Filtering P2P TV traffic

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

The importance of watching videos and TV programmes on Internet as possible ways of distribution of media contents is on the increase. The objective of my thesis is to focus on P2P networks with special focus on the various potential P2P-TV solutions, then to implement a set of processes that enables the division of already recorded Internet based data transmission into various traffic categories.

The first part of my thesis focuses on the goals, functions and types of P2P networks. I classify the already available networks and the ones under development along various criteria. Such criteria are the logical and topological structure of networks, the distribution solutions and the bandwidth demand. I give a description of media contents distributed over Internet and a classification of videos on the basis of their various characteristics.

Transmission over Internet results from a variety of applications and sources, is composed of a variety of protocols and it would be worth knowing what kind of data they contain in order to gain the right statistical data on the distribution of data. This information proves useful because it gives information about the data passing through the network and the user activities on the Internet as well as it allows us to scale the new networks to the significant data transmission. Therefore it is worth developing traffic filtering methods and applications that enables us to categorize already recorded data transmission on the basis of the nature of the transmitted data. I implement such a filtering algorithm in this paper. Traffic is recorded with the help of NetFlow, which allows me to record data transmission over a significant national Internet service provider. I use the text processing and text manipulating functions of AWK language during the analysis. By the end of the study, I can separate P2P, non-P2P, P2P voice and other types of transmission. I use the resulting data for analysis, to turn it into diagrams and to draw conclusions.

Downloads

Please sign in to download the files of this thesis.