The Highlights

• Eliminated dozens of process bottlenecks with self-deployment utilizing a container model.
• Improved the DevOps pipeline using an intuitive container orchestration GUI.
• Built a valued management layer on top of existing orchestration structure.

The Customer

As a global health research center, the Institute for Health Metrics and Evaluation (IHME) contributes to many research projects that can improve prevention and screening for health risks, lower the incidences of chronic conditions, and ultimately, save lives. Dozens of data analysts work with academics to craft research results in a dedicated analytics and learning platform supported by container virtualization. For this platform, IHME has a significant hardware architecture. The IT leadership have to figure out how to support all of this data-heavy research and allow stakeholders to work effectively through the network.

“Health research is best served by agile architectures. Systems where the data is backed up by great admin systems and where platforms use automation right. When we get that right, we unlock those top levels of productivity.”

Andrew Erns, Operations Manager

The Challenge

One of the biggest challenges that internal IHME staff faced early on had to do with the information pipeline, where researchers and web developers have to collaborate to produce results.

As the organization grew, top IT people realized that teams were micromanaging large numbers of servers. That’s where containerization came into the picture. IHME personnel were looking for a system that would manage containers in isolation, maximizing the effect of the self-contained “run anywhere” philosophy of container virtualization.

“Using Rancher, we’re able to get our deployment bottlenecks out of the way and that means we get our science done quicker.”

Andrew Erns, Operations Manager

The Results

IHME teams chose to build an enterprise container management platform based on Rancher. By using Rancher, the department created a self-service containerized environment that developers could deploy their applications on. Teams were able to get data out of silos by streamlining the pipeline and making it easier for research and development teams to work together.

With Rancher-driven container virtualization, developers are able to “self-deploy” all sorts of projects for testing in an efficient way. Projects move toward production with less hoops to jump through — this in turn accelerates the research that scientists are doing.

It also frees up resources that allow for growth and scalability. Leaders spend more time on core processes, and less time worrying about whether architecture navigation is intuitive. That’s a major part of what IHME get out of Rancher: An intuitive user interface that has made it easy for teams to adopt containers and orchestration.