A Refined AQM Approach for Providing an Improved Performance in Communication Networks

Abstract

Effective queue management for making certain of obtaining an improved network quality of service (QoS) is becoming a crucial challenge to the Internet owing to enormous increase in real-time communications. The widely reputed Random Early Detection (RED) and its offshoot amendment (Three-section RED, TRED) Active Queue Management (AQM) algorithms are considered inefficient for handling high congested network settings, leading to large end-to-end delay. In this paper, we propose a more sophisticated RED-based AQM design that leveraged TRED. We call this new scheme Ref-RED (an abbreviation of ‘Refined Random Early Detection’). Experimental investigations in ns-3 simulator demonstrated that Ref-RED's packet dropping strategy is more belligerent as it actively drops packets at heavy-load congested network, leading to the appreciable performance improvement.