Modeling and Forecasting of End-to-End Available Bandwidth in Wide Area Networks

This paper studied modeling and forecasting end-to-end available bandwidth in Wide Area Networks using time series models. We used the ARIMA(p,d,q) and SARIMA (p,d,q)(P,D,Q)[s] models to model available bandwidth data sets published by Stanford Linear Accelerator Center (SLAC) in 2009. We found that both models can be used to model available bandwidth along an end-to-end path. In addition, our results indicated that SARIMA outperforms ARIMA in forecasting future available bandwidth.