Home »Resources»Case Study

Case Study

The Institute for Health Metrics and Evaluation (IHME), University of Washington

IHME Boosts Productivity, Avoids Cost by Deploying Docker and HPC Workloads on Kubernetes.

Executive Summary

IMHE's Global Health Data Exchange (GHDx) provides a comprehensive catalog of surveys, census data and vital statistics, and in keeping with their core principles, IMHE shares their datasets and findings broadly in a spirit of open collaboration.
To keep pace with evolving applications, IHME needed a modern, shared infrastructure. For containerized workloads, IHME selected Rancher, a Docker container management platform that uses Kubernetes to manage containerized applications. While Kubernetes excels at managing scaled-out container applications, it lacked critical features needed by scientists running existing HPC workloads.

To allow HPC workloads to work seamlessly with Kubernetes, IHME implemented Univa's Navops Command, a solution that plugs into Rancher, extending Kubernetes with advanced workload placement and policy management.

"By using Navops Command and Univa Grid Engine in our Kubernetes environment we have the best of both worlds. We can easily share resources between container and non-container workloads and even create hybrid workflows, paving the way to further enhance our application environment at our own pace."

- Dr. Tyrone Grandison, Chief Information Officer

To access the case study .pdf, complete the form below.

Fields marked with (*) are mandatory.