Simulation Analysis of Active Queue Management for Internet Congestion Control

In order to prevent congestion, the current internet uses endto-end congestion control protocol like TCP. In congestion control issues, queue management employed by router has been utmost important. Active queue management (AQM) has been proposed as a router-based mechanism for early detection of congestion inside the network. AQM scheme helps for end-to-end congestion control by having routers detecting congestion and notify end-systems, so that the sender adjust transmission rate earlier and avoid unwanted packet drops. Aim of the paper was to analyze the performance of various active queue management (AQM) techniques like Random Exponential Marking (REM), Gentle RED (GRED) and Nonlinear RED (NLRED) with DropTail (DT). We study different characteristics of various versions of RED using NS2 simulator and outcome indicates that performance of NLRED is better than others in terms of goodput, packet loss rate, delay, link utilization, fairness index, average queue length.