网络流量预测方法和实际预测分析

Prediction Methods of Network Traffic and Real Prediction Analysis

随着网络带宽的迅速增加以及各种网络服务的广泛应用,针对网络流量的建模以及预测日益重要并备受人们关注。为了更好地对网络流量进行建模和预测,文章一方面将现有的一些流量模型分短相关和长相关两类进行总结和分析;另一方面分析了常见的流量预测方法,尤其是神经网络、模糊理论以及小波分析。并在实际测量中,使用FIR神经网络模型对网络流量进行了实际的预测。最后,提出了流量预测在网络安全领域的应用前景和方向。

With the rapid increase of network bandwidth and broad application of network services,modeling and predicting of network traffic is more and more important and attractive.To model and predict the traffic effectively,traffic models are reviewed and analyzed by classifying into SRD and LRD.Then,several traffic prediction methods are presented,espocially ANN,fuzzy theory and wavelet.FIR ANN model is used in measuring real network traffic.Finally,the application and some thoughts of traffic prediction in network security are explored.