VRED: An improvement over RED algorithm by using queue length growth velocity

Active Queue Management (AQM) plays an important role in the Internetcongestion control. It tries to enhance congestion control, and to achieve tradeoffbetween bottleneck utilization and delay. Random Early Detection (RED) is the mostpopular active queue management algorithm that has been implemented in the inInternet routers and is trying to supply low delay and low packet loss. REDalgorithm uses only the average queue length as a congestion meter to triggerpacket dropping or packet marking as a congestion feedback. Since the averagequeue length considers only long–term behavior of any queue, this approach fails tosee instantaneous changes of the queue length and hence its reaction is not fastenough. This paperincludes another meter i.e. queue length growth velocity tomeasure congestion level in the router. This leads to fast reaction to the congestionand hence improves the network performance. Simulation results show that theproposed algorithm outperforms RED algorithm in terms of number of droppedpackets and bottleneck utilization.

VRED: An improvement over RED algorithm by using queue length growth velocity Keywords:

Congestion control , active queue management , Random Early Detection , packet dropping , utilization