On the Effects of Non-Stationarity in Long-Range Dependence Tests

Abstract

Careful statistical analyses indicate that the measured traffic traces from live packet networks often contain non-stationary effects like level shifts or polynomial trends. In these cases several popular tests for long-range dependence can result in wrong conclusions and unreliable estimate of the Hurst parameter. In this paper both analytical and simulation investigations of the implications of these effects on several tests are presented. The results are also demonstrated with examples based on measured ATM traces. The use of these results can be utilized to avoid pitfalls in LRD traffic modeling.