As libraries strive to increase the impact of their collections and services to their communities, large-scale implementations of resource management systems and discovery environments provide a great opportunity to do so. A trend is already underway in which many libraries are shifting from individual implementations of ILSs to collectively shared infrastructure. This cooperative approach to technology resonates with the mindset of libraries, which are generally interested in supplementing their own local collections with resources available from partner institutions.
No library has the resources to directly acquire all of the content needed to fully serve its clientele. ILL and other resourcesharing arrangements are a well-established component of the broader library community. Moving toward shared infrastructure brings at least an initial level of resource-sharing into the core systems supporting the library. Participating in a shared environment can also be more economical relative to every individual library maintaining separate systems.
The current era of cloud computing provides opportunities for libraries and related organizations to implement technology infrastructure at massive scale, with the capability of increasing the impact of collections and services while reducing costs. In earlier phases of library technology development, limitations in computer ware, connectivity, and software architectures restricted the scale of system implementations in terms of the libraries served, size of collections supported, and volume of transactions possible. Current cloud-based infrastructure scales beyond almost any imaginable aggregation of library services.
Libraries strive for higher impact, efficiency, and economy in their technologybased systems. Impact, in this context, means the quantity and quality of materials available to the clientele of the library, with the best ease of access. An environment in which only the local library's own materials fall within the initial scope of search will have less of an impact than alternative arrangements that include materials from a large set of relevant libraries and that fulfill access to remote materials with the same ease as those held locally.
One of the basic models of shared infrastructure involves multiple institutions sharing an instance of a library services platform or an ILS. Although many different deployment options are possible, a common approach is a shared bibliographic environment, with data structures identifying the holdings and locations of each of the participating libraries. This arrangement enables patrons to easily search the aggregate collection, with appropriate access options presented for items available in a location that's local to the user or to request remote items for fulfillment. Such a shared system provides broad resource-sharing capabilities by taking advantage of request and routing features inherent in circulation functionality rather than the more complex intersystem interactions involved in an ILL or resource-sharing service.
Providing a larger universe of content behind the main search box presented by the library can benefit patrons, even if some items are not available on-site. Many libraries already manage fulfillment options for materials housed in off-site storage or in distant branch locations. Multi-location libraries already manage logistical infrastructure for expedited fulfillment, which can be extended for support of larger partnerships brought together by shared collection management infrastructure.
Shared infrastructure provides a larger core of aggregate resources available to users by default and is naturally further extended to other ILL or resource-sharing services. In most cases, the portion of resources fulfilled within the shared system will be dramatically higher, reducing the number of requests directed to fulfillment channels that may be significantly more expensive with longer fulfillment times.
Libraries will usually see savings when they pay into a shared system rather than fully fund their own dedicated implementation. Implementing a dedicated system is the most expensive deployment model, especially when counting the local technical and administrative personnel costs, as well as the costs of software licenses or computing resources. It's unfortunate that libraries with the least resources often end up with the most expensive and least impactful automation models.
Participation in shared systems involves trade-offs that may be uncomfortable or, in some cases, prohibitive. Implementations dedicated to a single institution mean full control of policies, preferences, and configuration options. Dedicated systems also enable locally curated bibliographic data. Those libraries with specific needs that require a dedicated resource management system gain fuller local control, but it comes at a higher implementation cost and presents lost opportunities for collaborative impact. Implementations of systems for a single library organization remain common and are not likely to disappear, although I anticipate that the numbers will diminish over time.
Shared infrastructure for resource management is an established trend. Many different types of library groups have implemented shared collaborative infrastructure. Some of the options include consortia arrangements, multicampus library systems, and statewide or national systems.
Consortia, which involve multiple legally independent libraries partnering to share a system, engage in collective materials procurement (or other collaborative processes) and are not new. Implementations of shared systems serving consortia have been around since the earliest years of library automation. Recently, there has been a trend toward larger consortia-shared implementations. Examples include the Illinois Heartland consortium, serving 476 library locations, and the London Libraries Consortium, which is shared by more than 300 library facilities in the greater London area.
Multi-campus library systems are useful. Many educational systems with multiple campuses have opted to implement a single comprehensive system rather than separate deployments for each campus. A striking example can be seen in California, in which each of the public educational systems has separately opted to move from individual campus to shared systems:
- University of California system: 10 campuses, 100 libraries, 52 million collection volumes, and a student population of more than 250,000
- California State University system: 23 campuses, 25 million collection volumes, and a student population of 460,000 students
- California Community Colleges: considered the largest higher educational system in the world, with 110 campuses, 2.1 million students, and 50,000 faculty members
- The State University of New York: 61 campuses and 470,000 students
Statewide or national systems serve all of the libraries within a broad jurisdiction. Some of the U.S. states with smaller populations have long-standing statewide systems for their public libraries, including Hawaii, Delaware, and Rhode Island. A few countries have nationwide systems for their public or school libraries. Denmark, for example, recently deployed a shared system for the 2,428 public and school libraries serving all 99 of its municipalities.
These are just a few examples of large-scale implementations deployed in recent years that displaced multiple independent implementations.
Especially in the context of a consortium, partnering with other institutions can go beyond any given institutional context. Libraries from diverse organizations can opt to cooperate to share systems and services. Many existing consortia include both public and privately funded organizations, with libraries ranging from small to very large. The mechanisms of governance- including processes for decision making with regard to cost distribution-may be challenging, but the issues are solvable. The parent organizations funding libraries are usually open to cooperative partnerships-for students, research funding, or other strategic interests- that result in savings even if it means partnering with other institutions with which they otherwise compete.
Consolidated discovery without consolidated resource management may not be economically or technologically feasible. A trend in previous times was to implement a shared discovery layer or catalog that addressed multiple libraries, each with its own ILS. This model was pragmatic when the scalability of ILSs was limited. Many examples of this approach remain, although they are steadily being replaced by shared infrastructure, including both back-end collection management and patron-facing discovery. Shared systems are a more elegant approach, in which collection materials are managed and indexed centrally, avoiding the need for complex and fragile interactions between a frontend interface and multiple independent ILS implementations.
The concept of shared infrastructure can be applied in contexts other than core library management systems. Many libraries also administer other content management platforms, such as institutional repositories. A large portion of university libraries operate their own standalone institutional repositories. Many are based on open source products (such as DSpace, Fedora, or Hyku) with fairly low operational costs. The fragmented deployment scenarios involved in the standalone implementation of institutional repositories may constrain their impact. Shared regional or national repositories are well within the scale of technology. A shared platform, segmented by institutions, would not necessarily appear that different from the current fragmented deployments, but would offer new possibilities for discovery, better options for digital preservation, and lower technical and administrative overhead. Current scenarios for the discovery of institutional copies of preprints or published scholarly materials mean great complexity for services that identify OA materials when publisher copies may not be available to the researcher.
As libraries look toward to the future, it is important to consider opportunities to participate in cooperative technology systems. When the time comes to replace existing systems, the path of least resistance will be to replace one standalone implementation with another. Although this approach may result in more modern technology and functionality, it may do little to increase impact. New systems and increased collection spending may produce only incremental improvements in collection access and impact. In contrast, moving from a standalone to a shared deployment model can result in a substantive increase in the resources provided, often expanding the default collection many times. The scale of technology no longer limits the possibilities for large-scale cooperation among libraries. Organizational or political concerns present the most challenging obstacles.
I anticipate an acceleration of the trend toward shared infrastructure as libraries seek to increase collaboration with their peers and partners, amplify the collection resources available to their users, and reduce technology overhead. Opportunities arise not only when libraries replace legacy systems, but also as they consolidate recently acquired deployments. Cooperation and consolidation in library system deployments can be seen as a natural response to the increased pressures on library budgets and expectations to continually improve services. Although not a panacea, massively shared technology systems are an important strategic option for libraries to consider.