Mountpoint for Amazon S3 now supports a high performance shared cache
You can now use Amazon S3 Express One Zone as a high performance read cache with Mountpoint for Amazon S3. The cache can be shared by multiple compute instances and can elastically scale to any dataset size. Mountpoint for S3 is a file client that translates local file system API calls to REST API calls on S3 objects. With this launch, Mountpoint for S3 can cache data in S3 Express One Zone after it’s read, making the subsequent read requests up to 7x faster compared to reading data from S3 Standard.
Previously, Mountpoint for S3 could cache recently accessed data in Amazon EC2 instance storage, EC2 instance memory, or an Amazon EBS volume. This improved performance for repeated read access from the same compute instance for dataset sizes up to the size of the available local storage. Starting today, you can also opt in to caching data in S3 Express One Zone, benefiting applications that repeatedly read a shared dataset across multiple compute instances, without any limits on the total dataset size. Once you opt in, Mountpoint for S3 retains objects with sizes up to one megabyte in S3 Express One Zone. This is ideal for compute-intensive use cases such as machine learning training for computer vision models where applications repeatedly read millions of small images from multiple instances.
Mountpoint for Amazon S3 is an open source project backed by AWS support, which means customers with AWS Business and Enterprise Support plans get 24/7 access to cloud support engineers. To get started, visit the GitHub page and product page.