Multi-Fractal Traffic and Anomaly Detection in Computer Communications

This insightful work explores mono- and multi-fractal traffic theories, enhancing understanding of traffic dynamics and anomaly detection in networks.

In Multi-Fractal Traffic and Anomaly Detection in Computer Communications, the author delves into a comprehensive examination of both mono- and multi-fractal traffic theories. The book begins by exploring the foundations of long-range dependent time series and the complexities of 1/f noise. It further discusses ergodicity and predictability of traffic, providing insights into traffic modeling and simulation techniques. Additionally, the text covers stationarity tests, traffic measurement methodologies, and the crucial topic of anomaly detection within communications networks.

The author presents compelling evidence that mono-fractal long-range dependent (LRD) time series are ergodic, demonstrating that LRD traffic can be considered stationary. A significant focus is placed on the stationarity of multi-fractal traffic, which is shown to depend on varying observation time scales. To address these complexities, the book introduces multi-fractional generalized Cauchy processes and modified multi-fractional Gaussian noise, broadening the understanding of traffic dynamics. Moreover, it provides practical guidelines for determining the appropriate record length in traffic measurement, enhancing the accuracy of data collection.

Scholars and graduate students engaged in the study of network traffic within the realm of computer science will find Multi-Fractal Traffic and Anomaly Detection in Computer Communications particularly beneficial. The book not only serves as a theoretical foundation but also offers practical approaches to traffic simulation and anomaly detection, especially in the context of distributed denial-of-service attacks.