New traffic analysis and engineering methods for emerging technologies

Recent developments in telecommunications technology have far outpaced progress in teletraffic methods and practices. While engineering problems, such as capacity management and traffic measurements, remain highly relevant, this paper argues that new methods are required to address these problems in emerging networks and services. As an illustration, we discuss several traffic management problems that we have encountered in recent years that required the application of novel methods drawn from chaotic dynamical systems and fractal theory. The first concerns a phenomenon known as traffic synchronization which resulted in sustained oscillations in the queue lengths of an access tandem (or via) switch. We discuss how traffic synchronization arose in this case as a consequence of changing traffic characteristics and switch processor scheduling; and we summarize dynamical systems techniques based on chaotic maps to model, analyze and control this behavior. Secondly, we will discuss how ideas from the sciences of chaos and fractals can be applied to address the complexity of traffic in packet based networks and services such as X.25, ISDN, Internet, Ethernet, frame relay, SMDS, and ATM, and the development of more accurate engineering methods. We will summarize some recent developments in teletraffic measurement studies, introduce basic concepts in chaos and fractal theories, and establish the connection between these concepts and traffic analysis and engineering