latency trade-off; learning-based approaches exhibit robustness to non-congestive loss but suffer from limited dynamic adaptability; and AQM significantly improves fairness and link utilization. The study exposes fundamental limitations of existing CC protocols in LEO environments and establishes a reproducible, empirically grounded benchmark for designing space-network-aware congestion control.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Search and Optimization: Sampling/Simulation-based Search
• Multiagent Systems: Agent-Based Simulation and Emergent Behavior

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization
• Security and Privacy: Large-scale security measurements
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

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. This paper presents the first comprehensive, emulation-driven evaluation of CC schemes in LEO networks, combining realistic orbital dynamics via the LeoEM framework with targeted Mininet micro-benchmarks. We evaluated representative CC algorithms from three classes, loss-based (Cubic, SaTCP), model-based (BBRv3), and learning-based (Vivace, Sage, Astraea), across diverse single-flow and multi-flow scenarios, including interactions with active queue management (AQM). Our findings reveal that: (1) handover-aware loss-based schemes can reclaim bandwidth but at the cost of increased latency; (2) BBRv3 sustains high throughput with modest delay penalties, yet reacts slowly to abrupt RTT changes; (3) RL-based schemes severely underperform under dynamic conditions, despite being notably resistant to non-congestive loss; (4) fairness degrades significantly with RTT asymmetry and multiple bottlenecks, especially in human-designed CC schemes; and (5) AQM at bottlenecks can restore fairness and boost efficiency. These results expose critical limitations in current CC schemes and provide insight for designing LEO-specific data transport protocols.

Problem

• Evaluating congestion control schemes for LEO satellite networks
• Assessing performance under handovers and dynamic RTT conditions
• Analyzing fairness and efficiency with multiple bottlenecks and AQM

Innovation

• Emulation-driven evaluation combining orbital dynamics with micro-benchmarks
• Assessed loss-based, model-based, and learning-based congestion control schemes
• Identified limitations and insights for LEO-specific protocol design