[Journals]

[Conferences]

[Internet Drafts]

[MS.c. Thesis]

[Ph.D. Thesis]

This page was last
updated by
Stefan Schmid on
Saturday, 02-Jan-2016 21:31:33 UTC
 

QoS based Real-Time Audio Streaming on the Internet

Submitted for the degree of M.Sc. in Computer Science

( “Diplom Informatik”)
 

University of Ulm, Germany

April 1999
 

Abstract

Live audio streaming is an important component of Internet multimedia. Today's Internet, however, offers only poor support for such streams due to the lack of Quality of Service (QoS) capabilities. The work presented in this thesis discusses the problems of real­time audio streaming and investigates solutions for improving the QoS provided in current and next generation IP networks.

The contributions of this work can be divided into three parts: First, the thesis provides an in­depth discussion on important Internet multimedia protocols and mechanisms intended to improve the QoS of real­time audio streaming in the Internet. The study evaluates these protocols and mechanisms for use within (interactive) real­time streaming applications and highlights a set of mandatory and recommended techniques. Current network level QoS mechanisms, especially DiffServ and IntServ, are compared and examined with respect to their prospective use in a future QoS framework for the Internet.

Second, WebAudio, a state­of­the­art real­time audio streaming application, is introduced. This uni­directional audio streaming application is designed in a platform independent fashion. The ``open'' stream control interface enables easy Web integration. The WebAudio server and client applications improve the streaming QoS by means of resource reservation and adaptation. The multi­streaming capabilities of WebAudio allow the applications to be used for audio conferencing. An extensive discussion on the application architecture and the implementation issues of the WebAudio server and client is presented.

Third, the thesis explores the benefits of IPv6 for network QoS mechanisms such as IntServ/RSVP. This work questions whether there is an efficiency gain in packet classification due to the employment of network level flow labels. Based on WebAudio and on a flow label­enabled and extended RSVP implementation, a series of experiments has been performed. The results indicate that flow label based packet classification performs of the order of 2­4 times better than standard IPv4 classification, and it outperforms standard IPv6 classification by about 3­6 times.

The combined results of the work presented in this thesis make a strong contribution towards understanding how to improve QoS for future multimedia applications running over the Internet.

[Download Diplom-Thesis.pdf (2MB)]

This site was last updated on [an error occurred while processing this directive]
Page Design and Content Copyright © by Stefan Schmid