Amazon SageMaker HyperPod now supports configurable cluster storage
Today, AWS announces the general availability of configurable cluster storage for SageMaker HyperPod cluster instances, which enables customers to provision additional storage for model development. This launch allows you to centrally automate the provisioning and management of additional Elastic Block Store (EBS) volumes for your cluster instances. With configurable cluster storage, you can easily integrate additional storage capacity across all your cluster instances, empowering you to customize your persistent cluster environment to meet the unique demands of your distributed training workloads.
Cluster storage on SageMaker HyperPod enables customers to dynamically allocate and manage storage resources within the cluster. Organizations can now scale their storage capacity on-demand, ensuring they have sufficient space for Docker images, logs, and custom software installations. This feature is particularly beneficial for foundation model developers working with extensive logging requirements and resource-intensive machine learning models, allowing them to effectively manage and store critical assets within a secure and scalable environment.
You can use configurable cluster storage with SageMaker HyperPod in the following AWS Regions: US East (Ohio), US East (N. Virginia), US West (Oregon), Asia Pacific (Seoul), Asia Pacific (Singapore), Asia Pacific (Sydney), Asia Pacific (Tokyo), Europe (Frankfurt), Europe (Ireland), and Europe (Stockholm).
To learn more, see the following list of resources: