Bottlerocket now supports Elastic Fabric Adapter for AI/ML and HPC workloads

Today, AWS announces the introduction of Elastic Fabric Adapter (EFA) support for Bottlerocket, the Linux-based operating system purpose-built for hosting containers, with a focus on security, minimal footprint, and safe updates. An EFA is a network device that you can attach to your Amazon EC2 instance to accelerate Artificial Intelligence (AI), Machine Learning (ML), and High Performance Computing (HPC) applications.

EFA support allows customers to use Bottlerocket with instance types such as G4, G5, P4, and P5 (see full list of the supported instance types). When an EFA device is detected, Bottlerocket automatically reserves huge pages and updates the locked memory limits for containers, ensuring the best possible EFA performance. This integration provides customers with a secure, efficient, and high-performance container hosting environment for their most demanding computational tasks.

EFA support is available on all first-party Bottlerocket Amazon Machine Images starting with version 1.28.0 and is accessible in all commercial AWS Regions and AWS GovCloud (US) Regions. To get started with Bottlerocket, see the Bottlerocket User Guide. You can also visit the Bottlerocket product page and explore the Bottlerocket GitHub repository for more information.