网络冗余流量的柯西-拉普拉斯多分形小波模型

为了刻画网络冗余流量在小时间尺度下的局部特性,提出了基于柯西-拉普拉斯多分形小波模型及其参数估计算法. 采用联合分布函数来描述冗余流量的局部表征,即分别用柯西分布和拉普拉斯分布计算冗余流量的重尾和尖峰参数乘法因子;通过概率比较方法获取小波系数和尺度系数的比例参数阈值,以界定两种不同分布的参数范围. 实验结果表明,提出的模型能准确且有效地描述网络冗余流量在小时间尺度下的多分形特性.

A Cauchy-Laplace Multifractal Wavelet Model for Network Redundant Traffic

In order to characterize local features of network redundant traffic on small-time scale more ac-curately, a new Cauchy-Laplace multifractal wavelet model was proposed. An algorithm for estimating pa-rameters of wavelets was also put forward. A joint distribution function was adopted to describe local fea-tures, i. e. , Cauchy and Laplace distributions were used to obtain the parameter multiply factors for heavy-tailed and spike features, respectively. A threshold for ratio of wavelets to scaling parameters, which decides how these two distributions affected redundant traffic modeling, was achieved by probabili-ty comparison. Experiments show that the proposed model can well characterize small-time scale multi-fractal features of network redundant traffic.