Find an Issue:

High-Altitude Technology

“Above-campus” computing services may offer the best option for responsiveness and reasonable cost. A task force from EDUCAUSE, NACUBO, and Internet2 is investigating the promise and peril of cloud computing for higher education.

By Richard Katz

See sidebar, “Forecasting Cloud Activity”

In the September 2004 issue of the Chronicle of Higher Education, Ira Fuchs, then vice president for research in information technology of the Andrew W. Mellon Foundation, argued that meeting higher education’s long-term software needs required the creation of a coordinating body. He suggested it be reminiscent of Bellcore, the telecom industry’s cooperative designed to set standards and conduct research and development. Supported financially by a large number of colleges and universities, Fuch’s “Educore” would coordinate the development and maintenance of open source software for the benefit of higher education, “giving universities and colleges the security they seek, but rarely find, in their relationships with corporate partners—who can raise prices and maintenance fees suddenly, or stop producing or supporting programs with little warning.”

Three major developments that impinge on Fuchs’s original vision are relevant here.

READ AN ONLINE EXTRA, "Looking at Clouds from Both Sides," in Business Officer Plus.

The dominant paradigm of software is shifting. Software is being reconceived, rewritten, and increasingly redeployed as a service. Even more, infrastructure—and entire business processes—are now becoming available as services. This development suggests that colleges and universities may be able to radically reconsider how they organize information technologies in support of their missions and the economics of IT and communications.

In a world where infrastructure, applications, services, and processes can be invoked and consumed on demand, providers no longer have to operate these services on campus nor provide sufficient capacity to meet “just-in-case” spikes in demand. While it is not yet clear whether total IT costs will decline, scale economies suggest that they will; and the shift to a services model will make it possible for IT to behave less like a capital good and more like an expense. Expenses are more predictable, more easily tracked, and—if desired—more easily recovered. The technology architecture enabling a shift in the locus of service provision is today called cloud computing.
Since 2008, much of the world has been in a recession. In the United States, the recession is thought by most economists to be the deepest since the Great Depression. College and university endowments have experienced losses—some substantial—requiring layoffs, furloughs, service curtailments, and a host of other actions in even the most-elite institutions. Widespread unemployment has resulted in a flood of enrollments at community colleges and state universities, outstripping the capacity of many institutions to meet constituent demand for their services.

These developments place new stresses on top of long-term trends in the cost and cost structure of higher education. Unable to add capacity rapidly—and badly in need of realizable economies of scale—institutions of higher learning are looking for new ways to deliver their missions. Supplementing or replacing face-to-face higher education services with services accessible via the Internet is an important class of alternatives under consideration by higher education’s leaders.
Many people throughout the world have been acclimated to taking courses, buying things, engaging in social exchanges, and so on over the Internet. The emergence of the Apple iPhone’s “apps store,” the success of the open source movement, and the rapid embrace of cloud computing by many software developers are fueling both tremendous innovation “in the cloud” and a robust consumer economy for services delivered “in the cloud.” Rapid growth in the supply of cloud services is drawing many students, faculty, staff, and other stakeholders away from institutional solutions and services.

This flow of consumer choice represents an even greater IT governance challenge to higher education than does the decentralization of IT to academic departments and laboratories. The unregulated movement of services off campus will pose a significant set of policy and competitive challenges to colleges and universities. Will academics use cloud-based tools to capture and store FERPA-protected information from classes without the institution’s knowledge? Are service providers’ IT environments secure? Who controls data in extramurally managed environments in cases of contract dispute, or mergers and acquisitions?

Fuchs’s original vision of Educore has only been strengthened and extended by intervening events. Higher education costs—including the costs of IT—must be reined in and higher education leaders must move to plan and organize a portfolio of infrastructure, applications, and services that is both premise-based and externally sourced, or they will inherit an unplanned patchwork quilt formed by constituents who have voted with their feet.

Finally, the failure to conceive and execute such a portfolio will increasingly consign colleges and universities to mediocrity as the locus of services innovation shifts to the cloud.

An Important Step

Task Force to Take Next Steps

EDUCAUSE, NACUBO, and Internet2 have created a task group to develop a coherent vision of a consortium to facilitate higher education’s adoption of cloud or above-campus services. To that end, the group will:

Recommend the roles, function, and scope of this consortium’s work and market.
Identify the network of business and academic relationships that will likely be needed to make the consortium an effective and competitive aggregator/broker of services.
Envision an operating model for this entity (in other words, “How does it work?”).
Develop a preliminary financial plan that demonstrates both the initial capital requirements of the aggregator, and a pro forma business plan that demonstrates financial sustainability (self-sufficiency) within a five-year time frame.
Identify strategies for forming the initial capital that will be needed, including specific partners that may be sought.
Evaluate governance alternatives and recommend a preferred strategy for guiding the corporate form that the aggregator entity should take, if the assessment of the opportunity, benefits, and economics all support a move-forward recommendation.

The task group will deliver a report to the presidents of EDUCAUSE, NACUBO, and Internet2 this summer. In coordination with this effort, Internet2 is exploring—with the advice of a panel of CIOs and others from the Tempe meeting—specific ways of brokering over its network access to infrastructure-as-a-service offerings, such as computing cycles and data storage.

These efforts are moving on a fast track. That’s because, notwithstanding the risks evident (and those not yet evident) in cloud computing—and the very real issues that make it hard for institutions to share risk, accountability, and gains—the time for higher education to seize the vision of above-campus services seems ripe. Today, too many of us are wrestling with the option of whether or not to offer a service at all. We clearly cannot keep all the trains running while retaining enough slack to monitor and develop those new technologies that will move our institutions forward. The promise of cloud computing and above-campus services is the promise of new capacity—capacity that comes from sharing the work.

In comments after the Tempe meeting, Nikki Krawitz, vice president for finance and administration, the University of Missouri System, captured those sentiments this way: “Higher education needs a new business model that includes leveraging scarce resources. We’ve all experienced the never-ending and growing investment in acquiring, implementing, maintaining, and upgrading hardware and software associated with technology solutions. The human resources required to do this reach all the way through the organization at a huge cost. This cost cannot be sustained given the growing constraints on resources. In the end, we must embrace change, have it foisted on us, or fall by the side of the road.”

Earlier this year, thanks to support from IBM, Pearson Higher Education, and SunGard Higher Education, 50 leaders from colleges, universities, corporations, NACUBO, EDUCAUSE, Internet2, and state networks met in Tempe, Arizona, to discuss cloud computing and the impending shift in the way infrastructure, applications, and services are sourced. This group identified a set of actions that colleges and universities could take to prepare themselves for what they concluded was a promising, risky, and inevitable shift. (See a comprehensive white paper jointly published by NACUBO and EDUCAUSE describing the meeting proceedings and outcomes.)

Participants also identified actions that EDUCAUSE, NACUBO, and other higher education organizations might undertake to educate their stakeholders and to identify practices that will mitigate some of the risks inherent in sourcing “above the campus.” Suggested actions:

Create a map of the cloudscape. Services such as Google Apps and Amazon’s Elastic Compute Cloud are well-known, but beyond a relatively short list, little is understood about the IT infrastructure, applications, or platforms that are available as “services” in the cloud. Research the cloud services market, especially in product-service areas that are high candidates for use in higher education. Specifically, develop a map that describes by category cloud-based infrastructure, platform, application services, and software that are available and the vendors that are competing in these categories.
Develop a guide to writing a business case for cloud sourcing. Based on findings for the cloudscape map, develop a short document that describes key questions to help frame an effective decision about hosting a system or service on campus, or “in the cloud.” Questions might include: What is impelling us to consider change?  Is cloud sourcing cheaper? Are cloud options more full-featured, reliable, robust, accessible, or nimble? What are the comparative risks? What is the risk of inaction?
Prepare a costing template. Identify uniform cost categories in both on-premises and cloud-based delivery alternatives and recommend preferred guidelines and methods for calculating said costs, when appropriate.
Complete a risk assessment framework and guide. Develop a document to guide practitioners and policymakers through  a risk analysis of premises-based and  cloud-based delivery alternatives.
Formulate audit guidelines to examine cloud-based infrastructure, systems, and services. This document will guide auditors through a risk assessment and audit of cloud-based transactions and portfolios.
Articulate what campus leaders need to know about above-campus systems and services. The result will be a short, strategically positioned document to educate trustees, regents, chancellors, presidents, senior staff, and academic leaders about the opportunities and challenges presented by cloud computing and the migration of infrastructure, systems, and services above the campus.
Identify new skills. IT professionals, business professions, lawyers, auditors, and others will need new skills to manage an infrastructure and service portfolio that is hosted elsewhere or invoked as needed as an Internet service. Skills such as contract management, creation of service agreements, and security management change as the portfolio changes. Institutions need to articulate a new curriculum and offer new professional development opportunities.
Transition to new governance. One of the key drivers of the need for higher education to act is the rapid evolution of the consumer market delivered via cloud computing. Increasingly, faculty, students, and eventually staff, are choosing to use a wide variety of consumer services in lieu of services being offered by the institution. This shift may have untoward economic effects and may create undocumented risks of security and privacy spills, data corruption and seizure, FERPA nondisclosure, and so on. The unregulated movement of IT and services off the campus represents an important opportunity to rethink IT governance. The deliverable would be an essay that would guide the “safe” consumerization of IT infrastructure and business and academic services.
Create model service-level agreements. A deconstruction or anatomy of a service-level agreement defines the purposes of an SLA, typical areas covered by SLAs, tips for negotiating service levels, and effective practices in managing contractual relations through such agreements. Ideally the model would include examples of effective service-level agreements.
Circulate a policy series. Short cloud-related policy perspectives might include topics such as privacy, data ownership, and exit strategies for above-campus services.

The unregulated movement of IT and services off the campus represents an important opportunity to rethink IT governance.

Most important, these leaders projected the timeliness of catalyzing the emergence of an entity or entities that would aggregate demand for above-campus services. Like Educore, a private higher education demand aggregator would be a coordinating body with a federated identity infrastructure, a common directory of services, standard contract terms, performance standards, standards of transparency, and so forth. The goal would be to aggregate demand from prospective college and university users in a way that simplifies the sourcing process and lowers the mistrust currently corroding the commercial sourcing environment.

Lowering the trust barrier to higher education’s widespread participation in a “cloud economy” entails the insertion of a trusted intermediary with community-informed rules of business and governance. Unlike Educore, the scope of this entity would not be restricted to software, but might evolve to become an ecosystem, access channel, or broker of software, solutions, infrastructure, and services. At the Cloud Computing Workshop, Shelton Waggener, associate vice chancellor and chief information officer at the University of California–Berkeley, described what was needed this way: “I want a consortium that facilitates targeted providers and standards, and makes the provisioning aspects of services customer friendly. It is the facilitation and provisioning of services—all done under master agreements—that will reduce the barriers to usage. I don’t have to negotiate each time, which will benefit suppliers and campuses alike.”

Conference participants asked EDUCAUSE, NACUBO, and Internet2 to evaluate the feasibility of creating such a consortium on behalf of higher education. (See sidebar, “Task Force to Take Next Steps” on plans for doing that.) Marilyn McMillan, vice president for IT and chief information technology officer at New York University, New York City, extended this visionary mission to accent an additional role: “We need an incubation function. Don’t let 1,000 flowers grow unattended, but rather operate something on a venture-philanthropy model.”

RICHARD KATZ is vice president of EDUCAUSE, Denver.

Forecasting Cloud Activity

After the cloud computing conference in Tempe, EDUCAUSE followed up with higher education and business leaders, asking them about the promise and peril of cloud computing and above-campus services. Following is a snapshot of their thoughts. To read the rest of the commentary, see "Looking at Clouds From Both Sides" in Business Officer Plus.

RICHARD KATZ, vice president of EDUCAUSE: Our colleagues are worried about the security of institutional data and processes and don’t trust third parties—especially commercial ones. What steps must we take to overcome this mistrust?

FRED ROGERS, vice president and treasurer, Carlton College, Northfield, Minnesota: Initial trials with intended partners or consortia will help. Clear designation of relative liabilities and responsibilities is essential to any trial or subsequent relationship. Institutions need to understand both the risks they assume in new relationships and the risks they carry with the status quo.

ELAZAR HAREL, vice chancellor for information technology, University of California–San Francisco: The mistrust is mostly an illusion; no data substantiates these worries. In fact, it is likely that our internal systems are often less secure than the commercial cloud services, primarily because they [commercial services] are usually supported by more-experienced and well-funded professionals. We do need to be careful, though, in how we select these vendors and what agreements we sign with them. For example, ownership of data must be clear and basic level of services, such as security and privacy, need to be addressed in those agreements.

JOANNE KOSSUTH, vice president for operations and CIO, Olin College, Needham, Massachusetts: We have to work to build effective third party-relationships. Relationships are built on trust, which takes time and good-faith investment of all parties. A good start would be to draft effective contractual language that responds to many of the questions that we all have. Where is our data stored? How is it broken down? How are we assured of getting all of our data back if a business is sold or goes into bankruptcy?

Since we likely will not have all the answers at one time, how do we pilot or try something in this space with a willing partner and then build on that? If these issues cannot be addressed, then perhaps higher education needs to create its own structure from an already-trusted group.

NIKKI KRAWITZ, vice president for finance and administration, the University of Missouri System: Fear of the unknown drives distrust. Having a clear understanding of how data is secured can go a long way toward overcoming the idea that data will be at risk. Knowledge, along with properly drawn contracts and service-level agreements that address security, among other things, will be critically important.

BRAD WHEELER, vice president for information technology and CIO, Indiana University, Bloomington: Corporations long ago learned to assess what risks could be managed by contract with third parties and what had to be held within the firm for risk avoidance. Institutions need to rethink what services can be contracted to commercial firms and what must be held within institutional control when there is a cost premium to do so. Fortunately, higher education also has a third option of institutional consortia that can achieve above-campus  economies of scale while operating more directly under institutional control than commercial firms.

KATZ: It seems that scale economies really favor cloud-based over premise-based solutions. What do the economics look like to you?

ROGERS: We have enough experience to know that new technologies are often underpriced in their early stages as providers seek widespread adoption. Successful corporations will eventually determine a viable business model for cloud computing and, as they do so, the true costs and benefits will become clearer.

HAREL: Cloud services are likely to use less-expensive energy sources and obtain better economies of scale than any of the universities can. As these services become commoditized, there is less and less justification for premise-based solutions. This does not mean that everything should move to the cloud. Services that are strategic resources to the university may be better provided in-house, and there will probably always be some situations where distance or network bandwidth limitations require premise-based solutions.

KRAWITZ: Increasingly, information technology is a utility. End users want a menu of high-quality, reliable services that are easy to use and meet their needs on demand. Like apps for the iPhone, they want the latest, and they want it now. They don’t care where the service originates or how it gets to them. Given the huge investment required to build, maintain, and meet this kind of demand, cloud solutions will win out over premise-based solutions except where that solution is truly unique to an institution.

KATZ: How do you think the higher education world might look in 10 years, as a result of cloud computing?

ROGERS: Institutions will increasingly ask themselves how they can access more service at lower costs across a wider spectrum of services than what we now think of as computing services. Media development and dissemination will be a much bigger component of the future of education and teaching. A development like cloud computing will change the issues we address at least as much as it will change how we address the problems we know today.

HAREL: I can’t think in 10 years of IT increments; it will look very different for many of us in 5 years. We will no longer have data centers on campus, we will not provide office services internally (such as, e-mail, e-calendar, and so forth), and many of our administrative services will be hosted elsewhere.

ANNE KEEHN, senior vice president, institutional sales, marketing, and strategic alliances, Pearson eCollege: Amazon spends millions and leverages data, models, and good science to help you find a movie or book that you will enjoy. The cloud can bring that degree of personalization to the ways we learn through prescriptive content, learning paths, and personalized experiences for each learner. Telepresence technology, simulations, holographic images, and 3-D will add up to a more multimodal and Internet-delivered learning environment. Some of it is available now, so in 10 years anything is possible.

KOSSUTH: Ten years from now, higher education should be a more collaborative, innovative place where unique resources, readily available to all students, will change the financial model and higher education modes of delivery.

DARREN WESEMANN, chief technology officer and chief product officer, SunGard Higher Education: Ten years is an eternity in computing. I believe it’s safe to assume that we’ll see more components on demand, similar to the ease of buying a book today. The accessibility of functional components will definitely improve, as will the interoperability of components into solution compositions—all from the cloud—which ultimately means less equipment on site to worry about and more direct value.

WHEELER: The greatest potential for change is in the mind-set of IT staff. Today, when something new like emergency notification systems captures attention, institutions spend about five years of complex, one-off institutional adoptions by leaders and laggards. Our industry can’t afford that approach, and in 10 years, we should look at provisioning a few above-campus services when a common need arises. Our historic obsession with institutional differentiation for common IT services is unsustainable.

^ Top