2006 | OriginalPaper | Buchkapitel
A Prediction Method of Network Traffic Using Time Series Models
verfasst von : Sangjoon Jung, Chonggun Kim, Younky Chung
Erschienen in: Computational Science and Its Applications - ICCSA 2006
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
This paper describes a method to derive an appropriate prediction model for network traffic and verify its trustfulness. The proposal is not only an analysis of network packets but also finding a prediction method for the number of packets. We use time series prediction models and evaluate whether the model can predict network traffic exactly or not. In order to predict network packets in a certain time, the AR, MA, ARMA, and ARIMA model are applied. Our purpose is to find the most suitable model which can express the nature of future traffic among these models. We evaluate whether the models satisfy the stationary assumption for network traffic. The stationary assumption is obtained by using ACF(Auto Correlation Function) and PACF(Partial Auto Correlation Function) using a suitable significance. As the result, when network traffic is classified on a daily basis, the AR model is a good method to predict network packets exactly. The proposed prediction method can be used on a routing protocol as a decision factor for managing traffic data dynamically in a network.