Capturing the Variability of Internet Flows in a Workload Generator for Network Simulators
Jörg Wallerich · mediaTUM – the media and publications repository of the Technical University Munich (Technical University Munich) · 2007
A rising number of traffic engineering and load adaptive routing mechanisms rely on the validity of Zipf's Law for the volume of network flows in order to control the majority of the traffic by treating a small set of large flows in a special manner. This approach requires these flows to show a significant amount of persistence. It has however been shown, that flows once classified as large can change their rate significantly over time. The goal of this work is the development of a workload generator for network simulators that captures the variability of flows an thus provides the means to evaluate novel approaches to traffic engineereing and load adaptive routing. To this end we develop a methodology to analyse network flows across the dimensions time and flow granularity. An analysis of several flow and packet level traces based on this methodology serves as the foundation for the generation of simulated network traffic with similar characteristics as found in our traces.