Repository Networks: Past, Present and Future

1 Introduction

Repositories represent critical infrastructure for open science / open scholarship. They preserve and provide access to a wide range of valuable research and educational resources, including articles, pre-prints, research data, images, software, and so on. And when resourced properly, repositories are sustainable and long-lived services because they are, for the most part, collectively managed by research institutions and their libraries.

While digital repositories have been part of the landscape for decades, it was only in the early 2000s that the notion of the institutional repository (IR) was proposed, setting the scene for a significant increase in the number of open access repositories across the world. IRs were positioned as tools that would both transform the scholarly communication system, as well as offer a window to the output of an institution. Cliff Lynch, in his seminal piece about institutional repositories wrote,

“The development of the institutional repository emerged as a new strategy that allows universities to apply serious, systematic leverage to accelerate changes taking place in scholarship and scholarly communication, both moving beyond their historic relatively passive role of supporting established publishers in modernizing scholarly publishing through the licensing of digital content, and also scaling up beyond ad-hoc alliances, partnerships, and support arrangements with a few select faculty pioneers exploring more transformative new uses of the digital medium.”^[1]

Early in the evolution of institutional repositories, it was clear that the value of any individual repository could be greatly enhanced when it becomes part of a distributed network, enabling resources to be more discoverable, linked with other related content, and part of the larger international corpus of international research. A 2011 report by the Confederation of Open Access Repositories (COAR) articulated the case for the networked repository, “The real value of repositories lies in their potential to become an interconnected repository network – a network that can provide unified access to an aggregated set of scholarly and related outputs that machines and researchers can work with in new ways.”^[1] However, despite this value proposition, the networking aspect of repositories has advanced relatively slowly.

2 The technical evolution of the network

3 Networks are more than technologies

Repository networks can be about a lot more than just technologies and services. While the specific technologies adopted by repositories are critical for the network, there is also an important non-technical component that should not be overlooked. Building a community around the services and technologies can significantly increase the value and impact of the network. Some repository networks, such as OpenAIRE, have recognized this important human aspect to the networks and actively seek to nurture the repository community through a variety of activities. A recent COAR-SPARC Expert Group^[10] developing a vision for a US network of repositories identified many possible ‘features’ of a repository network, including services such as discovery and shared infrastructure, but also a host of other characteristics such as communities of practice, common standards, collective investments, and advocacy. It is extremely helpful to have a compelling vision across the network about shared understanding about its role in the ecosystem. This can bring the community together and act as an incentive for repositories to engage more actively in the network services.

Given that many aggregators are heavily reliant on metadata, high quality and standardized metadata contributes to the quality of services this type of network can offer. This includes agreeing on a common format for exposing the metadata, the introduction of Permanent Identifiers (PIDs) and the use of controlled vocabularies, which ensure the same word is used to mean the same thing. New communities are being formed to collaborate on metadata and data curation activities across institutions. The Data Curation Network (DCN), for example, is a group of “professional data curators, data management experts, data repository administrators, disciplinary scientists and scholars that represent academic institutions and non-profit data repositories”^[11]. DCN brings together a pool of data curation experts to curate a wider variety of data types, bringing in subject expertise and expanding curation efforts beyond what any single institution might be able to offer.

4 Interoperability

Interoperability across repositories is a fundamental characteristic of a repository network. Since its inception in 2009, a major priority for COAR has been to advance repository interoperability. A 2012 COAR report argues that to fully optimize repositories, they must be interoperable, “specifically, that repositories follow consistent guidelines, protocols, and standards which allow them to communicate with each other; connect with other systems; and transfer information, metadata, and digital objects between each other”^[12]. Specific priority areas for repository interoperability were further articulated in the 2015 COAR report, Future Directions for Repository Interoperability, prepared by Friedrich Summann and Kathleen Shearer^[13].

In a similar sense repository aggregator should also be interoperable. Current aggregator services differ in their scope, operations and functionalities, and often have distinct requirements based on their specific user communities and jurisdictions. However, in order to support the global nature of research, collaboration and dialogue across networks is critical. As such, COAR has also been working to foster greater cooperation between repository networks by identifying common principles and areas of collaboration that can lead to concrete actions and the development of global services. The aim is to ensure a general level of interoperability, while still supporting the unique functions and use cases of each network.

5 Discussion

Despite the obvious value proposition for networks in the repository landscape, there are a few challenges that have contributed to their relatively slow progression. Because institutional repositories were initially conceived as “digital collections that capture and preserve the intellectual output of university communities”^[14]. This may have led to the emphasis on the value of IRs to the institution, as opposed to the repository as a node in a broader network. In addition, the technical architecture of repositories has likely contributed to the stand-alone nature of the services. Repositories were originally designed as digital libraries and content management systems, rather than resources-oriented, and therefore did not reflect the highly networked architecture of the web. As such, the networking functionalities of repositories have only been incorporated incrementally into the platforms over time, leading to relatively slow advancement over the past 20 years. Contributing to this issue, is that the communities of practice around repositories are mainly based on the software platforms, creating software siloes in the repository landscape, making interoperable networking functionalities more difficult to implement.

The role of COAR in bridging these silos has become apparent and is increasingly crucial to help advance the role of repositories in the ecosystem and enhance the networking capacity of repositories in an interoperable manner. This has been demonstrated by the widespread adoption of recommendations of the COAR Next Generation Repository Initiative^[15] in the different repository platforms, and subsequently with the widespread community interest in adopting the linked data notification technologies as proposed through the COAR Notify Project.

Undoubtedly, technologies will continue to evolve, as will the ways in which repositories are networked with each other. To that end, COAR will seek to maintain its role in progressing platform agnostic interoperability and the networking of repositories worldwide.

This article is dedicated to Friedrich Summann in thanks for his many contributions to the repository community throughout his career. Through his work with BASE and COAR, he has played a significant role in advancing the role of repository networks and repository interoperability over the last two decades.