Testing the Gaussian assumption for self-similar teletraffic models

Both the fractional Brownian motion (fBm) and the autoregressive integrated moving average (ARIMA) models have been applied to teletraffic scenarios in recent years. These models became popular after the discovery that Ethernet and VBR video data appear to possess the property of self-similarity. However the results presented in this paper suggest that Ethernet data is more impulsive than traffic generated by these models.