Slowly decaying traffic models

OData support
Supervisor:
Dr. Telek Miklós
Department of Networked Systems and Services

The paper consist of three parts wherein I describe the theory of network traffic modelling (specificly heavy tailed traffic) , a realization of the concept and its validation as a valid network traffic generator.

In the first part as an introductory chapter I describe the signifficance and difficulties of network traffic modelling, then i examine the properties of heavy tailed network traffic. With this knowledge I explore alternative models for trafic generation and the tools which with we can verify the qualities of heavy tailed traffic.

In the second part I describe the my chosen model, Markov Modulated Poisson Processes, its implementation and the road to the finished model (which is implemented in C++).

In the third part I verify the model with fitting the models parameters so thats its generated output fits a real-life measured data-set with statistical precision.

In the closing chapter I evaluate the goodness of the fitting by the aforementioned tools and give a recommendation on an improvement of the realization/application.

Downloads

Please sign in to download the files of this thesis.