摘要： 目前的流量抽样测量方法主要基于传统的数学理论，并没有考虑到实际网络流量的特征，基于此，提出基于FARIMA流量预测的抽样方法，根据流量预测值动态调整抽样率，既减轻了CPU的负载，又节省了存储空间。通过对比实际使用中的流量抽样测量方法取得的数据报文样本均值和Hurst参数，表明该方法能够正确体现原始数据的流量行为统计特征。
关键词:
网络测量,
流量抽样,
自回归分数整合滑动平均模型,
流量预测
Abstract: Current traffic sampling methods are based on mathematic theory without considering about real network traffic. So according to traffic prediction to dynamically adjust sampling rate, this paper proposes a new sampling method based on Fractal Auto Regression Integrated Moving Average(FARIMA) traffic prediction. The method not only reduces CPU load, but storage space as well. Compared with sample mean and Hurst parameter of traffic data produced from currently used sampling method, the experiment results can generate more accurate traffic statistics of sampling traffic.
Key words:
network measurement,
traffic sampling,
Fractal Auto Regression Integrated Moving Average(FARIMA) model,
traffic prediction
中图分类号:
潘乔, 罗辛, 王高丽, 裴昌幸. 基于FARIMA模型的流量抽样测量方法[J]. 计算机工程, 2010, 36(15): 7-8,11.
BO Jiao, LUO Xin, WANG Gao-Li, FEI Chang-Nie. Traffic Sampling Measurement Method Based on FARIMA Model[J]. Computer Engineering, 2010, 36(15): 7-8,11.