Multi-Fractal Traffic and Anomaly Detection in Computer Communications
This book provides a comprehensive theory of mono- and multi-fractal traffic, including the basics of long-range dependent time series and 1/f noise, ergodicity and predictability of traffic, traffic modeling and simulation, stationarity tests of traffic, traffic measurement and the anomaly detection of traffic in communications networks.
Proving that mono-fractal LRD time series is ergodic, the book exhibits that LRD traffic is stationary. The author shows that the stationarity of multi-fractal traffic relies on observation time scales, and proposes multi-fractional generalized Cauchy processes and modified multi-fractional Gaussian noise. The book also establishes a set of guidelines for determining the record length of traffic in measurement. Moreover, it presents an approach of traffic simulation, as well as the anomaly detection of traffic under distributed-denial-of service attacks.
Scholars and graduates studying network traffic in computer science will find the book beneficial.
Ming Li, PhD, is a professor at Ocean College, Zhejiang University and the East China Normal University. He has been a contributor for many years to the fields of computer science, mathematics, statistics, and mechanics. He has authored more than 200 articles and 5 monographs on the subjects.