Developing a Cloud Strategy

In an era when faculty and students often consume and implement cloud services without IT involvement, higher ed institutions need a strategy for the public, hybrid and private cloud.

Developing a Cloud Strategy

Photo: Shutterstock.com

When Howard Salis, a Penn State University assistant professor of biological and chemical engineering, developed a popular DNA Compiler application for researchers in his field, he opted to host it in the cloud using Amazon Web Services (AWS). As Salis explained, cloud solutions such as AWS allow researchers to develop highly scalable on-demand resources that are connected to the Web. (His work earned a Campus Technology Innovators award last year.)

Innovators such as Salis are forcing IT executives to develop strategies around public, private and hybrid cloud deployments. In a panel discussion on cloud strategy at last summer's Campus Technology conference, Kyle Bowen, Penn State's director of education technology services, said he looks at the cloud from three vantage points: publicly available services (such as social media tools); deployment of software applications on campus; and a larger "big data" infrastructure level.

Challenges and Opportunities

Often, faculty, staff and students come to cloud services on their own, consuming and implementing them independently. "We have seen a growth in social and mobile tools as a means of interaction, both inside and outside the classroom," Bowen said. While services such as Twitter and Facebook can help engage students, the relationships between faculty and students within those environments can be awkward. "There are certain aspects that are creepy and other aspects that are entirely innocent, but the rules for which is one or the other change on a regular basis," he said. "That becomes the slippery slope." It is not clear what is the latest social norm in these environments, he said. What is the latest environment that students have access to? Where is the data from those interactions going? "It is fraught with these challenges, but the advantage is that any faculty member or any student can engage these services immediately," he added.

Slightly different are tools or functions that historically the university would have deployed locally but that are now hosted in the cloud and delivered as a service. For instance, Penn State uses Yammer to facilitate interactions between working groups, Bowen said. It also uses services such as Qualtrics to provide surveys and other research instruments, and Box as a method for storing files. "We are also actively exploring delivery of video and videoconferencing out of the cloud," he said. "We are looking at how we can change how these services are offered in an effort to reduce the barriers to adoption."

At the infrastructure level, Bowen said, "the challenging part of the cloud is that anyone with a credit card able to pay a few dollars a month is now a sys admin, which creates its own interesting complexities." But it also creates the opportunity to rapidly innovate: The infrastructure is no longer a barrier to innovation, but the enabler of it. "It is no longer a matter of whether we can scale something," Bowen said. "We know we can. It is just a matter of how we architect a solution to do it. It is possible for us to store content in multiple geographic locations, and it enables innovation because no longer do we have to worry about whether we have infrastructure in place."

Cloud Strategy

Edward Mahon, CIO of Kent State University (OH), has led the development of a cloud strategy that goes hand in hand with the Information Services division's role as a service organization.

"The division is now viewing each activity as a service," Kent State's cloud strategy document states. "This increases our awareness of the effort and cost of all activities. Our activities are coming to be seen not as what infrastructure we own, but what services we provide. The cloud helps us to incorporate additional services that we can then offer to our users in combination with our on-premise services."

While most of its existing applications will not be cloud-ready in the near to midterm future (3-5 years), Kent State is preparing these applications for the cloud by moving them to a standard virtualized environment in the data center. "The division will continue to refine its cloud strategy, and, as the cloud model matures, evolves and innovates, the division will do the same."

Speaking on the same panel with Bowen, Dennis Ravenelle, release project manager for Harvard University Information Technology, reminded attendees that the cloud is not something new. "What is new is that it is reaching higher education and the public sector and being embraced there," he said. Harvard CIO Anne Margulies recently announced what Ravenelle described as an audacious goal: "Over the next three years, it is our intent to have all new services deployed in the cloud — and to migrate 75 percent of our existing services to the cloud."

Ravenelle said costs are a main and obvious driver. When Harvard weighed options for an IT service management (ITSM) solution, the cloud solution was almost $500,000 less expensive than the on-premise offerings.

He also contrasted the complexity levels and costs involved in upgrading software. "In 2008 we upgraded our on-premise ITSM tool suite. We were a couple of revisions behind. The project was estimated to take six months and cost $600,000. It ended up close to $1 million and took almost a year," he said. Fast forward to January 2014, when Harvard upgraded its cloud-based solution. After six weeks of testing, on a Saturday night they pushed the button and went live with it. "We had stood up expanded stabilization support for Monday. But by the middle of the day we called it off because it was a non-event. The upgrade cost us virtually nothing. These are some of the considerations."

Harvard has now moved to Office 365, uses OpenScholar for departmental Web sites and has enterprise agreements with Amazon Web Services and Google. "Departments and professors can take advantage of the enterprise agreement and stand up their own services in the cloud," Ravenelle said. Harvard also shares in a private cloud research infrastructure in a consortium with MIT, Boston University and the University of Massachusetts.

Ravenelle was quick to add that the cloud is not a panacea. "The devil is in the details," he said. You must make sure vendors have security certifications and are doing penetration testing. Portability is another big issue, he said. "If you deploy in Amazon's cloud and a year from now you decide that was the wrong decision, can you move it to Microsoft Azure or bring it back in house? Make sure in your agreements you have an exit strategy."

When negotiating contracts, he added, it is important to make sure you cover all the things you are going to need from the vendor. "Otherwise, they will nickel and dime you as you make changes and ask for things outside that scope." Universities must also keep in mind data sovereignty, he said. "If you host in a country other than United States, your data is subject to laws that are not U.S. laws. "Those are some of the things you have to consider."

When asked what happens to employees whose positions may be impacted by cloud service outsourcing, Ravenelle said he worked closely with the previous legacy platform administrator as the university transitioned to the cloud-based ServiceNow platform. "When we started heading down the path of going with ServiceNow, he was petrified," he said. "But we realized that his skills were transferable to helping administer the ServiceNow platform. It is a question of providing support and tools to learn new skills and make positions available to them. That is a key component of any responsible organization as it goes down this path."

Thinking About the Cloud

Like many university IT executives, Jill Albin-Hill, vice president of information technology and CIO at Dominican University in River Forest, IL, is evaluating how cloud-based services impact teaching and learning as well as the role of IT on campus. In an e-mail Q&A, she talked about her thought process.

CT: Dominican is implementing a cloud-based learning management system, Instructure's Canvas. Did that move make you and colleagues develop a more general strategy for cloud-based services at Dominican?

Albin-Hill: When we started the RFP process for a new LMS, we were in the middle of planning a five-year strategic technology plan. We articulated our operating philosophies, one of which is to "Provide a secure, stable, nimble infrastructure and leverage cloud resources for scalability." Going into the RFP, we felt that a cloud-based system would be a plus. For us, there is also a strong link to disaster recovery planning.

CT: What are some of the considerations with cloud offerings?

Albin-Hill: Certainly cloud partners must be considered carefully and contract terms must be closely reviewed and clearly articulated. A few of the key points we consider are data ownership, security and exit strategy.

CT: Are there other software types that you would consider cloud solutions for?

Albin-Hill: When a need arises we first look to see if existing systems or programs can perform the function — and then decide whether to build it, buy it or "rent it" in the cloud. We moved our student e-mail to hosted many years ago and have other specialty functions, such as faculty CVs, also with cloud solutions.

CT: Are there also implications for what your staff works on? For instance, if some applications and support move to cloud vendors, can your staff members shift to the most pressing IT needs on campus?

Albin-Hill: Staffing needs are certainly shifting as the environment becomes ever more complicated. As we leverage cloud services more, the work of the in-house IT shop changes. I need people who can integrate systems, not just perform software-patch management. The traditional database administrator needs to understand how to aggregate disparate data sources and what to do with "messy data," not just relational data structures. How we use all that data is another matter altogether. Business intelligence and data visualization tools are becoming critical for the business to make evidence-based decisions.

Featured