Jacco van Ossenbruggen
Jacco van Ossenbruggen, Lynda Hardman and Prof. dr. M.L. Kersten
Multimedia and Human-Computer Interaction group,
Information Systems, CWI,
PO Box 94079, 1090 GB Amsterdam.
tel: +31 20 592 4141
email: Jacco.van.Ossenbruggen@cwi.nl
Internet content providers need to deal with a wide range of different technologies to address the heterogenous infrastructure used by their target audiences. In addition to this variety of existing technologies, new ones are continuously being developed, and different applications and types of content require deployment of different techniques. In this dynamic environment, content providers face the challenge of delivering content for different applications to end-user devices (e.g high/low-end PCs, multiple generations of mobile phones and handhelds, etc) with a wide range of different capabilities (e.g. screens with different sizes, aspect ratios, resolution, color depth, etc).
For text-oriented Web content (e.g. HTML), recent developments address the problems that content providers face of keeping their content up-to-date with the latest technology while providing content that remains accessible across different (older) platforms and a wide range of devices. This includes the introduction of XML-related document transformation technology to support the semi-automatic production and maintenance of HTML content.
For media-centric Web content, however, support is still on a much lower level. Where high-level support for Web-content has focussed on text-oriented applications, multimedia research and standardization efforts have focussed on lower level issues such as audio and video encoding and compression techniques, high-bandwidth networking and quality of service management, streamed delivery of media content and multimedia document formats. As a result, there is currently a plethora of multimedia formats, protocols and encoding techniques, ranging from de facto and de jura standards and recommendations to proprietary and vendor-specific solutions. In addition, most content providers still rely on manually authored multimedia documents. This not only makes the production process expensive and time consuming, it also forces providers to target a relatively small portion of their potential audience. In particular, manual authoring processes require a decision early on in the production process as to which media encodings, document formats and exact bitrates a specific production will use.
The proposed research project aims at bridging the gap between the rich, but low-level functionality offered by state-of-the-art and next generation network and multimedia protocols on the one hand, and the high-level requirements of distributed multimedia applications on the other.
Dissemination of high quality media content over the internet requires multimedia applications that are able to adapt automatically and dynamically to the available service levels of the underlying infrastructure. In particular, research on the application layer and the boundaries with the network layer is needed to make intelligent use of the next generation network infrastructure. A framework is needed to support applications in making informed choices about the different available protocols, so they can (a) pick the most suitable and (b) adapt their content to the specific functionalities offered by the selected protocols. The framework needs to be aware of the extent to which these protocols can be controlled (e.g. by negotiating a guaranteed quality of service (QoS) level on behalf of the application). In addition, given a specific QoS level, the framework needs to provide support for deciding how the available resources can be most effectively utilized in the context of the application's specific requirements (e.g. by advising the application to reduce the bitrate of currently running, lower priority media streams to improve prefetching and buffering of higher-priority streams that will be needed in the next scene).
The research will complement and build on the results of two related projects CWI is currently participating in:
In addition, the proposed project will directly benefit from the activities of the group within the World Wide Web Consortium (W3C). In addition to the close involvement of CWI in the development of SMIL (W3C's XML-based, platform neutral multimedia document language for the Web), W3C is currently developing new activities that directly relate to the proposed research. Addressed issues include device independent authoring, specification of composite capabilites and platform preferences (CC/PP) and the modularization of different languages to better target specific devices.