Managing technical infrastructure has been a mainstay of library systems departments. Back in the day, the ILS and other critical library applications were usually implemented within the library itself and managed by its own personnel, including systems librarians and systems administrators. The heyday of the ILS meant tremendous responsibility for libraries to implement and manage almost all aspects of these complex applications, such as hardware installation, network support, systems administration, data migration, configuration, and customization. Such arrangements gave libraries great control of their technical systems, but likewise required significant investments in financial and personnel resources.
In recent times, the management of library systems has changed dramatically. These changes have come through successive cycles of organizational and technical paradigm shifts. As a result of each new phase of computing architecture and deployment models, libraries have needed to continually rethink the kind of work performed by their technology support departments.
Reacting to Enterprise Computing
One of the early phases of change transpired through the centralization of computing in educational and governmental institutions. As these institutions began to develop comprehensive enterprise networks and data centers, maintaining systems throughout different departments became increasingly problematic. Many categories of systems could be deployed at the institutional level in ways that were less expensive, more secure, and more sustainable than through decentralized installations. Many university departments, including libraries, managed their own email systems, file servers, and other components to support their computing needs.
The age of enterprise computing that came in full force by the early 2000s brought a major reconfiguration in the technical environment of most academic campuses and municipal or county governments. General commodity systems were the first to be centralized. The technical complexity of operating mail servers, for example, became increasingly onerous as these systems needed ever more complex defense from malware attacks and general security concerns. Institutions could better mitigate their risks and control costs through a centralized system with industrial-strength defenses rather than through multiple departmentally managed services, each deployed with limited resources and uneven technical expertise.
The shift from departmental to enterprise computing saw little resistance, at least in the realm of commodity services. Libraries and other university departments were happy to take advantage of institutionally provided services rather than expend their own resources and talents. Many of the departmental computer rooms and mini data centers were decommissioned in favor of largerscale institutional data centers. These data centers could be equipped with state-of-the-art firewalls and other security components, proactive systems monitoring, industrial cooling, and fire safety-all managed by professional engineers and operators. The larger scale of these data centers usually meant overall lower costs, which were usually absorbed centrally and not charged to individual departments.
This shift to enterprise computing did not necessarily mean the immediate demise of library computing. Specialized applications such as ILSs were often given particular consideration since they required specific knowledge of library processes and involved very complex configuration and operational procedures.
In the context of enterprise computing, management of library systems has taken a variety of forms. In some cases, the servers may be physically housed in an institutional data center. It's common to see shared approaches for their management, in which hardware, network, and OSs fall under the responsibility of the institutional IT department with the configuration and operation of the application software handled by technology personnel in the library. Some institutions have a dedicated individual or team for library applications as part of the institutional IT organization. There are others in which library systems continue to be physically housed and entirely managed within the premises of the library, although such arrangements are increasingly rare.
Moving to Vendor-Hosted Platforms
The trend toward vendor-hosted services for library applications represents an alternate path, with a similar outcome of less responsibility for technical infrastructure falling to library personnel. The idiosyncrasies of library systems often mean that the vendors developing them are better positioned to provide hosting and management for them than an academic or municipal IT department. Library vendors are able to manage many instances of the applications they created for their customers at sufficient scale to gain efficiencies in operations and cost. In most cases, the hosting fees that are charged by vendors will be lower than the local technical infrastructure and personnel costs that would be incurred for local deployments. For vendors, hosting fees bring added revenue and lower support costs. Hosted services mean more consistency and control with fewer problems related to on-premises equipment and networks. Recent years have seen a substantial shift in library systems from local installations to vendor hosting. Almost all new deployments of ILSs rely on vendor hosting.
Adopting SaaS and Cloud-Hosted Solutions
The more recent shift to true SaaS, as seen in library services platforms and discovery services, almost entirely removes responsibility for technical infrastructure away from the library. Since these applications are deployed through global multitenant platforms, their technical underpinnings are almost completely transparent to the library. All the hardware and software that comprise the service are entirely managed by the vendor. The low-level computing hardware may be even more abstracted when based on Amazon Web Services or other public or private cloud infrastructure providers.
Reliance on applications delivered through a pure SaaS model does not necessarily take all the work away from systems librarians or equivalent technical personnel in the library. These applications are complex and require considerable effort related to their configuration and operation. However, this work tends to center on policies, workflows, and metadata rather than on technical infrastructure.
The three trends of centralized enterprise computing, vendor-hosted services, and subscriptions to SaaS-based library services platforms have reshaped the realm of library computing support. Mostly gone are the days in which libraries manage basic infrastructure. Support for systems is focused more on the operation of the application and on patron-facing interfaces and services. While some seasoned systems personnel may miss the control they once had over library applications, most appreciate the opportunities to become more involved in technical work that has a more direct impact on staffers and patrons.
The current trend toward hosted systems and SaaS goes well beyond core systems (such as ILSs, library services platforms, and discovery services). Almost any new technology-based service libraries acquire will likewise follow these implementation models. Libraries still managing locally installed systems should expect change in their next phase of computing. While there may be some scenarios with ongoing requirements for locally installed systems, the trend toward hosted technology services is pervasive.
The movement away from support of computing infrastructure has major implications for the individuals and departments responsible for technology in libraries. The emphasis has moved from administration of infrastructure components to working with higherlevel software and services. The current technology milieu in which cloudbased services prevail brings with it demands for those working with library systems to hone new skill sets. Most of these skills revolve around the mastery of APIs to solve data problems, implement interoperability among related systems, and create new services. Metadata also remains central to most library applications. Involvement with established library standards such as the MARC suite of formats continues with no end in sight. New formats and concepts such as linked data, semantic web technologies, and AI continue to gain traction in the library tech arena. It's increasingly important to find connections not just among library-related systems and applications, but to build interconnections with the broader information ecosystem.
The basic mission of a systems librarian, or other professionals working with technology in libraries, has not fundamentally changed. The key objective revolves around applying technology to enhance the work of the library based on a thorough knowledge of all aspects of library operations and strategy, on awareness of available technologies, and on the technical skills to shape technology around the needs of the library. This era of increasing reliance on cloud computing provides more opportunities than ever before for fulfilling these fundamental goals.