Kenneth Green recently completed the latest annual Campus Computing Project survey, where he noted that there’s still “little movement to the cloud for the really ‘big’ tasks.” He concluded that the reasons include the risks institutions perceive coupled with their lack of trust in cloud providers as well as their own lack of control with the cloud paradigm overall. This post provides some thoughts on these risks and perceptions as well as cost considerations, and also adds some additional color to the preparations institutions could contemplate when considering the cloud.
The first, regarding the perception of cost savings, or lack thereof, has more to do with applying specific domain context in a higher education budget framework, than with generic cross-industry savings claims. Without proper context, it’s difficult to understand if money will actually be saved because of the uniqueness in institution spending versus a typical business.
Consider the statement: “I will save money moving to the cloud because I don’t have to incur the capital expense of buying hardware.” Some cloud providers fuel this statement by claiming their cloud offering releases most of an IT organization’s capital expense investments (referring to buying servers, networking equipment software licenses and other necessary IT infrastructure up-front vs. pay as you go). Now, if you’re an institution of higher learning, in particular a public institution, the opportunity to incrementally add operating expense may be difficult to impossible for some institutions (even if it means eliminating a substantial capital expense). For example, consider an institution’s budget with a $10 million capital budget in the current year, with an operating expense budget that is flat with the prior year (meaning no net-new increases to operating expenses). The ability to acquire the necessary computing gear via the capital budget would be compelling in this instance, conversely saving on capital while incurring incremental operating expense isn’t ideal.
Of course capital versus operating expenses is a minor part of the cloud analysis. Actually, cloud ROI for an institution (beyond the nuance highlighted above) may have little to do with capital expenses and operating expenses by themselves. Instead, it may have much more to do with the time-value of money, which again requires the institution’s own budget context to understand, because generically there may not be anything inherently beneficial in switching from an upfront investment to a cash-outlay over time. It all depends upon the particular institution’s situation and alternatives.
In calculating cloud ROI, time-value requires a multi-dimensional consideration, because paying for something now versus over time may include alternatives such as:
- What else can the capital expense be used for? Are there better uses for capital to further the mission of the institution?
- What else can the incremental operating expense be used for?
- Is our institution’s ability to budget for capital expense in out-years at risk?
- Over time, what is the value of the capital budget dollars today versus next year?
I posted some additional thoughts on this here, including a review of a great post by blogger Joe Weinman on Cloudonomics. In a nutshell, on a per time unit basis, it is more expensive to rent than to buy, so the utility paradigm needs to be applied in order to accurately know when cloud is cost-favorable, including variable usage requirements and wasted capacity with owned hardware.
Beyond the cost equation though, many more considerations should play a role in the drivers of cloud for Higher Education, including improving service levels, scalability, ease of implementation, reliability, maintaining technical and application currency, ability to respond and implement quickly, and spending less time on upgrades.
Further, when considering vendors compare their various pricing models, such as per user, amount of data transacted and the amount of processing power consumed. Additionally, the cost of moving applications themselves to the cloud should be understood, including an identification process of sorting out core and non-core applications, understanding all the sources of data for each, where it’s coming from, and all the integration points. Then the user interface should be analyzed (standard or proprietary, does it depend upon middleware, and are the APIs baseline or custom). And finally, determine if in-house expertise to support cloud applications exists, what it would cost to acquire and retain them.
With the multi-dimensional equation mentioned above, the risks should also be analyzed carefully, which include:
- Technical challenges
- Secure, store, and backup data
- Security plans and encryption
- Policy and compliance challenges
- Audits, access, and support
- Regulatory policies and institutional policies
- Privacy challenges
- Securing the privacy of your constituents
- Vendor challenges
- Vendor claims, assets, and contracts
- Contractual and organizational flexibility
- Contingency plans