Performance improvement of RED: Random early detection using input sensitivity with threshold modification

Active Queue Management (AQM) is congestion avoidance approach which provides an early notification of network congestion by actively dropping and marking packets which enables end hosts to adapt their transmission rates to network traffic conditions. For anticipitation of congestion notification Random Early Detection (RED) , well known AQM, uses average queue length as a congestion measure. In our work apart from using average queue length we have used combination of input rate and slight modification of threshold value and observed the impact of it on the performance of RED. Performance is observed in case of packet loss ratio and stability of average queue length. Mismatch behavior of average queue length and current queue length is improved. We have used ns-2 to implement our work.