Cloud computing provides access to software applications, digital storage, and other technical resources through services that are usually accessed with only a web browser and a connection to the internet. I've covered this family of technologies in many books, articles, and essays that provide more in-depth treatment of the technical details and the specific library products and services in this realm. In broad terms, it comes as the next phase in the evolution of computing following an era that depended more on locally housed servers accessed through software installed on desktop and laptop computers. This change from local to cloud computing has interesting implications for many aspects of a library's technology strategy.
The increased involvement in cloud computing enables libraries to focus more on their core areas of expertise and reduce time spent on routine technical activities, now available as commodity services. In these times when libraries seem to be busier than ever, it is important to make judgments on how to allocate resources, especially those involving technical infrastructure as well as the time and talents of personnel. By moving away from the operation of local servers and relying more on SaaS, libraries effectively have the opportunity to shift resources away from the plumbing of technical infrastructure toward higher-level activities with more immediate impact on their users. The adoption of technology services delivered through some flavor of external hosting or cloud computing brings a variety of changes in the shape of its infrastructure, including both technical and human resource components. This edition of The Systems Librarian explores some of the areas libraries might want to reassess as they move to the cloud.
The Increasing Burden of Local Computing
The responsible operation of servers and network components requires higher levels of expertise. It involves many layers of attention. Hardware components must be configured in ways that not only perform optimally, but that can withstand inevitable failures, which is generally accomplished through redundant components and clustering configurations. A well-designed hardware configuration can survive component failures with no data loss and with negligible diminishment in performance. Libraries that rely on local servers often do not have the resources to deploy redundant components making them more susceptible to failures, compared to the data centers of service providers that routinely implement multiple layers of redundancy.
Security remains more of a concern than ever. Servers must be defended through multiple layers of firewalls, hardened to withstand the ever-present threats of distributed denial of service (DDoS) attacks and malware intrusions designed to gain access to any sensitive information, among a host of other concerns. Attending to security concerns can be a full-time job in itself, and few libraries have the resources to proactively defend their local systems against the full range of possible attacks and intrusions.
The administration of operating systems and database management systems likewise requires constant vigilance. Proper tuning and optimization of OS parameters relative to the database management systems and application software can make an enormous difference in system performance. Only a small percentage of libraries have adequate technical personnel to manage all facets of their ILSs and other critical applications. Moving to cloud computing can allow libraries to step away from many of the complexities of hardware and software management that large-scale data centers are better equipped to manage.
Design and Control
A move to cloud computing does not mean ceding control of a library's strategic technology. Regardless of whether software applications operate on local servers or through externally provided services, a library needs to retain control of the high-level design of its technical environment to ensure that it supports its operational and strategic activities. This control comes in the form of making good choices in regard to which of the competing products to implement, in organizing the data flow among applications, and especially in designing an excellent user experience for the library's patrons.
The Role of the Technologist
libraries that do not have in-house personnel with deep technology skills will naturally find systems based on cloud computing more accessible. Not having to deal with local hardware and other complications makes web-based applications a good fit for libraries that do not have the luxury of dedicated technical staff. These libraries are often frustrated with locally installed systems where they may have to carry out complex routines such as daily backups, software upgrades, system restarts, or other tasks. They can take advantage of automation systems delivered through SaaS with almost the same ease as they have subscribed to web-based content services. Without the aid of local systems or technical staff, a library can work with the provider of the service for the initial configuration parameters, local branding, data migration, or other setup procedures. An essential component of a cloud-based service includes ongoing support that provides prompt assistance with any questions in the operation of the system, interruptions in the service, or any technical problems.
A shift to cloud computing does not necessarily mean that libraries do not benefit from having a systems librarian or other technology support staff. An increased reliance on applications delivered through SaaS and other flavors of cloud computing allows library technologists to work higher in the stack of hardware and software components comprising an organization's technical infrastructure. Even when an application is delivered through SaaS, it may require considerable attention relating to local configuration and customization, data management, or other tasks to ensure that it works optimally for the library.
That said, libraries that face the need to trim personnel budgets will see some aspects of technical support that are diminished as they move to externally provided services. Many routine tasks related to server and workstation maintenance will be obviated, which may reduce some areas of technical support the library needs to provide. The basic financial model of cloud computing involves replacing a cluster of aggregate costs that surround locally provided resources with a set subscription fee for a comprehensive service.
The work of the library's technical staff may be diminished in the tasks related to the installation and updates to the software on the library's computer workstations. Especially with complex applications such as an ILS with workstation-based client software, installation and upgrades can be very time-consuming for libraries that have dozens or even hundreds of staff computers. Web-based systems provide a welcome relief to these kinds of tedious tasks.
But I would suggest, even when the automation system and other key business applications are hosted externally, that libraries continue to benefit from the expertise of their technologists. These personnel can become involved in more interesting and important activities that allow them to tap their technical skills and creativity to help library personnel work more effectively and efficiently and to improve services to patrons. Many libraries have a web presence that is underdeveloped in its capabilities to deliver access to collections and related services, and it has a less than optimal user experience. The library's website often receives too little attention, with technical personnel consumed with other behind-the-scenes activities. By moving to cloud computing, libraries can redirect efforts to areas of more direct concern.
Many of the new-generation cloudbased products provide a set of APIs that enable the library to extend its functionality. By writing a small application or script that takes advantage of these APIs, a library programmer might be able to create tailored services to fulfill some need not provided by the base functionality of the system. Engagement with the APIs of each of the systems and services can be a great opportunity for library technologists to make extensions and improvements for the benefit of the library.
Connectivity Is King
When libraries become increasingly reliant on cloud computing, network connectivity rises above all else as critical technical infrastructure. Areas of concern include the total amount of bandwidth available and the reliability of connectivity. The inherent design of cloud-based technologies provides almost no operational backup capability for any episodes when the system is down or inaccessible. In my experience, problems happen much more frequently due to connectivity issues between the library or its users and the service provider than with the service actually being down. Many service providers maintain multiple hosting locations, each with redundant internet connectivity options.
Libraries likewise should deploy the strongest internet connectivity strategy that they can afford. The move to cloud computing will decrease spending for local computer hardware, but some of the savings should be allocated toward connectivity. When both access to content subscriptions and automation services are delivered through the internet, this component of technical infrastructure deserves special attention.
Many libraries (especially those in educational institutions, government agencies, or corporations) are able to take advantage of internet connectivity provided through their parent institutions. In these scenarios, the library will need to take a proactive role to ensure that the connectivity provided meets crurent and anticipated future needs and intervene accordingly should service levels fall below expectations. Libraries that must provide their own connectivity face additional challenges. They must work more directly with ISPs to subscribe to the optimal level of bandwidth to meet increasing requirements. One of the major impediments to cloud computing lies in rural areas where there are limited opportunities for having high-performance bandwidth. Internet connectivity options need to be re-evaluated every year or so given the increasing demands for bandwidth and the occasional opportunities to take advantage of new connectivity options which may offer more bandwidth at a lower price. Given these dynamics, it makes sense to avoid any long-term contracts that might lock the library into service plans despite the availability of more favorable alternatives.
A Good Match for Libraries
Given this current phase of technology, we find ourselves in a period of transition toward cloud computing. I anticipate that many libraries will continue to rely on at least some locally managed system for quite some time, but in broad terms, most library - specific and general-purpose cloud applications will see wide adoption. Most of the newer automation systems such as library services platforms and webscale discovery services have been designed to be deployed through SaaS. As libraries continue to rely on serverbased ILSs, an increasing proportion will choose a vendor-hosted configuration. Even when they're not ready to change systems, I observe that many libraries choose to move to a venderhosted configuration when their local servers are ready for replacement.
In many ways I think that cloud computing is well-suited for libraries. It's a model of technology that allows them to focus more on their core strengths, such as managing multifaceted collections, organizing information, and providing a rich variety of service to their users. I think that technology strengthens the ability for the library to perform its core services. Among the many approaches to technology, cloud computing fits nicely into the library realm when it delivers appropriate functionality with a lower overhead of equipment and personnel resources. *