An Empirical Study of the Congestion Response of RealPlayer, Windows.pdfVIP

An Empirical Study of the Congestion Response of RealPlayer, Windows MediaPlayer and Quicktime Sven Hessler Michael Welzl Institute of Computer Science Institute of Computer Science University of Innsbruck University of Innsbruck 6020 Innsbruck, Austria 6020 Innsbruck, Austria sven.hessler@uibk.ac.at michael.welzl@uibk.ac.at Abstract ness of streaming media is given in [7, 8, 9, 15]. While the authors of [8, 15] investigate the performance of RealPlayer, Properly reacting to congestion is known to be a require- and Windows MediaPlayer [9] in an isolated environment, ment for maintaining the stability of the Internet; in addi- only Li et al. [7] compare two streaming applications (Me- tion, UDP based applications should do so for their own diaPlayer vs. RealPlayer). Though, all the aforementioned sake if they are delay- or loss-sensitive. We measured the research lacks the answer to the practical question a con- responsiveness of the three popular streaming media appli- tent provider might ask: “What program (RealPlayer, Me- cations RealPlayer, Windows MediaPlayer and Quicktime diaPlayer, Quicktime) is the best if congestion occurs?”. As with a varying amount of cross traffic and pr esent a com- a sensible extension to the research already done and to pro- parison of the results. vide content providers with applicable inform