Managing Queue Stability Using Art2 in Active Queue Management for Congestion Control.

ABSTRACT Congestion occurs on a link when the traffic is exceeding the maximum capacity of that link which is a major problem in today's world of internet. If the incoming link is larger than that of the outgoing link, it leads to congestion. Active Queue Management (AQM) schemes allow router to proactively respond to congestion by increasing the average length of its queue. Random Early Detection (RED) AQM algorithm a solution for congestion has difficulties in setting the parameters for bursty traffic. Hence an intelligent AQM technique is needed to reduce the packet loss by stabilizing the queue. The proposed Adaptive Resonance Theory 2 (ART 2) RED based AQM an unsupervised intelligent neural network stabilizes the queue and reduces the queue oscillations. The results are tested using network simulator ns 2 and prove that the proposed ART 2 RED AQM best suits for maintaining the queue stability by reducing the packet loss