LFQ: Online Learning of Per-flow Queuing Policies using Deep Reinforcement Learning
This work addresses network congestion management for flows with incompatible congestion control algorithms, offering an incremental improvement over static or traditional AQM methods.
The paper tackles the problem of managing buffer sizes in fair queuing systems by designing LFQ, an Active Queue Management mechanism that uses deep reinforcement learning to dynamically learn optimal queue sizes per flow online, resulting in significantly smaller queues while maintaining or improving throughput compared to existing schedulers.
The increasing number of different, incompatible congestion control algorithms has led to an increased deployment of fair queuing. Fair queuing isolates each network flow and can thus guarantee fairness for each flow even if the flows' congestion controls are not inherently fair. So far, each queue in the fair queuing system either has a fixed, static maximum size or is managed by an Active Queue Management (AQM) algorithm like CoDel. In this paper we design an AQM mechanism (Learning Fair Qdisc (LFQ)) that dynamically learns the optimal buffer size for each flow according to a specified reward function online. We show that our Deep Learning based algorithm can dynamically assign the optimal queue size to each flow depending on its congestion control, delay and bandwidth. Comparing to competing fair AQM schedulers, it provides significantly smaller queues while achieving the same or higher throughput.