Effect of Load and Threshold Variation on Performance of RED: Random Early Detection

To support end to end control mechanisms in Internet Active Queue Management (AQM) is proposed. Random Early Detection (RED) is one such prominent congestion avoidance mechanism in Internet architecture. Since it is proposed in 1993 several modification and enhancements have been made so as to make it more responsive for congestion avoidance. RED parameter setting has a great impact on its performance In this paper we are investigating the effect of varying threshold parameter and additionally updating average queue size value depending on current queue value. We are analyzing the impact of these changes using network simulator ns-2 and got improved results in terms of packet loss, throughput, better link utilization. Using extensive simulation our strategy evolution demonstrates superiority over original RED. .