A Dual-Queue Active Queue Management Scheme to Enhance Overall Performance of Data Flows

Shorter latency improves performance of delay-sensitive Internet applications. However, for less stable networks such as WLAN, reducing target queuing delay or queue length in Active Queue Management (AQM) algorithms easily creates empty queues and lower throughput for bandwidth-intensive flows. Fortunately, bandwidth-intensive data flows are usually insensitive to latency, while delay-sensitive ones transmits less data. Therefore, the paper proposes a fairness-oriented AQM scheme, which administers different data flows using different queue management strategies. The scheme assigns data flows to queues managed by different AQM algorithms based on their respective bandwidth usage. Simulation results show that in unstable network environment, the proposed scheme decreases queuing delay for delay-sensitive data flows and shortens Average Flow Completion Time, when maintains high traffic throughput at the same time.