Weighing the Cloud: OpEx versus CapEx
Cloud vendors tout the benefits of paying for cloud services as operating expenses, rather than raising capital funds for infrastructure projects on campus. But the calculus is not as simple as that.
Many considerations inform a university's decision to move services to the cloud. But, in these lean economic times, budgetary factors are always going to play a major role. One of the key benefits touted by cloud vendors is the ability of institutions to categorize their services as an operating expense (OpEx). If colleges and universities build their own infrastructure (like a data center), on the other hand, the cost is considered a capital expense (CapEx). Money is money, though, so why should it matter which bucket it comes from?
From an accounting standpoint, the biggest advantage of the OpEx approach is probably flexibility. Operating expenses are usually paid monthly and can be adjusted more easily in a volatile budget environment. In general, CFOs have a lot more leeway in how they disburse operating expenses--a freedom that often does not extend to capital expenses.
Indeed, capital expenses often come heavily encumbered. When a university decides to invest in a major infrastructure project, such as buildings or equipment, the CFO must often seek financing. In addition, the whole project may be required to pass through a series of committees and approval processes, all of which take time. And even after the infrastructure is built, the school is still on the hook for operating expenses.
Yet, the decision to transition to the cloud is not a slam dunk, by any means. While the calculus may be easier for businesses, higher education institutions receive their funding from a host of sources, including state and federal governments, municipal bonds, and grants. Usually, these funds are earmarked for specific uses or items, and are often categorized as capital expenses. Few CFOs are going to leave these funds on the table.
The other compelling argument behind outsourcing involves obsolescence. Technology is changing so quickly that capital projects are in danger of being outdated almost before they're brought on line. And if technological innovation outpaces the rate of depreciation or amortization, a university can be left paying off infrastructure that is obsolete.
And the speed of innovation is truly dizzying right now. Moore's Law states that the number of transistors that can be placed on a chip doubles every two years. While this rate of growth is expected to slow over the next decade (doubling only every three years), it still illustrates how quickly the tech landscape is changing.
Given all this, the argument for transitioning services to the cloud makes a lot of sense. But it would be a mistake to consider major infrastructure and service decisions through such a narrow lens.
Indeed, universities need to step back from such arguments and take a more holistic view of the situation: "What do we have in place already? What do we do well here? What will simplify our lives by letting someone else do it? What are the greatest needs, assets, and strengths of our community?
It is highly unlikely that an all-or-nothing approach will suit any institution. Instead, colleges and universities need to strike the right balance between providing and procuring, building and buying, capital expenditures and operating costs.
It's an approach that works well for North Carolina State University. Before any technology acquisition or transition, says Marc Hoit, vice chancellor for information technology and CIO, the university looks at the set of requirements; explores the options for building, buying, hosting, or using the cloud; and evaluates the relative costs, pros and cons, and strategic directions. A decision is made using an established governance model and taking into account input from stakeholders.
For NCSU, the shift from CapEx to OpEx for certain services, like Google Apps, was an obvious choice because the value was clearly there. "The OpEx costs were small--in many cases, smaller than the OpEx of maintaining our own infrastructure," notes Hoit.
And by shifting to the cloud, the school was able to free up resources--specifically IT staff. "Being able to shift capable individuals to middleware development from hardware and application maintenance was critical," continues Hoit. "We also needed to shift more of the effort to our support services."
But NCSU isn't putting all of its eggs in the OpEx basket. "We continue to hold some of our CapEx as 'end of the year' funds," explains Hoit. "This provides flexibility for the CFO, but it also means a risk if funds are not available. We have had to postpone acquisitions for a year. However, we have always had funds for critical infrastructure."
NCSU has run own internal cloud system since 2004, which has provided "a huge benefit, very great cost savings and a very good learning environment for cloud services," says Hoit.
But perhaps pondering whether to go to the cloud or not is the wrong question altogether.
"'Cloud' is a terrible term for assessing the correct economic questions regarding efficiencies around demand aggregation," says Brad Wheeler, vice president for information technology and CIO at Indiana University. Wheeler and Shelton Waggener, associate vice chancellor for information technology and CIO at University of California, Berkeley, co-authored a paper that outlines three models for aggregating IT services: commercial sourcing, institutional sourcing, and consortium sourcing.
In their view, consortium sourcing is an ideal solution in academia where collaboration and communication have a long-standing tradition. And members of a consortium are still free to choose which IT to aggregate.
Utilizing this model, it made sense for IU to build and capitalize a hardened data center. "We are a 'cloud' provider to eight campuses, part of the state of Indiana, IU Health, and now several other higher ed institutions in the state," says Wheeler. "Still, we source some things with others (student e-mail to Google or Microsoft) when they make sense. We did an on-premises deal with Penguin On Demand for High Performance Computing needs that spike beyond our own or NSF HPC systems."
There is certainly no one-size-fits-all solution for institutions. And what works for public universities may not apply to private schools. "As a public institution, our funding cycle means we need to have flexibility to shift between CapEx and OpEx for some portion of our portfolio," explains Hoit. "The institution appreciates having this flexibility and has rewarded it, provided I can provide a strong business case for the need."