Skip to main content

Distributed Multimedia Operating System for the 1990s


Distributed systems, multimedia and developments in processor and memory technology are changing the way in which systems should be designed to such an extent that the systems technology in common use in academia and industry today is rapidly becoming inadequate.

To incorporate full-speed and high digital audio and video in a distributed system means building from scratch, as there are no systems around with sufficient support for high-performance, low-overhead soft real time.

PEGASUS aims to design and build a prototype of the sort of system likely to become commonplace in research environments by the end of the decade. The goals are to design and build a distributed storage service that can work on a world-wide scale, that provides high levels of reliability and availability, and that is capable of storing multimedia documents, including digital video, in real time; to design and build the multimedia infrastructure that will allow activities such as video conferencing, multimedia document exchanges and multimedia communication; and to design and build a demonstration application.


The project brings together partners with expertise in multimedia, high-speed networks, security and naming and with expertise in distributed operating systems, real time, and file storage. This expertise will be put to use to construct an architecture for a distributed multimedia system centered around the concepts of desk-area networks, ATM devices, wide-address-space processors, and the use of typed data ( ie object orientation).

The project will develop both hardware and systems software. The software will not be based on Unix, although Unix will be integrated with the system (there will be Unix-based servers). The overall systems architecture will be developed in parallel with small rapid-prototyping experiments as contributions to a more detailed architecture. In the second half of the project, a demonstration prototype application will be developed as proof of the concept.


The systems research community has already expressed interest in the work.


Drienerlolaan, 5217
7500 AE Enschede

Participants (1)

University of Cambridge
United Kingdom
Free School Lane
CB2 3RF Cambridge