Research and development issues for large-scale multimedia information systems

During the last few years a convergence in the industries of computers and software, telecommunications, publishing, and consumer electronics has been observed. This is of major industrial importance. The convergence of the relevant technologies will allow information to be distributed on demand from very large information servers to home electronics devices (called Set-Top Units, STUS) through a variety of high-capacity telecommunication media. Private and public organizations from these industries have formed the Digital Audio-Visual Council (DAVIC). DAVIC has produced a draft standard that describes the structure for information servers, telecommunication systems, STUS, and the communication protocols. The objective is to achieve “rapid, industry-driven consensus” in avoiding problems of incompatibility while facilitating the fast development of the infrastructure from different manufacturers. The multimedia information servers will store very large amounts of multimedia data to support a large number of applications, such as movies (and other audiovisual data) on demand, teleshopping, news on demand, telemedicine, and the like. In this article we focus on the research problems in constructing such large-scale multimedia information servers. In our discussions we present the approach we have taken in building the KYDONIA multimedia server [Christodoulakis et al. 1993] and we highlight the open research problems being addressed within the context of the ESPRIT Long-Term Research project HERMES, solutions to which are being implemented in a DAVIC-compatible multimedia server within the context of the ACTS SICMA project. The presentation is organized into database-system issues, storage-system issues, and issues in supporting delay-sensitive data streams.