Building the Info Grid
The Danish Electronic Research Library (DEFF) [1] offered a two-day event, Building the Info Grid [2], focusing on the recent and upcoming developments in digital information management, more specifically on the possibilites and challenges of providing integrated access to scholarly content and communication, via distributed technological services and infrastructural software.
In this report we will not cover all aspects of the conference, but rather focus on the specific topics that were the binding glue throughout the conference: Service-oriented Architecture (SOA); the Grid/Information Grid; Rights Management; Single Sign-on; and Google Scholar [3] development. The conference gave both overview and in-depth insights into terminology as well as more specific choices of solution such as Fedora [4] and Shibboleth [5].
The overall perspective of the conference bridged interests of libraries, computing centres, administration, scientific institutes, funding bodies, national research organisations and vendors, which contributed to a blend of participants and delegates from as many as nineteen countries.
Global, National, Local: Collaboration and/or Competition
The first theme of the conference, "Global, national, local - collaboration and/or competition", as well as the introductory remarks by Kim Østrup, Chairman of DEFF Steering committee and Deputy Director of IBM Denmark, dealt with one of the general challenges for building the information grid, namely on which level and under what disposition it is optimally tackled. A collaborative disposition became evident throughout the presentations, by the massive and manifold co-operative scenarios between public service providers and between nations. The issue of competition was predominantly broached by stating that vendors and companies should be 'taken on board'. Correspondingly, presentations of companies like Google and Elsevier stressed developmental activities and the ease of integrating their products. Anurag Acharya, Distinguished Engineer at Google presented Google Scholar and, later, Ale de Vries of Elsevier's Science Direct shared experiences from implementing new authentication technologies (i.e. Shibboleth).
A general scenario for building the information grid from a library perspective, was proposed by Peter Brantley, Director of Technology at California Digital Library, namely a complete reconceptualisation of libraries' self-image towards being service providers committed to evolving in a changing information world. Peter focused on the theme of 'Deploying Services and not Libraries'. Many telling points were made on the focus of digital libraries and the surrounding community: 'Libraries need to change radically, engage in institutional stakeholders and different allies, participate in the open-source community and to deploy service-oriented architectures and not more digital library silos.' Or, as it was simply stated in the end: 'Service-oriented Architectures are about putting down the paving, the yellow median stripe, and the shoulders for a fast road.'
Google Scholar's Anurag Acharya gave a vivid presentation on the ongoing Google Scholar work where the ultimate goal is to provide worldwide visibility to all research. The first and foremost task is not only about indexing full texts but to be indexing any parts of a document available, which might in the end be just a citation. The reality of achieving these goals seems however to be a long way off yet, as only a fraction of all research is currently available online. Furthermore, identifying duplicates and normalising the many shapes of a citation is a difficult task requiring many hours of analysis within the Google Scholar department - where would we be today if researchers had come across a common standard many years ago?
The Knowledge Exchange Office (KEO), a new joint collaborative initiative between the research funding organisations in Denmark (DEFF), Germany (DFG, the German Research Foundation), the Netherlands (SURF Foundation) and the United Kingdom (JISC - Joint Information Systems Committee), was presented by the newly appointed Programme Director, Diann Rusch-Feja. She gave an in-depth presentation on the plans and forthcoming years of work within the KEO. The KEO initiative has its official launch in December 2005, and will be aiming at developing closer working relationships between national agencies and bodies within Europe which are responsible for the strategic development of infrastructure and services. The KEO aims to be a point of contact for other organisations which are attempting to co-ordinate national activities in parallel areas, and facilitate joint policy development and identification of best practices, e.g. collaborative development and implementation of common standards for interoperability, including digital rights management and authorisation.
As part of the second day overviews of current activities in the KEO partner countries were given by Liz Lyon, UKOLN (UK), Norbert Lossau, Bielefeld University (Germany), Martin Feijen, SURF (Netherlands) and Mogens Sandfær, DTU Center for Knowledge Technology (Denmark).
Service-oriented Architectures: The Framework of the Grid?
The idea of Service-oriented Architectures (SOA) as a framework of the grid, was elaborated in the technical sessions. The repository software Fedora was presented by Sandy Payette, Researcher at Cornell Information Science, as a basic infrastructural element in SOAs that is generic enough to represent arbitrary data structures and yet specific as well as flexible enough to provide the technological interfaces and concepts for connecting heterogenic tools, services and functions that provide the basis of the structure. Carl Lagoze, Senior Research Associate at Cornell Information Science, provided a proof of concept for this approach by explaining the Fedora based project National Science Digital library funded by the US National Science Foundation. A paradigm of contextualising resources in an information network overlay architecture should supplant the currently prevailing "Search and Access" approach. Herbert Van de Sompel, Los Alamos National Laboratory, extended the quest of managing complex and distributed digital content in a semantically significant manner by retrieving and integrating content from different repositories deploying specifications and standards such as MPEG-21 and IMS-Content Packaging. It is worthy of note that these approaches lead away from standards particular to the library world towards a more direct connection with industry and business standards.
Identity and Rights Management: Single Sign-on to the Grid?
A final group of contributions raised the issue of user and rights management, i.e. authentication and authorisation, which is, obviously, a tremendously significant and simultaneously extraordinarily difficult task in a network of distributed systems operating on SOAs: seamless connection between system components and elegant transfer mechanisms for individual user profiles must be an essential characteristic of services in order to prevent confusing or even overtaxing end-users with ever-changing GUIs (Graphical User Interfaces) and authentication procedures while navigating between bibliographic systems, administrative tools or distributed directory services.
Shibboleth, emerging from the Internet2 Project - the technical-organisational development co-operative of over 200 universities in the USA and other partners across government and industry was put forward by speakers such as Peter Brantley as a promising technical approach. Vendors such as Elsevier are also running trials with these technologies. Further implementations of Shibboleth and alternative solutions were presented on behalf of the London School of Economics (John Paschoud), the Joint Information Systems Committee JISC (Terry Morrow) and DEFF (Arne Sørensen).
As part of the spin-off activities and further collaboration after the conference, the first European Fedora User Meeting [6] took place at the Center for Knowledge Technology at the Technical Knowledge Center of Denmark [7].
Most of the presentations [8] can be found at the conference Web site.
References
- The Danish Electronic Research Library (DEFF) http://www.deff.dk/default.aspx?lang=english
- Building the Info Grid http://seminar.deff.dk/
- Google Scholar http://scholar.google.com/
- Fedora http://www.fedora.info/
- Shibboleth http://shibboleth.internet2.edu/
- European Fedora User Meeting 2005 http://fedora.cvt.dk/meeting-2005/
- Technical Knowledge Center of Denmark http://www.dtv.dk/English.aspx
- Building the Info Grid: Speakers and Abstracts http://seminar.deff.dk/index.php?content=speakers