Mount S3 as a Filesystem — No SDK, No Data Duplication.

On April 7, 2026, AWS launched Amazon S3 Files — making Amazon S3 the first and only cloud object store to provide fully-featured, high-performance file system access to your data.

Before S3 Files, applications that needed to work with S3 data as files had two options: make thousands of API calls with S3-specific logic in your code, or duplicate your data to a separate filesystem and keep it in sync. Both added complexity.

S3 Files eliminates that tradeoff. You mount an S3 general purpose bucket on your EC2 instance, Lambda function, or ECS/EKS container using the NFS v4.1+ protocol — and your application just does open(), read(), write(). Your data never leaves S3. Built on Amazon EFS, it delivers ~1ms latency for actively accessed data, with bidirectional sync between the filesystem and bucket.

In this video, I walk through the full setup — from creating the file system and mount target, to reading and writing S3 objects with standard Linux commands like cat and ls, to verifying bidirectional sync in real time.

📺 Watch the full 12-minute demo:
AWS S3 Files — Mount S3 as a Filesystem | BeSA Cloud Academy