Fakultät für Informatik und Mathematik


MIP 0703

Paper Description

BibTeX entry

@ incollection{MIP-0703,
author="I. Dedinski, H. de Meer",
title="Advanced Application-Level Crawling Technique for Popular Filesharing Systems",
institution="Fakult{\"a}t f{\"u}r Informatik und Mathematik, Universit{\"a}t Passau",


P2P filesharing systems are causing the largest traffic ammount in todays Internet, which explains the interest ot the research community. On the other hand, most of the filesharing traffic is caused by the exchange of illegal content. That makes research participation in such systems hard, since the systems try to protect themselves from observation. This paper presents an application level crawling technique for current filesharing systems that exploits the minimal openness of the filesharing system to perform a broadband content scan with minimum ressource usage. Such a technique can be used to continuously scan a filesharing system. The gathered information can be used by researchers for studying the dynamics of P2P systems or by companies trying to protect their copyrights. Is also could be usefull to influence the behavior of such P2P systems, e.g., by an ISP traffic engineers. The technique was extensively evaluated through a series of measurements in the eDonkey filesharing network. The information gathered gives interesting insights about the behaviour of the users doing filesharing. Some behaviour patterns were found that influence the performance of the suggested technique in a very positive way, proving its feasability. These patterns inidcate that a filesharing system should not only be regarded as a technical system, but has to be also viewed as a social network. 

Paper itself