The impact of random early detection on the performance of different queuing disciplines

Queuing disciplines are part and parcel of packet-switched network study and research. Network performance is deeply affected by the queuing discipline being used. In this paper, we study the impact of using random early detection (RED) policy on the performance of several queuing disciplines as compared to drop-tail policy in a mixed-traffic packet-switched network. We consider four widely used and implemented queuing disciplines: First-Come-First-Serve (FCFS), Priority Queuing (PQ), Weighted-Fair Queuing (WFQ) and Deficit-Weighted Round Robin (DWRR) and evaluate their performance using OPNET network simulator. Under each queuing discipline, many important network performance metrics such as queuing delay, packet drop rate, end-to-end delay and delay jitter are reported and compared for both droptail policy and RED policy. The simulations results verify the major role of RED in improving the measured performance metrics of all considered queuing disciplines.