Looking at Clouds From Both Sides
At an EDUCAUSE-NACUBO conference in early 2010, higher education leaders discussed the promise and peril of above-campus computing. After the meeting, EDUCAUSE followed up with some of the participants to dig deeper into the discussion.Edited by Carole Schweitzer
In the feature article “High-Altitude Technology” in the July-August issue of Business Officer, Richard Katz, vice president of EDUCAUSE, analyzes the benefits and risks of moving higher education to a totally hosted software, infrastructure, and business services model. The technology architecture enabling such a shift in locus of service provisions is today called cloud computing.
Katz's article highlights the discussion that took place in Tempe, Arizona, early in 2010, among 50 leaders from colleges, universities, corporations, NACUBO, EDUCAUSE, Internet2, and state networks. The meeting was also supported by IBM, Pearson Higher Education, and SunGard Higher Education.
After the conference, EDUCAUSE queried higher education and business leaders about specific issues related to cloud computing—or what some are calling “above-campus services.” Here's what they had to say.
RICHARD KATZ, vice president, EDUCAUSE: We have long considered using hosted or shared services, and e-commerce is widely in practice among higher education practitioners. What's new and different about the cloud?
FRED ROGERS, vice president and treasurer, Carleton College, Northfield, Minnesota: Perhaps the newest thing is the single name for such a wide variety of potential services, and thus confusion about the source of cost and value. The trick seems to be identifying the time and scale at which an institution or groups of institutions can derive value from shared or cloud services.
ELAZAR HAREL, vice chancellor for information technology, University of California-San Francisco: It's not all new and different, but cloud services seem to provide a more granular set of options (for example, subscribing to selected services from a large menu of items); services that can start and end at any time (for example, accessing a massive amount of data storage for just a short period of time); and the ability to take advantage of the high-speed networks that did not exist until recently.
ANNE KEEHN, senior vice president, institutional sales, marketing, and strategic alliances, Pearson eCollege and Pearson Learning Solutions: What's new is the ability to use the cloud for robust applications that can support the institution's mission and business. It is now possible to combine applications and services that create a superior user experience, rich functionality, and reliable service at a fraction of the cost. The burden on us is to demonstrate that we offer the stability, capacity, performance, scalability, state-of-the-art data centers, and massive bandwidth required to meet an institution's needs for trusted service.
NIKKI KRAWITZ, vice president for finance and administration, the University of Missouri System: From the perspective of a business officer, cloud computing seems to provide the foundation for moving to a new business model. In this new model, each institution no longer needs to own all of the technology and related expertise that supports the mission. The implication is that through economies of scale, access to cutting-edge technology would be available to all institutions in real time at a lower cost.
DARREN WESEMANN, chief technology officer and chief product officer, SunGard Higher Education: One might think of the modern cloud as hosted shared services on steroids. That's because everything in the cloud is shared (network bandwidth, multitenant database instances, and virtualized computing as well as other common services). But more important, the cloud services and the infrastructure itself are “provisionable” (and even procurable, in many cases) on demand. That means quicker access to computing through a browser and more cost-effectiveness because of the more efficient use of resources and economies of scale.
BRAD WHEELER, vice president for information technology and chief information officer, Indiana University, Bloomington: The real potential of above-campus services is improved costs for higher education through economies of scale and leverage of IT services that are beyond the grasp of even large institutions. Sure, some commodity cloud services are no different than decades of sourcing decisions, but we will miss the real opportunity if we allow a procurement mind set to pursue the elusive fallacy of single institution optimization for common services.
KATZ: While many people use consumer cloud services like Gmail, Flickr, or Picasa, most people do not want their information mined, modeled, profiled, and sold. What about privacy in the cloud?
ROGERS: This is a timing issue. As the business models are worked out, the issues will become clearer. Assumptions made at the beginning of this cycle of adoption may not be sustained as usage and suppliers evolve.
ROGER BRUSZEWSKI, vice president for finance and administration, Millersville University, Millersville, Pennsylvania: This could the major drawback to cloud computing services. The issue of who controls and manages privacy, security, and data integrity is a serious concern for all of us. What will be the rules and regulations that providers will need to follow in order for institutions to feel comfortable? The corporate world is addressing these issues. We should take a good look at the practices of very large firms and pick the best of the best for higher education.
And, what happens if you make the move—and eliminate hardware, software, and personnel while at the same time upgrading your network—and two years down the road, your goals are not being met? Before we move to the cloud, I suggest we have an exit strategy.
HAREL: This is mostly an illusion. As long as confidentiality, security, and privacy are maintained and as long as the price is right, why should we worry about others mining generic data?
KEEHN: Privacy is important! Users must trust the environment in order to have successful learning experiences. The trade-off, of course, is with personalization. While some personalization can occur with anonymous information, at some point users may choose to share some personal information in return for improved experiences.
JOANNE KOSSUTH, vice president for operations and chief information officer, Olin College, Needham, Massachusetts: Privacy is a major issue whether on premises or in the world of cloud computing. Everyone has differing needs for privacy. If most people do not want their individual information identified with them (except for friends), can we assume that aggregating information addresses privacy for the majority? Does aggregated information drive innovations in technology, especially in the marketing and social software areas? Would we change our minds about personal data, if we got discounts from loyalty cards our constituents use? How will technology drive our options in the future and how can we be protected from the consequences of our own choices?
KRAWITZ: Contracts for cloud services need to specify the extent to which an institution's information should be available to be mined, modeled, profiled, and sold to third parties. The cloud vendors need to create an environment that prevents these actions to the extent that they are preventable. At the same time, I wonder why we think the technology environment that we all operate in now fully protects against these actions by third parties.
WESEMANN: Because cloud computing includes the sharing of infrastructure and services, the data for each tenant can be kept secure. Most cloud vendors provide strong safeguards and security to keep data secured while it resides in the cloud, and protected when it is in transit. An easy way to bypass the concern is to leverage cloud computing for the services that don't require persistent data (of which there are typically plenty of scenarios in any given enterprise system). For those scenarios that require data, ensure your cloud provider follows strict security controls, implements standards, and allows audit of its security operations.
KATZ: One of the key arguments for moving to cloud computing is to alleviate the risk of vendor lock-in. Does sourcing infrastructure, platforms, or applications in the cloud reduce, eliminate, or perhaps amplify the risk of lock-in?
ROGERS: In part it will depend upon how differentiated the services become and how many suppliers survive. Institutions need to consider what they might outsource to cloud suppliers and thus where they want to assume mission risk. Typically, programs that differentiate or uniquely define an institution's ability to meet expectations and mission are not good candidates for outsourcing.
BRUSZEWSI: To quote Joni Mitchell, “I look at clouds from both sides now.” I believe there are cost benefits to moving from the current state to clouds. For example, you definitely save by moving from campus-based e-mail to Gmail or Microsoft mail. But what happens when the business model does not work and the free service is not longer viable? What is the cost of picking it all back up? We have been involved with some private cloud software for a couple of years with mixed reviews. The costs are higher at this point, but we do see more economies kicking in as the software matures.
HAREL: There is always a risk of lock-in, regardless of the service being external or internal. Therefore, systems and data should be designed in modular fashion, using industry standards so that transition from one vendor to another is as easy as possible.
KEEHN: Vendor lock-in, of course, is not a risk that is limited to cloud computing. Closed and proprietary approaches that promote lock-in not only discourage adoption and reduce potential customers, but they narrow the breadth of learning technologies. The diversity of learning technologies is so vast that no single provider—cloud or otherwise—can meet all the needs on an institution. Cloud users should insist on standards-based APIs (application programming interfaces) that leverage reliability for core functions and allow for the construction of rich applications from other suppliers-commercial or open. This strategy not only reduces the risk of vendor lock-in but promotes learning ecosystems that can integrate with one another and promote the notion of personalized learning.
KOSSUTH: It is not clear that the cloud will alleviate vendor lock-in risks. In fact, given the current players in the cloud, one could make the counterargument that the cloud players are merely locking us into vendors at a different level, and perhaps a riskier level. The risk involved would be in the storage, distribution, and protection of our data. In my opinion, if the interoperability standards do not exist then the risk of vendor lock-in is not much different than the current state.
KRAWITZ: The vendors want lock-in; it doesn't matter whether it is technology or some other good or service that they are selling. It's up to the customer to understand, demand, and choose services that minimize the ability of a vendor to lock in terms and conditions and maximize the control over their destiny.
WESEMANN: Certain cloud providers offer proprietary coding and configuration (either as an option or a requirement). Cloud lock-in occurs when those offerings are embedded in the applications hosted, as it makes the application nonportable. As long as the application placed in the cloud follows popular standards with no proprietary extensions unique to the cloud vendor, there is little risk to being locked in.
WHEELER: I'm with The Economist on this one with its online article “Cloudy With a Chance of Rain.” Switching costs rise rapidly as institutions move from Infrastructure to Platform to Software as a Service. There will be no incentives for cloud providers to ease switching to other vendors, so a colleague of mine offers the essential question: “What's my exit strategy?” for any agreement.
KATZ: What about integration? More than ever I want my institution's Web presence to look and feel well integrated and to create a user experience that builds loyalty and a sense of affiliation. How will a mixed portfolio of premise- and cloud-based service affect this?
ROGERS: There are already examples of Web presences that include a mixture of local and hosted components. The issue of branding and identity is an important objective, but it seems likely that we will develop models for hosting specific pages and content mixed into an environment of local presence and identity. Today's students are not put off by links that take them to another environment, especially if there are convenient links that retain routes back.
HAREL: This is likely to become a difficult trade-off. The integration issues can be somewhat alleviated by using common core services, such as a single sign-on, a single portal, single content management system, and so forth. Fortunately, the younger generation is now accustomed to a variety of look and feels, so this may be less of a problem in the future.
KOSSUTH: Integration of services is among our largest challenges and likely takes most of our time and resources. I do not see that changing. If anything, the availability of new services in the cloud and the drive to incorporate those services will increase the challenge. Standards will play an important role in easing the integration concerns, especially if the vendors commit to true and unadulterated standards. New technologies in authentication and mobility will provide tools, as will social networking software evolutions. We will need to be responsive to customer demands (while deciding what we will no longer do) and provide the most adaptable, personalized approach to integration possible. Much of the savings from utilizing cloud technologies may very well be spent on the continuous integration of emerging tools.
WESEMANN: The key to integration is found in Web standards. If the common services offered by the cloud vendor—as well as the hosted applications adhere to interoperable standards—the extendability of the application in the cloud becomes manageable. Some cloud vendors offer tools to make this easier, but beware of vendor lock-in if the tool does not export portable bits. The best scenarios are those in which the tools with which your institution is familiar work with those provided in the cloud.
KATZ: Imagine that you and I are headed into the president's cabinet meeting, with a chance to win significant funding for an enterprise cloud computing initiative. What does our pitch consist of?
ROGERS: We should highlight the benefits to academic mission and the areas where savings will be achieved in the current organization, including staffing and capital investments.
HAREL: I'd say flexible spending pattern, quick adoption of advanced technologies, lower overall cost, high customer satisfaction, and green solutions.
KOSSUTH: The cloud will increase our agility and responsiveness, especially with regard to research agendas and collaborative opportunities. The greatest impacts won't be savings, but improvements in the research and collaboration areas so that students, faculty, staff, and external experts can innovate quickly, fail quickly, and restart even faster. The potential for achieving game-changing results is enhanced and students are better prepared to deal with the world's problems through these experiences. Using the cloud as a platform allows us to use the latest and greatest technologies with limited investment.
KRAWITZ: If we want to be on the cutting edge, we can no longer afford to go it alone. The technology that supports the teaching and research missions and administrative functions is changing rapidly. The rapid change drives up the cost to stay current and challenges our ability to meet faculty, student and staff needs with high service quality. A cloud computing initiative will allow us to be responsive to faculty and student needs now and in the future at a reasonable cost while improving service quality.
WHEELER: Research and education are driven by innovation at the edge—faculty, researchers, staff, and students in the many disciplines that comprise the academy. IT has unquestioned economics that favor aggregation and scale at the core. An enterprise cloud computing initiative will create greater efficiencies in that core and enable more, not less, scholarship at the edge.
KATZ: What happens if higher education (or my campus) chooses to sit this one out?
ROGERS: That's like asking people whether it was worth adopting personal computers or the Internet, since those early offerings were so slow and limited in their initial apparent value. This is a question of time, not a question of whether or not it will take hold.
KEEHN: Those institutions would miss out on an opportunity to improve their IT budgets, provide business continuity, and take advantage of a new way of supporting their blended and online learning environments with a more reliable, flexible, and scalable infrastructure.
WESEMANN: The downside to ignoring the cloud is the missed benefits of greater efficiencies. But, the cloud doesn't need to be adopted wholeheartedly to reap some benefits. For example, a small service or two can be outsourced to the cloud and over time utilization can grow. Starting small is good advice for most new things anyway; get a couple of early wins and prove the benefits before going deeper.
WHEELER: I don't think anyone sits this one out for the long term. The issue may be one of only early and late adopters. This is not only an IT evolution but an economic one. No one sat out the PC era forever, and we see few, if any, departments of word processing anymore.
CAROLE SCHWEITZER is senior editor of Business Officer.