Search   |   Back Issues   |   Author Index   |   Title Index   |   Contents

Conference Report

spacer

D-Lib Magazine
January/February 2007

Volume 13 Number 1/2

ISSN 1082-9873

Report on CORDRA @ WORK

Held November 13-14, 2006, in Memphis, Tennessee

 

Nigel Ward
Director, Australian ADL Partnership Laboratory
<nward@adlaustralia.org>

Heather E. Walls
Editor, Workforce ADL Co-Lab
<hewalls@workforcecolab.org>

Red Line

spacer

The CORDRA @ WORK [1] meeting was held at the Workforce ADL Co-Lab, located on the University of Memphis campus, on November 13 and 14, 2006. The meeting brought together international stakeholders in Content Object Repository Discovery and Registration/Resolution Architecture (CORDRA) [2], a model for repository federations. Attendance numbered 71, and included representatives from North America, Europe, Asia and Australia.

A highlight of the meeting was seeing what a vibrant and diverse implementation community had developed since the initial CORDRA workshop in 2005 [3]. While no formal papers were presented, leaders of the various projects gave status reports on CORDRA and related repository federation projects. Two panel discussions provided opportunities for an exchange of information and ideas about the future of CORDRA.

The meeting was divided into three themed sessions. The first session focused on projects positioned as part of the CORDRA initiative, the second focused on non-CORDRA repository federation and distributed discovery projects, and the final session addressed related technologies.

CORDRA and CORDRA registries

Larry Lannom of the Corporation for National Research Initiatives (CNRI) kicked off the meeting with an update on CORDRA, which he succinctly described as a collaborative effort to define a logical architecture to federate repositories of learning content. He said the aim was to create an enabling infrastructure for finding, accessing and reusing learning content. Using the Internet as an analogy, Larry asked what was the learning content infrastructure equivalent of the TCP/IP layer? Good infrastructure must not get in the way of future developments and "layers" that will be superimposed on it. While CORDRA is currently learning-centric, it has the potential to spread beyond the learning community.

Avron Barr (ADL) and Henry Jerez (CNRI) spoke about the ADL-Registry (ADL-R) [4], a US Department of Defense (DoD) implementation of CORDRA that is progressing toward large-scale operation. The ADL-R technical operations have recently been supported by a DoD policy that stipulates all DoD agencies must look for existing learning content before authoring new content. Effort is now focused on developing a community of early adopters through the establishment of the ADL-R User's Group (ADL-RUG) [5] and implementation of an interface that supports safe and secure access.

ADL-R development has focused on producing a generic registry implementation that can be released under an open source license in the future. The implementation is deliberately agnostic about content and metadata and is modular, allowing replacement of any of the four main components (the registration engine, the validation logic, the data repository and the indexing engine). Current development is focused on REST and Web Services interfaces to the registry and on a prototype for the "registry of registries" functionality.

The session also included progress reports on the following CORDRA development efforts:

  • The DoD-R implementation of the CORDRA model being developed for the counterterrorism community. The implementation differs from the ADL-R in that it must support diverse content types and must respect each repository's access control policies at discovery time. The implementation is building on the experience of the ADL-R implementation to tackle new problems such as access control, declarative policies and versioning of content.
  • FeDCOR [6], a federation of DSpace [7] institutional repositories based on the CORDRA model.
  • The TILIS (Transport Integrated Learning Information System) [8] project, a repository federation infrastructure being developed for the transportation and logistics industry in Australia on a commercial basis.

Other registry and discovery projects

The theme for the second session was non-CORDRA registry and distributed discovery projects. Presentations included:

  • Global Learning Objects Brokered Exchange (GLOBE) [9], an SQI-based [10] implementation of federated search across international repositories in Europe, Australia, Canada, the U.S.A., and Japan.
  • Eleonet (European Learning Object Network), an EU registry of learning object metadata. The solution mandates SCORM [11] as a content type, LOM [12] for metadata and DOI [13] for digital identifiers.
  • The EUN (European Schoolnet) Learning Resource Exchange [14], which aims to support discovery of learning content in digital repositories in over 20 European countries.
  • The NIME-glad gateway (National Institute of Multimedia Education - gateway to learning ability development) [15] for e-Learning in Japan, which aggregates LOM for Japanese repositories and provides the GLOBE federated search service to access international repositories.
  • The China Digital Museum Project [16], which is federating DSpace institutional repositories to replicate content and metadata for preservation purposes. The federation uses Handles [17] to uniquely identify content and track multiple content locations. Metadata and content is shared by harvesting METS (Metadata Encoding and Transmission Standard) [18] Dissemination Information Packages via OAI-PMH [19]. The project is also currently investigating a peer-to-peer approach to content and metadata replication.
  • The Australian Federated Repositories for Education (FRED) project, which is developing a service-oriented representation of repository federations and software toolkits to help Australian education communities develop federations.
  • Canadian Department of National Defense plans for displaying ADL-R registered content in a learning management system (LMS). Non-technical barriers to this approach included determining content sharing, licensing and classification arrangements with allies.

The session highlighted that there are two common approaches to federated discovery: collecting and indexing metadata in a central registry, and distributing searches across distributed repositories. During his GLOBE presentation, Stefaan Ternier from Katholieke Universiteit Leuven (KU Leuven) suggested that the federated search approach was useful when repositories were not willing to allow their metadata to be collected in a registry. David Massart (European Schoolnet) noted that an IMS Special Interest Group has just been established to compile a 'cookbook' of solutions for federated discovery.

Many of the projects reported on practical challenges when establishing repository federations. These included establishing a critical mass of content in a federation, cheaply creating quality metadata so content could be effectively discovered, and harmonizing vocabularies across a variety of communities, countries and curricula. Suggestions for creating quality metadata included a combination of educating publishers, automation and social tagging.

An emerging requirement for most of the projects was the need to support content pricing, licensing and access restrictions. A technical requirement included the need for discovery based on access, classification or licensing criteria.

Some of the projects represented have been addressing issues of content replication such as understanding when multiple copies of the same content appear in different repositories, and how content should be shared across multiple registry federations.

Related technologies

The final half-day session was devoted to technologies related to CORDRA.

Rob Wilson (RWCS Ltd.) and Larry Lannom (CNRI) discussed persistent identifiers, a key component of the CORDRA model. Rob noted the value of creating infrastructures that can be repurposed over time; the importance of separating notions of identification, location and metadata; and the value of putting identifier management at the core of discovery and access infrastructure. Larry described the Handle System, an identifier resolution service for Internet resources that is used by current CORDRA implementations. The system is logically centralized but physically distributed. It associates an identifier with one or more typed values. The system is based on an open, well-defined protocol; the public implementation of the protocol is optimized for speed of retrieval and reliability.

Larry noted some new projects investigating Handles as their core infrastructure, including:

  • a Grid incubator project examining the inclusion of Handles into the Globus toolkit [20],
  • the Transient Network Architecture project, which is part of the National Science Foundation's Future Internet Network Design (FIND) effort [21], that is examining what the Internet might look like if all network components, including mobile devices, were given persistent identifiers.

Lyle Winton (Department of Education, Science and Training [DEST] Australia) expanded on Larry's mention of the Grid, providing an overview of the Grid infrastructure that many research communities use to integrate shared computing resources and for storing large amounts of scientific and research data.

Chi Nguyen (Meta Access Management System/Macquarie University's E-Learning Centre of Excellence [MAMS/MELCOE]) [22] and Kerry Blinco (DEST Australia) discussed standards-based approaches to federated authentication, authorization and access management. They noted that these were all critical issues to consider when discovering and providing access to content. While technical solutions exist, more work needs to be done on translating between technical terminology and business terminology.

Kerry closed out the session with a report on the e-Framework [23], a service-oriented approach to process and system integration across domains. The e-Framework provides a common technical vocabulary for describing interfaces between services. The common vocabulary enables better communication between vendors and consumers. Kerry suggested that the e-Framework approach could help CORDRA answer questions such as: How does CORDRA differ from other repository federation approaches? What role does policy play in the differences? What interfaces already exist? What interfaces do we need to work on?

Summary of other discussions

Throughout the meeting, many presenters asked the questions: "What is CORDRA? What should it become? How should we get there?" Participants discussed whether CORDRA should simply be a community of implementers developing repository federations, whether CORDRA should be an open model defining approaches to repository federation, or whether it should be a detailed repository federation specification with compliance criteria. While no consensus was reached, it was noted that if CORDRA becomes an open model or detailed specification, then it should describe requirements without dictating business rules. Participants also generally agreed that CORDRA should be generalized beyond the ADL-R implementation.

Allyn Radford (HarvestRoad) explained requirements from the commercial world to incorporate multiple types of content – including both legacy and virtual collections – into federations. Allyn suggested that "necessity is the mother of deviation", pointing out that the commercial world will develop solutions with or without CORDRA.

Several attendees lamented that there is currently no "off the shelf" CORDRA software solution available to organizations that are ready to federate.

More than one speaker mentioned that learning content is different from standard digital library content, mainly because it is not static: the creator engages with the content and metadata after publication, reworking it to suit flexible learning processes and goals.

Larry Lannom (CNRI) noted that federating repository federations was part of the original CORDRA vision. He suggested that there is a growing set of requirements and understandings about federating federations, and that perhaps it was time to start experimenting with possible solutions.

While CORDRA @ WORK officially ended at noon, the newly formed ADL-R Users Group (ADL-RUG) met informally in the afternoon.

Next steps

During the final panel session, participants discussed ideas for maintaining the community and progressing CORDRA development. Two practical ideas included further development of documentation defining the CORDRA model, and fostering interoperability experiments between existing CORDRA projects.

In response to these suggestions, the meeting organizers have established a mailing list for those interested in developing CORDRA documentation or establishing CORDRA interoperability experiments. The mailing list can be found at <http://groups-beta.google.com/group/CORDRA?hl=en>. The meeting organizers can also be contacted directly regarding other suggestions for progressing CORDRA by emailing <planning@cordra.net>.

Acknowledgements

The authors wish to thank Nick Nicholas and Lyle Winton from the Australian Department of Education, Science and Training (DEST) for taking extensive notes during the CORDRA @ Work meeting. Without their attention to detail, this report would not have been possible.

Meeting organizers were Daniel Rehak, Co-Director, Workforce ADL Co-Lab [24]; Philip Dodds, Chief Technical Architect, and Avron Barr, ADL Initiative [25]; Larry Lannom, Director of Information Management Technology, Corporation for National Research Initiatives [26]; and Nigel Ward, Director, Australian ADL Partnership Lab [27].

References

[1] CORDRA @ Work
<http://cordra.workforcecolab.org/>.

[2] CORDRA
<http://cordra.net/>.

[3] Issues in Federating Repositories: A Report on the First International CORDRA™ Workshop
<http://dx.doi.org/10.1045/march2005-kraan>.

[4] ADL Registry
<https://adlregistry.dtic.mil/>.

[5] ADL Registry Users Group
<http://adlcommunity.net/course/view.php?id=7>.

[6] FeDCOR: An Institutional CORDRA Registry
<http://dx.doi.org/10.1045/february2006-manepalli>.

[7] DSpace
<http://dspace.org/>.

[8] TILIS - Transport Integrated Learning Information System
<http://www.talc.com.au/TILIS/tabid/220/Default.aspx>.

[9] GLOBE
<http://globe-info.org/>.

[10] Simple Query Interface (SQI)
<http://www.prolearn-project.org/lori>.

[11] Sharable Content Object Reference Model (SCORM)
<http://www.adlnet.gov/scorm/index.cfm>.

[12] IEEE LOM 2002, IEEE Standard for Learning Object Metadata, IEEE Standards Department, Institute of Electrical and Electronic Engineers, Inc. (2002). IEEE-SA Standard 1484.12.1-2002.

[13] The Digital Object Identifier System
<http://www.doi.org/>.

[14] EUN Learning Resource Exchange
<http://lre.eun.org/>.

[15] NIME Glad
<http://nime-glad.nime.ac.jp/en/index.php>.

[16] Building a Distributed, Standards-based Repository Federation: The China Digital Museum Project
<http://dx.doi.org/10.1045/july2006-tansley>.

[17] Handle System®
<http://www.handle.net>.

[18] Metadata Encoding and Transmission Standard (METS)
<http://www.loc.gov/standards/mets/>.

[19] OAI-PMH (Open Archives Initiative Protocol for Metadata Harvesting)
<http://www.openarchives.org/OAI/openarchivesprotocol.html>.

[20] Globus Toolkit
<http://www.globus.org/toolkit/>.

[21] FIND initiative/Transient Network Architecture
<http://nsf-find.cs.umn.edu/>.

[22] Meta Access Management System (MAMS)
<http://www.melcoe.mq.edu.au/projects/MAMS/>.

[23] e-Framework
<http://www.e-framework.org/>.

[24] Workforce ADL Co-Lab
<http://workforcecolab.org/>.

[25] ADL Initiative
<http://adlnet.gov/>.

[26] Corporation for National Research Initiatives
<http://cnri.reston.va.us/>.

[27] Australian ADL Partnership Lab
<http://adlaustralia.org/>.

Links to presentations

Copyright © 2007 Nigel Ward and Heather E. Walls
spacer
spacer

Top | Contents
Search | Author Index | Title Index | Back Issues
Previous Opinion | In Brief
Home | E-mail the Editor

spacer
spacer

D-Lib Magazine Access Terms and Conditions

doi:10.1045/january2007-ward