LRED: A robust active queue management scheme based on packet loss ratio

Abstract

Active queue management (AQM) is an effective method to enhance congestion control, and to achieve tradeoff between link utilization and delay. The de facto standard, random early detection (RED), and most of its variants use queue length as a congestion indicator to trigger packet dropping. The proportional-integral (PI), use both queue length and traffic input rate as congestion indicators; effective stability model and practical design rules built on the TCP control model and abstracted AQM model reveal that such schemes enhance the stability of a system. In this paper, we propose an AQM scheme with fast response time, yet good robustness. The scheme, called loss ratio based RED (LRED), measures the latest packet loss ratio, and uses it as a complement to queue length in order to dynamically adjust packet drop probability. Employing the closed-form relationship between packet loss ratio and the number of TCP flows, this scheme is responsive even if the number of TCP flows varies significantly. We also provide the design rules for this scheme based on the well-known TCP control model. This scheme's performance is examined under various network configurations, and compared to existing AQM schemes, including PI, random exponentially marking (REM), and adaptive virtual queue (AVQ). Our simulation results show that, with comparable complexity', this scheme has short response time, better robustness, and more desirable tradeoff than PI, REM, and AQV, especially under highly dynamic network and heavy traffic load.