Please log in to watch this conference skillscast.

The question is no longer whether to migrate HPC workloads to the cloud, but where and how to start. While the industry has been discussing migrating workloads to the cloud since the launch of AWS in 2006, concerns about latency, regulatory issues and lock-in have slowed HPC adoption until recently. In the past year, new application deployment models, cloud-service offerings, and advances in management software have emerged to help serve as a valuable tool for HPC workloads looking to transition to the cloud. Containers in particular have been distinguished as breakthrough technology for IT professionals looking to move HPC workloads the cloud. In the past, HPC applications often relied on pre-requisites, such as specialized math libraries or various MPIs, making it tedious to move workloads to cloud-provider machine images without customization. With continued advances in virtualization and container technologies, like Docker and Kubernetes, HPC applications can be packaged along with dependencies, ultimately speeding deployment and side-stepping challenges related to version control and configuration management.

Fritz will share with you how IT professionals can leverage container technologies to move HPC workloads to the cloud and how containers can be used to build seamless hybrid cloud environments that support a wide variety of HPC workloads, provide a clear path to future cloud native environments and enable capabilities like cloud-bursting.

YOU MAY ALSO LIKE:

• Espresso Patronum: The Magic Of The Robot Pattern (SkillsCast recorded in October 2019)
• Effective Reactive Architecture (SkillsCast recorded in October 2019)