Multiple paths to Open library infrastructure

---

This is the fourth in a series of blog posts exploring the Open landscape and OCLC’s place within it. This post presents some reflections on library management systems and different approaches to delivering ‘Open’ (i.e., extensible and interoperable) infrastructure as open-source or as proprietary software. It explores some of the market dynamics at play in this environment and the rationale behind OCLC’s approach to delivering Open solutions.

I had the pleasure of participating in Marshall Breeding’s annual Executive Perspectives panel at ALA Annual this past June. Marshall has been organizing these sessions for many years, bringing in a mix of technology and library leaders to address timely topics. OCLC has been a regular contributor to these discussions in the past, but I’ve been a spectator up to now rather than a speaker, so this year was a refreshing change of pace.

The topic for 2023 was Multiple Paths to Openness, with a panel representing a mix of organizational perspectives. In his introduction, Marshall noted that the marketplace for Open solutions is quite broad, encompassing a range of providers and business models. He designed the panel to include representatives from four types, which he described as:

• a for-profit, open-source software provider (Index Data)
• a non-profit open-source software provider (Equinox Open Library Initiative)
• a for-profit, proprietary software provider (Ex Libris/Clarivate)
• a non-profit, proprietary software provider (OCLC)

This typology, while somewhat reductive, did some useful rhetorical work by distancing Open from choices about how software is developed or licensed, to focus on how solutions are supported and sustained.

The session format – typically a moderated Q&A on a handful of agreed topics – underwent a last-minute change when Marshall opted to open the floor to questions from the audience at the beginning of the session, rather than the end. I thought this ‘unconference’ approach worked well, allowing the interests and concerns of attendees to shape the flow of conversation. The result was a much more animated and customer-centric dialogue driven by the concerns and questions of librarians and developers in the room rather than pre-set questions. I think the panelists learned as much from the session as the audience, which is exactly as it should be.

I liked the way Marshall framed the discussion, acknowledging at the outset that commercial motivations are an important part of the open-source ecosystem. There is no cost-free technology solution, there are simply different ways of accounting for and distributing costs. The costs of developing and maintaining open source code – the effort, expertise, and time that community members invest in supporting a common code base – are real, even if they are not recorded in a ledger or directly billed to customers.

Some institutions will absorb these costs by allocating existing development capacity to support or extend the code base. Others will outsource that activity to a service provider and/or hosting partner – this creates business opportunities for organizations (for-profit and non-profit alike) to concentrate capacity and offer it as a shared service. This is what Index Data and Equinox do, along with many other organizations (e.g., ByWater, K-Int, 4Science, PTFS Europe…). Think of Lyrasis, which provides hosting services as well as an organizational home to multiple open-source projects. Think of the Open Library Foundation, which provides back-office business services to selected open-source projects. Think of EBSCO, which is subsidizing some FOLIO development – not as a charitable contribution, but as part of a broader business strategy. Each of these organizations is managing direct and indirect costs associated with the very real work of sustaining open-source software. Some of those costs are inevitably passed on to libraries.

Then there is the matter of profit (AKA revenues in excess of expenses). Marshall included for-profit and non-profit providers on the panel not in the interest of sparring over motivations or margins, but as evidence that Open solutions can be supported by a variety of business models. Generating revenues in excess of expenses is in the interest of any provider who expects to stay in business. It’s how that surplus is used – to pay dividends to shareholders/owners or to reinvest in community infrastructure and public benefit – that differentiates for-profit from non-profit entities. As I’ve written elsewhere, OCLC is in the community infrastructure business, so sustainability is a primary objective. We see this as the best way to ensure that software and services evolve with and for libraries.

I appreciate Marshall’s pragmatic position on the benefits of Open – in the library software environment, it’s generally agreed that extensibility and interoperability deliver benefits to all players. Even competing software providers benefit from the ability to integrate solutions built on different platforms so that the benefits of market choice can be maintained. The question of how you deliver the benefits of Open solutions and why those choices matter is more relevant than subjective purity tests about what Open is.  

There are benefits and tradeoffs associated with all ‘paths to Open’ library infrastructure. Open source offers some advantages (a high level of transparency, opportunities for deep community collaboration and co-creation, the prospect of a codebase that may outlive its organizational home), as well as risks (uncertain sustainability given funding and reinvestment challenges, inefficiency in decentralized operations, lower incentives to exercise cost controls, governance challenges, difficulties in scaling data security…). Proprietary software can offer a clearer path to sustainability, practical advantages in managing security and scalability (including regulatory compliance in multiple jurisdictions), and efficiency and cost controls. Solutions built on proprietary code can be extensible and interoperable, but developer access is mediated. Community involvement in governance may be limited, especially in for-profit firms and even more so in privately held companies. The right approach to delivering Open software solutions will look different in different organizations.

I can’t speak to the choices or motivations of other organizations that participated in the panel. But I can clarify why OCLC’s path to delivering Open solutions is the right fit for our organization. OCLC is a global technology provider focused on advancing the interests of libraries of all types and sizes. We have built and sustained community infrastructure for more than half a century. Our first duty is to ensure that this infrastructure – a global data network and services that connect libraries, publishers, and information-seekers on the web – continues to evolve to meet library needs. We favor proprietary software solutions because they enable us to:

• scale services globally while delivering the level of security, regulatory compliance, and up-time that libraries expect and deserve;
• ensure the sustainability of shared infrastructure, upholding our commitment to libraries worldwide;
• meet library and vendor expectations for interoperability and extensibility via standards-based integrations and a mix of open and proprietary APIs.

OCLC supports integrations with every major library service provider and has done so for years – because we recognize that the vitality of shared infrastructure increases as participation grows, whether libraries are using management services provided by OCLC or some other partners. We also support community-driven development in the OCLC developer network. This goes beyond documenting and providing access to developer APIs. Since the launch of the first WorldCat Search API in 2008, OCLC has hosted annual events where library developers can learn from OCLC engineers and from one another. The most recent DevConnect attracted hundreds of registrants, a testament to the high level of community interest in extending OCLC services to meet specific local needs.

As a mission-driven organization, OCLC has a unique responsibility to stewardship of community resources. From a technology perspective, that means that we deliver solutions that maximize sustainability and security while supporting integration and interoperability. We support standards-based integrations to meet the needs of library vendors and proprietary APIs that enable non-library partners to work with us to make library collections and services more visible on the web. We work closely with library and non-library partners to facilitate the flow of information across platforms. For an organization like OCLC, developing solutions in proprietary, extensible code is the right strategic choice.

In the weeks since Marshall’s panel, there have been some interesting developments in the library open source sector. A project to deliver a consortial borrowing system on top of the open-source ReShare platform has separated from the ReShare community and rebranded itself as OpenRS (the ‘Open Resource Sharing Consortium’). Several consortia previously affiliated with the ReShare initiative have now thrown in with OpenRS, along with K-Int which has committed to developing and maintaining the software, and EBSCO which will be providing some support for the project. OpenRS was swiftly folded into the portfolio of open-source initiatives managed within the Open Library Foundation, which also provides some organizational support for FOLIO. ReShare joined OLF in 2019 and was identified as an official project as recently as February, but is no longer on the roster of member initiatives, apparently supplanted by OpenRS. (This week ReShare announced that the Boston Library Consortium – an early implementer of the ReShare platform – will serve as its fiscal sponsor going forward.)

The open-source library management space is becoming more crowded, part of a more overtly competitive agora. Open source offerings – in the ILS/LSP category certainly, but also for resource sharing, repository management, and other core library operations – compete against proprietary solutions in what is already a highly consolidated market, and increasingly against other open source providers. As Marshall noted in his opening remarks at the Executive Perspectives panel, the FOLIO initiative itself comprises several different competing interests: FOLIO hosted/supported by EBSCO is different from FOLIO hosted/supported by ByWater, which is different from FOLIO hosted/supported by Index Data, all of which are different from FOLIO when it is self-hosted. Even if much of the underlying software code is shared, the customer experience will vary widely depending on the mix of services (and service level agreements) that are supported by different hosting providers.

This is not a bad thing, of course. But like the splintering of ReShare into separate offerings with different hosting providers and development partners, it introduces new complexity and additional coordination costs for libraries. Interestingly, Lyrasis recently announced that it would cease hosting Islandora while maintaining hosting services for DSpace, an alternative (i.e. competing) open-source repository solution. In early October, ByWater Solutions, a leading open-source hosting/service provider in the US, announced that it’s acquiring shares in PTFS Europe, which offers similar services in the European market.

These patterns simply confirm that competitive interests and market dynamics – including vendor consolidation – are equally at play in the open-source and proprietary software environments. Marshall’s panel provided a welcome opportunity to engage directly with the library community to explore multiple paths to Open library infrastructure without any presumption about the intrinsic superiority or purity of different approaches. I was glad to represent OCLC’s position in a frank and open dialogue about the benefits and tradeoffs of the choices we are making.