自相似流量的主动队列管理算法

Self-Similar Traffic Control in Active Queue Management

该文提出了一种替代随机早期检测(RED)的主动队列管理算法:信号能量的小波分解(wavelet-decomposedsignalenergy,WDSE)算法。WDSE能够有效地对自相似流量进行拥塞检测和管理。采用小波的二阶子段滤波测量输入流量的能量,基于小波的多分辨率分析可以对到达流量的短期和长期波动进行预报,重新调整主动队列管理的丢弃概率。WDSE可以避免先前主动队列管理所难以解决的参数选择问题,如RED等。该文比较了WDSE和RED的动态性能,显示WDSE能够有效改善系统的暂态特性。仿真证明WDSE比RED能更快地响应负载的变化。此外,WDSE在控制瞬时队列长度的稳态值也具有更好的鲁棒性。

In this paper, an alternative AQM（Active Queue Management） is designed： WDSE （Wavelet-Decomposed Signal Energy） approach is capable of detecting and managing self-similar traffic. This approach uses wavelets and measures the energy of input traffic at the output of each dyadic subband filter. The scheme is based on the wavelet multiresolution analysis, which can be used to predict short and long term fluctuations in the incoming traffic pattern and readjust the marking drop probabilities without the difficulty of parameters selection in advance unlike RED. Comparing the WDSE and RED about the dynamic performance, improved transient performance of WDSE is illustrate. As evident from the simulation, the WDSE performs better, responding more quickly to load variations than the RED. Moreover, the WDSE is superior in robustly regulating the steady-state value of the queue level.