Large-scale distributed computing systems (LSDCSs) comprising several thousands to millions of individual computer systems (nodes) are coming into existence both as a consequence of technological advances and the demands made by applications. The BROADCAST project aims to develop the principles for understanding, designing and implementing LSDCSs.
The principles for understanding, designing, and implementing large scale distributed computing systems (LSDCS) have been developed. The 3 areas addressed are fundamental concepts, systems architecture, and systems engineering.
% The project has been particularly productive with respect to fundamental concepts. These results provide a firm grounding for the rest of the project. Deeper understanding of problems of failure detection, message ordering, group management, replica management and so on has been developed, leading to some promising algorithms and system structuring approaches. Several design and implementation efforts have been carried out on building distributed systems and applications. Project contributions in the areas covered by systems architecture and engineering include efficient schemes for remote referencing of objects, wide address operating systems, storage management, object support systems and application building tools.
APPROACH AND METHODS
The project is addressing the following tasks:
- Fundamental Concepts: evaluate existing and novel computational paradigms; design and evaluate fundamental distributed algorithms.
- Systems Architecture: develop the architecture of LSDCS, taking into account future application requirements and anticipated technological developments in high bandwidth communications, processors, peripherals, and storage systems.
- Systems Engineering: develop techniques and mechanisms to efficiently support the key elements of the architecture by exploiting the concepts and algorithms developed earlier; develop programming support and tools for building distributed applications.
The project places equal emphasis on building sound theoretical foundations and experimental work for demonstrating the validity of the principles developed. This includes building prototype systems and tools.
Existing wide-area communication networks in Europe and elsewhere can be viewed as forming the basic infrastructure necessary for building LSDCSs. Furthermore, increasingly large numbers of IT applications will depend on effective, coherent sharing of information as well as the possibility of collaboration between large groups of geographically dispersed users. The increase in scale adds a new dimension to the problems of building distributed systems, and requires rethinking of many of the existing computational paradigms, algorithms and structuring principles for distributed computing. The BROADCAST project will address these issues and will share results and insights gained from the theoretical and experimental work with colleagues in academia and industry through the normal channels of research publications, reports and workshops. In addition, the software developed within the project will be made available widely. The consortium expects to be able to transfer some results to industry through its industrial contacts.
78153 Le Chesnay
7500 AE Enschede