The Internet traffic characteristics of a sub-network are analyzed from the correlation point of view. It is easy to recognize that the traffic histogram of the sub-network shows the seasonal variation phenomenon with a 24-hour period due to the daily usage. This property is important for traffic modeling. The autoregressive integrated moving average (ARIMA) model is applied to characterize the properties of sun-network traffic. The modeling performance is evaluated by examining the match of histogram and moving average behavior. Our experimental results illustrate that the proposed model can effectively capture the characteristics of Internet traffic and can then be used as a suitable traffic model for the analysis of Internet performance.