Learning to Sail Dynamic Networks: The MARLIN Reinforcement Learning Framework for Congestion Control in Tactical Environments

TL;DR

This paper introduces MARLIN, a reinforcement learning framework designed for congestion control in tactical networks, demonstrating superior performance over traditional methods like TCP Cubic and Mockets in complex, unstable environments.

Contribution

The paper presents a novel RL-based congestion control framework tailored for tactical networks, with an accurate emulation environment and refined evaluation methods for better generalization.

Findings

MARLIN outperforms TCP Cubic and Mockets in file transfer tasks

The RL framework effectively adapts to fluctuating tactical network conditions

Results show improved throughput and reliability in simulated tactical scenarios

Abstract

Conventional Congestion Control (CC) algorithms,such as TCP Cubic, struggle in tactical environments as they misinterpret packet loss and fluctuating network performance as congestion symptoms. Recent efforts, including our own MARLIN, have explored the use of Reinforcement Learning (RL) for CC, but they often fall short of generalization, particularly in competitive, unstable, and unforeseen scenarios. To address these challenges, this paper proposes an RL framework that leverages an accurate and parallelizable emulation environment to reenact the conditions of a tactical network. We also introduce refined RL formulation and performance evaluation methods tailored for agents operating in such intricate scenarios. We evaluate our RL learning framework by training a MARLIN agent in conditions replicating a bottleneck link transition between a Satellite Communication (SATCOM) and an UHF Wide Band (UHF) radio link. Finally, we compared its performance in file transfer tasks against Transmission Control Protocol (TCP) Cubic and the default strategy implemented in the Mockets tactical communication middleware. The results demonstrate that the MARLIN RL agent outperforms both TCP and Mockets under different perspectives and highlight the effectiveness of specialized RL solutions in optimizing CC for tactical network environments.