Amazon announces SOCI-indexed images to speed SageMaker jobs
Amazon has introduced Seekable OCI (SOCI) indexing for its SageMaker AI training and processing jobs. The feature lets SageMaker lazy‑load container images, fetching only the files needed to start the container and streaming the rest in the background. This removes the full image download from the critical path, allowing jobs to begin in seconds instead of minutes.
Key points
- SOCI indexing cuts SageMaker image download time from minutes to seconds
- Adoption requires only a –soci image tag or a custom SOCI index
GPU idle time during image pulls dropped by up to 90 %.
To adopt SOCI indexing, users can switch to pre‑indexed AWS Deep Learning Containers that carry a –soci suffix, or build an index for a custom image using the SOCI snapshotter. No changes to job definitions or data pipelines are required, and the feature is available in all AWS regions.
Announcing SOCI-indexed images for faster SageMaker AI jobs
AWS Machine Learning Blog · 27 October 2026
Loading the full article…
This text was published by AWS Machine Learning Blog and written by Kanwaljit Khurmi. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Enterprise & Industry
All →- Oracle launches Media Fabric AI platform for media production and streaming · 1 src
- Oracle integrates Snowflake AI Data Cloud into OCI as first-party service · 1 src
- Oracle launches AI Security Mesh with OpenAI cyber models for enterprise risk testing · 1 src
- Microsoft report reveals AI driving cyber threats · 1 src
- Invesco links portfolio tools to Anthropic’s Claude for financial advisors · 1 src
Comments
via GitHub Discussions