Managing network congestion with a Kohonen-based RED queue

TL;DR

This paper introduces Kohonen-RED, a neural network-based Active Queue Management method that stabilizes queue lengths and reduces delay jitter without complex parameter tuning, addressing limitations of existing RED enhancements.

Contribution

It proposes a novel Kohonen neural network approach to improve RED's stability and parameter robustness in network congestion management.

Findings

Kohonen-RED stabilizes queue lengths effectively.

It reduces delay jitter in network traffic.

The method requires no complex parameter tuning.

Abstract

The behaviour of the TCP AIMD algorithm is known to cause queue length oscillations when congestion occurs at a router output link. Indeed, due to these queueing variations, end-to-end applications experience large delay jitter. Many studies have proposed efficient Active Queue Management (AQM) mechanisms in order to reduce queue oscillations and stabilize the queue length. These AQM are mostly improvements of the Random Early Detection (RED) model. Unfortunately, these enhancements do not react in a similar manner for various network conditions and are strongly sensitive to their initial setting parameters. Although this paper proposes a solution to overcome the difficulties of setting these parameters by using a Kohonen neural network model, another goal of this study is to investigate whether cognitive intelligence could be placed in the core network to solve such stability problem. In our context, we use results from the neural network area to demonstrate that our proposal, named Kohonen-RED (KRED), enables a stable queue length without complex parameters setting and passive measurements.