An Active Queue Management Algorithm to Enhance RED Stability

Abstract

RED is an active queue management algorithm to solve the end-to-end congestion problem of TCP. In order to further reduce the fluctuation of the queue length in the RED algorithm, improve the end-to-end throughput performance, and reduce the end-to-end delay jitter, an active queue management algorithm ARRED that enhances the stability of RED is proposed. The algorithm replaced the original linearly increasing packet drop probability function with the larger-scale ascending ridge function, which made the change of the packet loss probability function between the minimum and maximum queue length thresholds smoother. Considering the problem that the sharp change of the packet loss probability causes queue jitter when the maximum threshold is exceeded, a new parameter was set to adjust the packet drop probability. The simulation results show that the algorithm can effectively maintain the stability of the queue length, reduce the packet loss rate and delay jitter, and improve the network throughput performance.