ERCIM News No.45 - April 2001 [contents]
GRIDs in ERCIM
by Keith G Jeffery
GRIDs is both a new and an old concept. Many of the components have been the subject of R&D and some exist as commercial products. The GRIDs concept represents many different things to different people: metacomputing, distributed computing, advanced networking, distributed database, information retrieval, digital libraries, hypermedia, cooperative working, knowledge management, advanced user interfaces, mobile and pervasive computing and many others. More importantly, end-users see the GRIDs technology as a means to an end to improve quality, speed of working and cooperation in their field. GRIDs will deliver the required information in an appropriate form to the right place in a timely fashion. The novelty of GRIDs lies in the information systems engineering required in generating missing components and putting the components together.
In the Summer of 1999, the UK Director General of the Research Councils (John Taylor, whose article follows on page 10) was considering the Research Councils Strategic Review - the plan for the next few years. I was asked by the CLRC Chief Executive to produce a brief IT Strategy paper. This paper proposed a 3-layer architecture of computation / data grid with networking as the base, an information grid for integrating heterogeneous information into a homogeneous presentation above and on the top a knowledge grid for extraction of knowledge from information and assists for user control of processes (see Figure 1); and it suggested a business proposition that scientific application requirements could require the development of middleware and metadata which would be technology transferred to commercial e-business.
Of course, the basic ideas in that paper were not new and came from previous work my department (and the sister department in CLRC, Computational Science and Engineering) had done in a European context with industry, academia and especially ERCIM colleagues over many years, and especially our work with W3C. We also noted the US work on a computation / data GRID (equivalent to the lowest of the 3 layers in the UK proposed architecture) and especially the book The GRID by Ian Foster and Carl Kesselman known colloquially as The GRID Bible (see Figure 2).
Knowing of work on GRIDs component technologies among colleagues in ERCIM Institutes over several years, I proposed to the ERCIM Directors in November 2000 that we set up a GRIDs Task Force. It is not a traditional Working Group because it relies on contributing middleware and metadata technologies from Working Groups on Digital Libraries, (including Cultural Heritage), Database, User Interfaces for All, Environmental Modelling, Health and Information Technology, e-Commerce, e-Learning and even from other teams who have not yet formed an ERCIM Working Group (eg metacomputing). Each community has its own view of GRIDs technology, emphasising different aspects or layers.
This Task Force initiative was stimulated partly by the emerging ideas on ERA (European Research Area) from Commissioner Philippe Busquin (http://europa.eu.int/comm/commissioners/busquin/index_en.htm) from January 2000 to the resolution of June 2000 and beyond (see ERCIM News no. 42, page 5). Part of the push came from the developments in the USA and also from the EC-funded Datagrid Project mainly for particle physics but involving several ERCIM institutes and associated universities and industries.
The Task Force has been set up as follows: the ERCIM member institutes have proposed a contact person or persons and we are busy now collecting data for an up-to-date picture of GRIDs in the ERCIM-represented countries. We are also collecting plans for the future. The general aim is to coordinate activity in ERCIM and associated academic and industrial partners such that we maximise European potential in GRIDs development. Discussions with EC (European Commission) officials indicate that perhaps they are considering measures to leverage national efforts in GRIDs which in aggregate are much larger than the resources the EC has at its disposal for this activity.
Figure 1: GRIDs 3-layer Architecture. Figure 2: The GRID Bible.
The following articles present a flavour of the dynamic GRIDs scene in Europe: John Taylor, Director General of Research Councils sets out the UK national initiative that he has championed for the community. There are articles on GRIDs in some representative countries, which indicate the breadth of approach yet the commonalities that exist. There are articles on metacomputing and related technologies in the computation/data GRID layer. Applications of the GRIDs technology in the Information layer show a rich diversity touching almost all areas of scientific research and development using advanced information technology. The article on a postgraduate course in GRIDs technology surely a demonstration of abstracted knowledge for the knowledge layer - demonstrates that the technology is coming of age. The other major technologies for the knowledge layer data mining (to generate knowledge) and e-publication (to disseminate knowledge) have been documented elsewhere.
Lord Sainsbury the UK Minister for Science and Innovation in his keynote article (page 3) provides the rallying call. We in ERCIM have the framework to coordinate and co-develop with industry and academia the GRIDs technology on a 3-layer architecture (as distinct from the current US GRID technology concentrating primarily on the Computation / Data layer). The Information and Knowledge layers require traditional European strengths and competencies including multicultural multilinguality. They have commonality with Tim Berners-Lees vision of a semantic web (http://www.w3.org/ 2001/sw/) and the web of trust (http://www.w3.org/Talks/9704WWW6-tbl/slide22.htm), but go beyond and emphasise further the importance of middleware and metadata technologies. These technologies are prerequisites not only for successful e-business but also for successful e-science in a joined-up academic and commercial world such as in bioscience, information technology, materials science, concurrent engineering, healthcare, environment and others. They are no less prerequisites for e-Learning, e-Culture and other disciplines to complete the e-Society.
Europe, working with our colleagues worldwide, should make its contribution such that European science, business and society all benefit. ERCIM has a key role.
Keith G Jeffery - Director IT, CLRC
Rutherford Appleton Laboratory
Tel: +44 1235 44 6103