This post was originally posted to the Collections Trust blog in October 2011.
I want you to imagine a scenario with me. Picture your museum. Now imagine it with no servers, no in-house IT team, no Collections data onsite at all. Imagine that all of the software you use to manage your Collections is accessible through your browser, with your data held in a secure, stable server farm somewhere far, far away.
How do you feel?
The idea of applying the principles of Cloud Computing and Software as a Service (SaaS) to Collections Management is not new. Indeed, in some ways it is much older than most people think – harking back as it does to the very early days of using remote terminals to access processing power provided centrally on a mainframe. Back in the early 1980’s, museums all over Europe would ship their hard-copy record cards to the MDA Computing Bureau to be transcibed into electronic records, again, making use of the economies of centralised computing power.
Fast forward 30 years, and the scale, complexity and richness of electronic recording of information about museum artefacts have expanded exponentially. Documentation and cataloguing are a global business, supported by a thriving community of highly specialised software applications and standards such as SPECTRUM.
The prevailing model of computing in museums is (as it is in most other types of enterprise) to have a number of relatively highly-powered machines connected to a Local Area Network which acts both as a conduit to the Internet, a firewall, a communications layer and a shared filestore. Until relatively recently, most Collections Management Systems were engineered to operate in this environment – either installed across the network or on a dedicated internal server.
This model had its advantages. A LAN could include multiple physical locations, enabling, for example, curators onsite and conservators in offsite stores to interact with a common dataset. It empowered the museum to run its own kit, and to establish policies for things like backup and disaster recovery. In essence, it put the information about the collections inside the same physical context as those collections.
But the networked-application model also has some profound drawbacks. It is inherently inefficient, requiring expensive user support and onsite technicians. It requires physical space appropriate to the operation of high-powered and energy-hungry machines. It places a barrier to the upgrade path for the software, requiring manual intervention, downtime and occasionally bespoke development. The knock-on effect of this is that innovations funded by one user of the software are seldom cascaded automatically across the entire user community (or where they are, the time-delay involved can be considerable).
Collections Management Software as a Service, in which Collections data is held on low-cost, scalable offsite storage and manipulated using browser-based interfaces, appears to hold the solution to many of these issues.
It makes sense from the developers point of view – hugely reducing the complexity and therefore cost of maintaining 2000-3000 separate installations of their software. It provides a relatively painless path through which upgrades and new features can be cascaded out across the entire user-base of a given system. Not only this, but it offers a development path for new features and functionality – perhaps in the form of integration with tools for Digital Asset and Digital Rights Management, Workflow Management, Visualisation or Digital Preservation.
It also makes sense from a managers point of view – onsite IT infrastructure (and the skills to maintain it) can be very expensive. The space required for dedicated onsite kit can be given over to storage, administration or other hard-pressed functions. CMSaaS can be cheaper, offering the flexibility of subscription-based models which scale with the content and its uses, and in the process, offering a means of futureproofing against the evolution of the museum and its information management needs.
It makes sense, too, from a professional point of view. A community of clients using a common Cloud-based Collections Management System forms, almost by definition, a community of practice. It leads to the sharing of knowledge and expertise, to mutual support and other forms of collaboration. Not only this, but it naturally tends towards the principle of openness with museum datasets, encouraging the more proactive use of data through API and 3rd party applications and websites.
For all of these reasons, the majority of the leading Collections Management Systems have either already developed fully-hosted versions of their applications or are in the process of developing them. And business, it seems, is booming – at least partly in response to the strictures imposed on museums by funding cuts. Most vendors are reporting significant increases in the uptake of their Software as a Service offers, with some confidently predicting a full transition to Cloud-based, browser-based Collections Management within the next decade.
And yet there are many for whom the idea of Cloud-based Collections is unpalatable, to say the least. Some point to the inherent instability and insecurity of the Cloud as being too much of an operational risk to entrust their Collections data (which is intended to be the canonical and authoritative record of the Collection) to it. Others point to the fragility of their connection to the web – raising the prospect of downtime, lag or failure. Others point to the creativity and innovation which comes from having technologists, technology, curators and data under one roof. Others, too, suggest that this places them in a position of total dependency on their vendors and software providers, effectively rendering them hostage to the vendors charging model. It is also suggested that each and every installation of a Collections Management System represents a bespoke tool, specific to the needs of that museum and that SaaS would fail to deliver the kind of local configurability that museums and their Collections demand.
It is too early to say whether all of these concerns are equally valid. Certainly many museums run antiquated hardware and suffer from poor Internet connections – although people suffering with a bad PC should in theory benefit from the lighter processor load of using browser-based applications. Many are also governed under a Local Authority IT Policy which stipulates requirements that make CMSaaS all but impossible. It is also fair to say that many Browser-based applications are not quite as responsive as their locally-installed counterparts (although the gap is closing).
As to the risk inherent in Cloud storage, I suspect this may be based on an overestimation of the stability and security of locally-based physical media. After all, the data is still bits encoded on a spinning disk (or solid-state drive) whether it is under your stairs or sitting in racks in a data warehouse somewhere. I also suspect that the inexorable and inevitable rise of mainstream consumer-focussed applications which run in the Cloud will gradually help people to overcome this concern.
As to the creativity and innovation that comes from having museum technologists on the team – this has doubtless led to the flowering of ideas and applications, but it may simply not be sustainable in the face of the economic realities confronting museums in most countries worldwide. This is, of course, a gross simplification, but we may come to see a relatively small number of museums (outside of the larger institutions) that are able to maintain this kind of staffing and capital overhead. Many museum technologists – disaffected in some cases by the inability of the museum to enable them to deliver what they know they are capable of, have themsevles defected to commercial vendors and development houses.
In reality, we are unlikely to see a total transition to Collections Management in the Cloud anytime soon. This is a tremendously diverse sector, with a huge range of different types and scales of institution. There will always be those who need an installer and a locally-based application, for whom this is the simplest and most effective option. What I do think we’ll see, particularly over the next 4-5 year cycle, is two significant trends:
1. A significant swing in the medium-to-larger end of the museum community towards full-service online Collections Management Software and;
2. The emergence of more online tools with a very low barrier to entry (both in terms of cost and complexity) which are suited to the needs (and budgets) of smaller local and community museums.
Is it time for Collections in the Cloud? I’d love to know what you think!