Internet delay statistics: Determination and measurement of social connectedness using a dichotomous Hurst parameter

As Internet delay times become more important to the social status of recreational Internet users, the ability to differentiate among sites and users using statistical measures beyond average packet delays may be of commercial value. One such statistical measure is the Hurst parameter of a long-range dependent process. A Hurst parameter can be used as a measure of a communication system's burstiness or peculiarity. The notion of measuring a dichotomous Hurst parameter from segmented day and night delay data is introduced in order to capture changes in Internet delay statistics caused by human activity, site performance, and the autonomous systems of the Internet. This thesis documents the development and implementation of an analytical tool that can measure dichotomous Hurst parameters. Confirmations of dichotomous Hurst parameters were obtained at the 95% confidence level for several popular Internet sites from around the world. The theoretical basis and subsequent usefulness of a dichotomous Hurst parameter are also discussed in this thesis.