The doctoral dissertations of the former Helsinki University of Technology (TKK) and Aalto University Schools of Technology (CHEM, ELEC, ENG, SCI) published in electronic format are available in the electronic publications archive of Aalto University - Aaltodoc.
|
|
|
On Traffic Classification and its Applications in the Internet
Mika Ilvesmäki
Dissertation for the degree of Doctor of Science in Technology to be presented with due permission of the Department of Electrical and Communications Engineering for public examination and debate in Auditorium S4 at Helsinki University of Technology (Espoo, Finland) on the 3rd of June, 2005, at 12 noon.
Dissertation in PDF format (ISBN 951-22-7692-5) [11607 KB]
Dissertation is also available in print (ISBN 951-22-7691-7)
Abstract
In this work, the methods and applications of traffic classification in the Internet are examined in detail. First, we define and discuss the conceptual environment of traffic classification. We then discuss the performance issues of traffic classification and define a method of visualization to compare the performance of traffic classification implementations.
Previously introduced methods of traffic classification: the static applications, the packet count and the list classifiers are compared with each other. We find these methods to perform quite well when analyzed as performing in an IP router, but to be rather ambiguous as to the effect they cause to the user.
We introduce an implementation of dynamic traffic classification to two classes using learning vector quantization (LVQ) for flow analysis data and find it to perform well in a simulated environment using flow analysis made on traffic measurements. In comparison to the previous methods of traffic classification, we see that the LVQ classifier has adequate performance. We also study a method of traffic classification using consecutive flow analysis with varying values of the parameters of the flow and find that we are able to classify traffic to 2 or 3 different classes. Within the classes the applications are similar in measured behavior and thus may provide help in realizing some advanced Internet service architectures.
Finally, we also observe the application of the dynamic classifier in an Internet router and in the Internet itself. We argue that the implementation of the dynamic classification method is feasible in the network.
Keywords: traffic measurements, traffic classification, Internet, neural networks
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
© 2005 Helsinki University of Technology
Last update 2011-05-26