While the most acute phase of the economic crisis may be behind us, the reality of the "new normal" in funding (or revenue) levels has caused multiple industries to retool themselves. Many believe that part of the solution to achieving lower run rates may lie in the latest trend in IT — cloud computing. On the basis of the hype — unlimited capacity, lower costs, and minimal barriers to adoption — the promise of cloud computing is particularly enticing given the extreme financial pressure many higher education institutions currently face. As seasoned information technology leaders who have seen many IT game-changing trends come and go, we know that for every new advance that actually pans out, there are as many or more cases where the reality just doesn’t match the promises or expectations. So the question we all struggle with now is, where will cloud computing fit into that spectrum — and how do we take advantage of the benefits while avoiding as many of the pitfalls as possible?
Beginning with this column and over the course of the year, I will present four topics that higher education IT leaders evaluating how to move into the cloud should keep in mind:
- Organizational impact: changes in organizational structure and culture needed to support moving to the cloud
- Cloud structure: transitioning to cloud services from siloed solutions
- Data management: new and challenging issues for IT in a cloud-sourced world
- Contracts and agreements: rethinking the role of the contract when working with commercial providers and partner institutions
Before examining these impacts and aspects of moving to the cloud model, we need to be clear on some terminology. Marketing machines are in full force around "cloud," with each purveyor of high-tech wares offering a definition of cloud computing that best suits its purposes, resulting in a general lack of clarity about what cloud computing really entails. Rather than letting vendors define the term, we can refer to the emerging consensus driven largely by the broader community, which helps break cloud computing into three smaller components:
- Infrastructure as a service: IaaS
- Platform as a service: PaaS
- Software as a service: SaaS
Brad Wheeler, the CIO of Indiana University, and I endorsed these concepts in our article, "Above-Campus Services: Shaping the Promise of Cloud Computing for Higher Education," in the November 2009 issue of EDUCAUSE Review. I will continue to reference these components over the coming year.
Transitioning to a Shared Services Organization and Culture
For many years in higher education, we have seen a trend toward decentralization of IT service delivery. How many IT departments exist on your campus? More than ten? More than twenty? Thirty or more? In most cases, this organizational model arose not from planning but from local necessity. While central IT units were generally organized around large central systems such as Student, Finance, or HR, or in some cases around network and voice infrastructures, it wasn’t uncommon for most schools or colleges, administrative functions, or research units to create their own local IT departments. This has resulted in the current structure of many small-scale IT providers with high per-unit costs, where the ability (or incentive) to leverage solutions between groups is very small or nonexistent.
This environment has also encouraged the selection of technologies on the basis of what the individual technologists know and how quickly a given technology would allow them to provide a solution. The result is hundreds of technical solutions meeting duplicate or very similar needs. At Berkeley, IT professionals report into more than 100 different departments, many with four or fewer IT staff, each working to build and provide technology solutions optimized to local needs, often with little awareness of identical work being done in the other units. For smaller institutions, this fragmentation might not be as widespread, but it can be just as severe in terms of suboptimal development of institution-wide, scaled solutions and in considering cloud services to meet these scaled needs.
Providing a single approach to IT solutions is very difficult given the diverse constituencies of a higher education campus. Discussions about solving the economy-of-scale issue frequently lead to complaints about the dreaded "centralization" of IT services and the perceived loss of features and autonomy. If adopting cloud services is on the horizon for your institution, I believe the time has come for you to move away from the siloed, local delivery model and consider organizing and training your IT teams around the concepts of IT demand and supply. Rather than centralized versus decentralized, the real discussion for higher education IT needs to be about demand planning and service delivery, and where those two activities most appropriately belong to achieve maximum benefit at both the institutional and local levels.
The IT Demand Organization
Historically, demand planning and service delivery (supply) activities were combined and handled by department-level IT staff — business analysts and technical experts who understood the local requirements and built solutions to meet them. The first major change many of us need to consider to successfully integrate with and expand upon services the cloud offers is to provide training to staff to specialize in demand-planning activities. The goal is to have individuals with the same or greater level of technical sophistication as IT generalists, but with a sole focus on the demand rather than supply side. Critical skills to develop include business analysis, capacity planning, requirements gathering, documentation, negotiation, project management, and financial planning and portfolio management. A high level of technical competency is also crucial. These IT professionals will need to understand how to work with process owners to create clear requirements and how to map available solutions to meet the documented need. Most IT staff currently must perform these functions to one degree or another, but often these skills take a back seat to real-time problem solving, code development, systems administration, network engineering, or similar tasks.
The IT Supply Organization
IT supply will also undergo a transformation, one even more radical than that of demand. The days of a single individual or small team owning and controlling all aspects of an IT solution are gone. As technical options in virtually every domain (application, infrastructure, data) continue the exponential growth of the past decade, we can expect to have even more choice, and we need to leverage those options well to provide the best, most cost-effective solutions. When using cloud-operating models that allow for near instant provisioning of services, you don’t even need to download software. That’s a radical departure from past models where large offerings of a central department used to be delivered over a period of years to captive audiences with little variation or customization. The local solutions attempted to respond to the feature gap created by large, slow-moving systems, but could not achieve their economy of scale. Today, cloud computing provides both: thousands of similar customizable offerings, often at substantially lower costs.
To move campus IT organizations toward a design that will provide support using the cloud, you have to break with the functional, control unit, or centrally designed organizational models for delivering solutions. The best starting point is usually the central units, where sufficient scale exists that structuring the organization around service delivery is possible. Instead of separate HR, Finance, and LMS environments, align them by horizontal service layers. Bring together all the network design aspects, the systems management activities, the storage services, the visualization and display technologies, the like application development activities, and the database management professionals.
Initially, there will be a significant service hit because the groups’ operating skills might not be optimized for the new roles. You will also have significant customer engagement challenges as long-standing relationships ("When I have a problem, I call Jane in the Student Applications group and she will take care of it") are redefined (when Jane takes a new role). Each of the horizontal layers will need to redefine who their customers are, understand the new challenge of joint accountability when the service isn’t meeting expectations, and develop new models to address problems. This is a more complex model than the past dedicated-team-by-customer approach, but offers substantially greater opportunity for cost savings through scale and will position you to take advantage of cloud offerings at the horizontal layer (such as Amazon EC2) and also vertically (such as the salesforce.com application). The end result should be teams of specialists dedicated to improving services that can be provided at scale across functional needs. The model also can provide much clearer career paths for technical staff, who can see the specific skills needed to advance in a specialty, compared to the less clear career development options for generalists.
Putting It All Together
In creating demand and supply teams, both groups will need clear career paths and new job descriptions. Generic "System Analyst" or "IT Engineer" labels will no longer suffice as staff specialize in specific service layers or demand-planning activities. Generalists are always needed, but they should focus on either the demand or supply side rather than trying to do both. At Berkeley, we undertook a process to entirely remap every one of the more than 1,000 IT staff to new job families, with descriptions and salary scales based on IT industry standards. The objective was to map IT talent away from generic templates to more detailed and meaningful descriptions, as well as to create clear roadmaps for career development among and between the various job families. While this work was not specifically undertaken to enable a demand- and supply-based IT organization, the work we did will be critical to our ability to meet that goal. (More information about Berkeley’s job mapping effort is available at http://careercompass.berkeley.edu/.)
While the job mapping and other HR work will position you to take advantage of cloud services, many in your community will still be concerned about giving up the perceived efficiencies of local, one-stop shops for all IT needs. You need to create a critical mass of the IT organization dedicated to the supply side. While we worked across campus on the job mapping, we also undertook a comprehensive restructuring of the central IT organization at Berkeley, moving away from siloed customer groupings to a services-based model. The new model supports delivering IT as a service by creating four specific service teams: Infrastructure, Applications, Data, and Client Services. The units rely on each other and must work together to develop and deliver solutions for large or small customer needs. The short-term advantage is that smaller projects can use the environments developed for larger solutions if the environments were originally architected to scale both up and down. This moves the central IT groups into operating as local private campus cloud providers rather than as exclusive owners of siloed solutions.
The cultural transition to demand and supply organizations, with both the subtle and dramatic changes to how teams interact with clients, vendors, and each other, is a significant change-management challenge. However, it puts internal IT teams in a very good position to consider and respond to the fast-moving pace of IaaS, SaaS, and PaaS operating models. Initially during our transition, many indicated that moving away from the function-oriented model would break customer relationships or slow the delivery of solutions. Instead, we have improved uptime through enhanced change management (another requirement for this kind of model) and improved response times for groups now populated with focused teams with high skill levels, all while lowering costs. The savings have been particularly pronounced in our IaaS and PaaS offerings, where we can quickly take advantage of technology gains to lower costs.
The above transformational approach is very difficult to achieve on a project-by-project basis. Your IT organization, as well as all IT staff at your institution, should be given clear direction as to the future delivery model for technology. If you aren’t sure about making such a radical departure from your current organizational or operating model — or don’t feel you can begin the process of adopting the demand and supply approach for running internal cloud services or taking advantage of third-party services — you can be assured that there is an alternative. Your customers will simply start adopting commercial cloud offerings on their own. Likely they already have.
In the next EQ, I will discuss the specifics of creating private cloud services and the benefits and challenges of the model.
© 2010 Shelton M. Waggener. The text of this article is licensed under the Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 license.