Why Cloud?

ITS Case for Change

History

Back in 2017, ITS made a decision to move as many ITS systems as possible to the cloud in three (3) years. This decision was being driven by two critical demands: 1: ITS was experiencing a number of significant incidents, where critical systems were experiencing performance issues and affecting our ability to deliver reliable service to the U-M campus. Moving systems to the cloud was viewed as one way to help build resiliency. 2: A need to provide 1 Megawatt (MW) of power at the Michigan Academic Computing Center (MACC) data center to accommodate growing computing needs of the U-M research community within three (3) years.

In concert with the intention to migrate ITS systems located in the MACC to either the North Campus Data Center (NCDC) or to public cloud environments, ITS embarked on a review of fifty-three (53) critical infrastructure systems for resiliency and disaster avoidance (SRDA). Through the SRDA effort, it was confirmed that ITS had accumulated significant technical debt that required investment to ensure resiliency. Some services might benefit from moving to a cloud service to increase resiliency. Some services might achieve resiliency on-premise, without the need to move to the cloud. In addition, these efforts uncovered a large number of systems that could be retired and removed from the MACC data center, allowing access to the 1 MW of power that was needed for the research community.

Changing Landscape

Both of these developments have changed our approach to adding a significant cloud component to our strategy.

ITS leadership continues to support an effort to utilize cloud technologies for ITS systems, as is reasonable. Unlike other institutions, this change is not being driven by an immediate need to free up space in U-M data centers, which allows us to be more deliberate in our decisions on what should be moved to the cloud.

Current Drivers

The current approach is being driven by the following factors:

Prepare for the New IT Paradigm

In years past, the model most universities adopted for their enterprise and productivity applications was to purchase, configure, and deploy applications and associated infrastructure on university owned hardware in on-premise data centers. This included HR, finance, student enrollment, email, learning management systems, patient care, and other applications, in addition to the storage, compute, database, and other infrastructure needed to run them. This approach is becoming increasingly inefficient and expensive compared to alternatives.

With the advent of ubiquitous high speed networks and hyperscale computing environments, the IT paradigm is changing. Companies such as Amazon, Google, and Microsoft now offer access to their infrastructure as a service to meet the needs that were previously fulfilled by on-premise virtualization, storage, and database infrastructures. Furthermore, software vendors are now able to operate and deliver their systems at scale in their own data centers (or those of a public cloud provider’s), delivering their own software as a service to individuals and enterprises.

U-M has been steadily adapting to this paradigm shift using applications such as Concur for Travel, Canvas for Teaching & Learning, and Google Apps for email and productivity. At the same time, leveraging public cloud provider’s infrastructure as a service offering for email routing and Unizin Data Platform. U-M schools, colleges and departments are interested in exploring how these models may provide for other key administrative, infrastructure, and research needs over time. Many of our faculty and researchers are already working in a cloud environment to manage their big data projects. We’ll be most effective at engaging with campus if we are fluent in cloud technologies and providing direction, consultation and support. This will require that we provide staff with appropriate training and development opportunities.

Newer Technology/Features

Most IT industry innovation and investment is now being made in cloud technologies. ITS will position U-M, as a world class institution, to take advantage of these cutting edge innovations. The need for features such as machine learning and handling of big data will grow in importance for a research institution such as U-M. 

Flexibility/Agility

The proliferation of cloud services along with automation and orchestration allow for quick deployment of IT resources that can be scaled up and down with ease and speed. As new applications and options become available U-M does not need to sustain their operation and can also use them for a short period of time. The cloud is becoming an ecosystem where people can pick what they want and pay only for what they need. This allows staff to focus on mission critical activities.

Resiliency

Cloud technologies can provide significant advantages in making our systems more resilient. In order to ensure that we are providing robust and reliable IT services to the U-M community, it is important that we leverage these benefits.

Data Center Capacity

While the immediate need for additional data center capacity has passed, on-premise space will continue to be in demand for some time. U-M’s commitment to research and discovery, academic innovation, data and computational sciences, precision medicine, high performance computing, and preservation of knowledge will continue to drive demand for data center space. It is quite possible that a researcher will need more access to an on premise data center like the MACC.