By 2020, the company migrated all of its microservices-based applications and data pipelines to container platforms, where they currently run in production.
“We have made the container-first approach as an architecture standard across R&D technologies at our company,” the DevOps engineering manager said. “We also started deploying our AI/ML training models onto containers, and all this work is happening on our (Mirantis Kubernetes Engine) platform.”
Transitioning to Hybrid Cloud
In 2020, the pharma began expanding their R&D data platform beyond on-premises infrastructure. Utilizing Docker images as immutable artifacts in their build process made this flexibility possible.
“As part of our R&D platform’s hybrid, multi-cloud journey, we started enabling container and Kubernetes-based platforms on public clouds,” the DevOps engineering manager explained. “Now, going into 2021 and the future, enabling our R&D users to easily access data and applications in a platform-agnostic way is very crucial for our success,” he said.
Results
Early in their container journey, the pharma experienced clear benefits with Mirantis Kubernetes Engine. “By migrating all our web services into containers, we not only achieved horizontal scalability for those specific services, but also saved more than 50% in support costs for the applications which we have migrated,” the DevOps engineering manager said.
Additionally, with edge nodes on demand, prototyping has become significantly faster. Some tasks that previously required 3-6 months now take only a few minutes.
With a hybrid, multi-cloud platform in place, scientists at the pharma now have greater flexibility to conduct large-scale experiments within a world-leading data and computational environment, using a broader range of toolkits and technologies. The company has successfully implemented containers for artificial intelligence and machine learning and is now looking to use MIrantis Kubernetes Engine to securely deliver virtualized data in containers or Kubernetes volumes to scientists on demand.