摘要
针对网络业务流自相似参数的离散小波估计方法中大尺度下小波系数能量估计的不精确问题,提出了一种使用多孔算法估计自相似参数的方法,该方法可以提高自相似参数估计的精度。分析比较了多孔算法区别于离散小波变换的特点,剖析了自相似参数小波估计方法的机理和存在的问题,通过对真实网络业务使用离散小波和多孔算法估计方法估计自相似参数的试验,表明多孔算法在提高自相似参数估计精度方面比离散小波方法提高一个数量级,而且参数的95%置信区间有所减小。当然,多孔算法的机理限制了其在实时网络业务流管理方面的使用,更适合网络性能评估和网络规划。
In view of the wavelet- coefficient energy estimation coarseness at larger scale in wavelet estimator for network traffic self- similar parameters, this paper proposes a trous algorithm used instead of the ordinary discrete wavelet transform（DWT） to improve the self - similar parameter accuracy. After delving into the characteristic of a trous algorithm different from DWT and the mechanism of wavelet estimator, explaining the inner deficiency, the simulation experiments are taken with the simulated data- flow and the true network traffic using the two wavelet approaches comparatively. The results showed that a trous algorithm can improve self- similar parameter accuracy in a magnitude of one order more than DWT, and that the parameter 95% confidence intervals become less correspondingly. The principle of a trous limits its application to network real - time dispositions, but more suitable for network performance evaluation and network planning by estimating self- similar parameters with much more available network traffic data.
出处
《江西农业大学学报》
CAS
CSCD
北大核心
2005年第6期920-924,共5页
Acta Agriculturae Universitatis Jiangxiensis
基金
国家"863"网络安全管理与测试技术资助项目(863-301-05-03)
关键词
网络业务
自相似参数
多孔算法
小波估计方法
计算精度
network traffic
self- simialr parameters
atrous algorithm
wavelet estimator
caculation accuracy
