The method of surrogate data is applied to the traffic data of the Internet in order to analyze important characteristics such as self-similarity. By using the iterated amplitude adjusted Fourier transformed (IAAFT) algorithm, the stochastic time series, which preserves the empirical histogram and autocorrelation function of the given data, can be generated. This means that the self-similarity of the original data is also preserved by this surrogate, since it depends on the autocorrelation function. By comparing the influences of the real world traffic and its IAAFT surrogate data, it is shown that both have almost the same effects to a transmission model. Thus, the IAAFT surrogate data can be considered useful for modeling of the network traffic.