What is Cloud Computing
Image by MichaelMarlatt via Flickr

With the current economic climate, most financial news coming out of the academy is terrible.  In the small segment where I personally have some involvement — scholarly communications — all the talk is of cuts, cuts, cuts.

Libraries and their representative consortia and professional associations have circulated strongly worded memoranda about the nature of what is being called a crisis, putting their vendors, principally publishers, on notice that some subscriptions will be canceled, no price increases will be tolerated, and new products are to be avoided.

I find this last point the most troubling. No new products?  But that’s what’s being said.

It’s difficult for anyone to find anything heartening in all this.

While no one can make all the problems go away, we can mitigate some of them with careful analysis and a shrewd sense of how we spend our money.  In the increasingly digital world of communications, new services arise all the time, and not only in the academic community.  Some of these services may provide a means to reduce costs in the academy without compromising services.

I have in mind the (mostly) consumer service Scribd, with which readers of the Scholarly Kitchen are already familiar.  Scribd is a cloud-computing alternative to institutional repositories (IRs).  By using Scribd, libraries can shut down their own institutional repositories, saving money and losing no services.

Scribd is best thought of as “a hard drive in the cloud.”  It’s a form of storage with a Web interface, and it resides in the Internet cloud rather than on anyone’s personal computer.  Scribd is often called “the YouTube of documents,” a good description, as it allows anyone to upload documents at no cost and makes them available (subject to the author/uploader’s stipulations) for free.  That is, Scribd is a free Open Access service, a repository waiting to shoulder some of the growing costs of a library’s own information management.

It is not known to me how much IRs cost libraries, but the number cannot be small, as the management of an IR typically requires some involvement from IT staff and related activity such as the attendance of conferences and trade shows.  Most of the cost is in personnel, as hardware and bandwidth are relatively inexpensive and much IR software is open source (e.g., ePrints and DSpace).

But whatever the cost, it is now an unnecessary cost.

If a service can be provided for less money or no money outside an institution, then working with that outside service is in the institution’s interest. Doing so will free up resources to pursue other aspects of an institution’s mission.

It will be argued that IRs provide services that Scribd does not.  I’m sure that is true, but the question is how long will that remain true and whether Scribd can be improved for academic users.  For example, some IRs stipulate that every document must be accompanied by extensive metadata.  To get that same level of metadata into Scribd simply requires the creation of an online template or checklist to guide authors on what to include with their documents.  To create such a template is a one-time cost, while the management cost of an IR is perpetual.

Let’s imagine an alternative way to pursue the much-publicized announcement that Harvard is going to create an IR for the scholarly articles created by the Harvard faculty. Now, the point of that announcement was not to create an IR but to make the scholarly output freely available to anyone.  With Scribd this can be done, and Harvard can still effect its policy but at a much lower cost.  Instead of instructing the faculty on how to deposit articles into Harvard’s own IR, the faculty would be instructed on how to deposit (and tag) articles for Scribd.  These articles would still be open access, and they would still be discoverable through their metadata and the full-text search of Google and other search engines.  In working with Scribd, Harvard would save money and lose nothing in terms of services.

IRs are part of a backward-looking vision of information technology, a variant of the enterprise-computing model, where it’s believed to be necessary to “put everything in one place.”  On the Internet, however, an institution — whether Harvard, Princeton, or Cabrillo Community College — is represented by nothing more than a tag. It’s not relevant where the data resides; insisting that Harvard or any other institution host its own services is akin to setting up a retirement account in which you plan to stuff all your money under a mattress.

In order to fulfill their mission, libraries should begin to close their IRs.  The money saved can then be put toward other things.

Perhaps libraries will even use the money to buy new books.

Reblog this post [with Zemanta]
Joseph Esposito

Joseph Esposito

Joe Esposito is a management consultant for the publishing and digital services industries. Joe focuses on organizational strategy and new business development. He is active in both the for-profit and not-for-profit areas.


12 Thoughts on "How Libraries Can Find Money in Clouds"

If I were setting up an institutional repository, I’d be hesitant to use a for-profit platform with an unproven business model in a nascent and highly competitive field. It makes you dependent upon their success for continued access to your archive. The whole concept of cloud computing is one that’s worrisome, at least from an archival point of view. The complete meltdown of Ma.gnolia should serve as a cautionary example.

If one were to take this path, the smart way to go would be to use multiple, redundant services. And, to constantly make backups of the entire archive that are stored somewhere else on the cloud and in hard copy as well. This complicates the process somewhat, as you’ll now have to support multiple platforms. And if one of those companies fails, all the work you’ve done customizing your archive for their particular service has been wasted.

Joe, Lots of interesting points, but your call to close IRs and switch to Scribd falls down on your main example. Harvard faculty has an open access mandate. It could not mandate deposit in Scribd. The objective of OA mandates is to approach 100% open access within faculty and institutions. That could not be achieved through Scribd.

I have frequently wondered how libraries can simultaneously complain about serials costs/budget cuts while spending large sums developing repositories. My understanding, however, is that this often involves a different department with a different budget. If so, it wouldn’t be the first example of this sort of apparent financial doublethink in a university – or any other organization for that matter.

Joe is also right that the ‘all in one place’ solution is backward looking. But I suspect there is an additional motivation here: control/possession of the information and the fear that ‘clouds’ are rather transient in nature…

Whatever are the arguments , print resource support to teachers and students is vital and we must find out ways to overcome current problems.

Institutional repositories represent a commitment to preservation, as David Crotty suggests. Scribd may be an interesting and useful access and dissemination technology, but I don’t see how the preservation objectives of repositories could be assured with it.

Also, repositories aren’t necessarily associated with an “all in one place” credo, as multiple copies of works often exist with the publisher, the LOCKSS (Lots of Copies Keep Stuff Safe) program, etc., in addition to the repository copy.

Another lesson in the transience of popular internet sites can be gained from watching the current disappearance of Myspace. For archives that are supposed to have some permanence, jumping on a trendy and currently successful business is a risky move.


If universities were to prove foolish enough to scrap their own Institutional Repositories, renouncing their efforts to reclaim custody of their own research assets at long last, to heed instead the siren call to entrust them yet again to 3rd parties — and commercial ones like scribd, to boot — then, frankly, they are unsalvageable and deserve everything that’s coming to them.

I don’t for a minute, however, believe that the Academy would fall for this, having been once bitten, now twice shy, any more than they are falling for the concerted bid by some publishers to “leave the open-access archiving to us!” http://bit.ly/bYjB3

Rather, this is a highly anomalous and dysfunctional era of academic “outsourcing” that is happily nearing its well-deserved end.

I’m afraid Professor Harnad is once again simply uninformed and incurious. This is not a matter of outsourcing vs. “build your own.” Many academic IRs are already outsourced. Professor Harnad should take a look at the client list at http://bepress.com, which markets the Digital Commons service. This is a “white label” outsourced (SAAS or Software as a Service) solution. Among the many clients of BEPress is the University of California, a leader in open access publishing. Let’s focus on the real problem, which is the high cost of maintaining services in house that would otherwise be free.

It seems, though, that several commenters are making the suggestion that these services are not “free”, but rather may contain their own hidden costs. Digital Commons may very well be such a “white label” solution because it contains preservation methods and non-exclusive licensing agreements similar to those that would be established by libraries, if libraries were setting up their own local repository.

In other words, how do you, as a responsible archivist, ensure your document will exist 100 years from now? 50? 10? You’d need assurances about data back-up, metadata maintenance, format conversion, and more. Does Scribd provide those services? Would anyone be able to provide all that and still be free?

To be honest, I’m not sure if Scribd addresses these issues. I haven’t had time to look through the user agreement — maybe it does. But I believe we need to address those details if we’re recommending Scribd as a legitimate alternative.

It’s not really about who does the work, so much as who makes the rules.

It’s the difference between hiring a company to promote and distribute your bands music, verses signing your rights over to a record label and being at their mercy.

What is important is where the power lies, and who owns the data. Commercial companies can’t and shouldn’t be trusted with something so important. They’ve already tried other iffy tactics: http://www.wired.com/wiredscience/2007/09/yesterday-new-s/

Comments are closed.