Effects on random early detection of the packet drop probability function with an adjustable nonlinearity

This paper proposed a packet drop probability function with an adjustable nonlinearity parameter in random early detection (RED) for active queue management of router to control network congestion. We investigated the effect on average size, throughput, retransmission rate, round-trip time, and fairness index two widely used loss-based congestion algorithms: Reno CUBIC. Simulations were performed Tail-Drop original RED clarify under different traffic loads. The results showed that nonlinear did not aggravate performance statistics because achieved high throughput while maintaining low-queuing delay, such as linear extremely heavy condition. Under light conditions, increasing bending degree accomplished by preventing excessive discarding packets.