Prism improves mobile network speed and cuts delay using RL
PRISM: RL-Based Rate Control for Low Latency and High Data Rate in Mobile Edge Networks
Networking and Internet Architecture
Summary
Many interactive services like virtual reality need internet connections that are both fast and have minimal delay. Traditional methods to control data flow in mobile networks rely on signals that can be slow or unreliable. The authors propose PRISM, a method that uses machine learning to adjust data rates early by watching how network resources are used, keeping delays very low and speeds high. Tests show it performs much better than common methods when multiple users share the network.
What this means in practice
- •For mobile network engineers: Optimize data transmission settings to reduce delays and boost throughput in crowded mobile networks using early congestion signals from resource usage.
- •For cloud gaming platform operators: Manage network congestion in mobile edge computing to maintain smooth, low-latency game streaming during high network contention.
Authors
Hamid Hassani, Staji Wilson, Kishor C. Joshi, Sonia M. Heemstra, Ignas G. Niemegeers, George Exarchakos
Abstract
Emerging delay-sensitive interactive services, such as extended reality and remote operation, demand advanced congestion control algorithms (CCAs) capable of maintaining high data rates while keeping latency low in mobile networks. Existing CCAs typically rely on congestion signals such as packet loss, delay, or explicit congestion notification (ECN) marks, which can be delayed, imprecise, or require extensive upgrades to the network infrastructure. To address these limitations, physical resource block (PRB) utilization has recently been proposed as a reliable early congestion signal in mobile networks [1]. This paper introduces PRISM, a reinforcement-learning-based CCA that uses peak PRB utilization to regulate the send rate under challenging mobility and fading conditions. NS3 simulation results show that, in a mobility scenario where two PRISM flows coexist with two TCP Cubic flows and one TCP BBR flow, PRISM maintains a per-packet mean one-way delay (OWD) of 6.3ms and a $p_{99}$ OWD of 44.6ms, with negligible (0.02%) packet loss, over a 20min high-contention interval. However, over the same interval, the coexisting TCP flows experience 60x higher mean OWDs of 371-390ms and $p_{99}$ OWDs of 563-578ms, with 0.1% packet loss. PRISM achieves a 31x improvement in goodput-over-latency performance, indicating a favorable tradeoff for delay-sensitive applications.