AWS Blurs the Line Between Object Storage and File Systems

Amazon Web Services has introduced a new service, S3 Files, that effectively turns S3 buckets into high-performance file systems. This move addresses a long-standing division in cloud architecture, where teams had to choose between the scale and durability of object storage and the interactive, shared access of traditional file systems.

The service allows any general-purpose S3 bucket to be mounted directly to Amazon EC2 instances, containers, or Lambda functions using a standard NFS interface. Changes made through the file system are automatically synchronized back to the underlying S3 objects, and data can be shared across multiple compute resources without duplication. AWS achieves this by leveraging its Elastic File System (EFS) technology under the hood, delivering latencies around one millisecond for active data.

For business leaders, the implication is a potential simplification of data infrastructure. S3 can now act as a unified data layer for workloads that previously required separate storage solutions, from machine learning training and AI agent development to legacy applications expecting a file interface. The service intelligently manages data placement, keeping frequently accessed files on high-performance storage while serving large sequential reads directly from S3 to optimize costs.

S3 Files is available now across all commercial AWS regions. Pricing is based on the amount of data stored in the file system, file system operations, and related S3 request costs. This release signals a strategic effort by AWS to consolidate its storage narrative, reducing the need for complex data movement and synchronization between different storage tiers.

Source: AWS