Introducing Amazon S3 Files: Seamless File System Access for Amazon S3
- Amazon Web Services (AWS) has launched Amazon S3 Files, a new file system designed to connect AWS compute resources directly to data stored in Amazon Simple Storage Service...
- In standard object storage, data is typically handled as discrete objects that must be replaced entirely to be edited.
- S3 Files is built using Amazon Elastic File System (Amazon EFS) and delivers latencies of approximately 1ms for active data.
Amazon Web Services (AWS) has launched Amazon S3 Files, a new file system designed to connect AWS compute resources directly to data stored in Amazon Simple Storage Service (Amazon S3). The service allows S3 buckets to be accessed as high-performance file systems, removing the traditional technical divide between object storage and file systems.
In standard object storage, data is typically handled as discrete objects that must be replaced entirely to be edited. S3 Files changes this dynamic by enabling users to read, write and organize S3 data using standard file and directory operations, providing full file system semantics and low-latency performance.
Technical Architecture and Performance
S3 Files is built using Amazon Elastic File System (Amazon EFS) and delivers latencies of approximately 1ms for active data. It supports Network File System (NFS) v4.1+ operations, including the creation, reading, updating, and deletion of files.
The system operates by presenting a traversable view of S3 bucket objects as files. When a user navigates directories or opens files, the associated metadata and contents are placed onto high-performance storage. S3 Files utilizes an on-demand loading mechanism, meaning it does not duplicate the entire dataset but loads contents as needed.
To optimize throughput and costs, the service employs different data handling paths. Files requiring low-latency access are served from high-performance storage, while files requiring large sequential reads are served directly from Amazon S3. For byte-range reads, the system transfers only the requested bytes to minimize data movement.
The service also includes intelligent pre-fetching to anticipate data access needs. Users have granular control over synchronization, allowing them to decide whether to load full file data or metadata only to optimize for specific access patterns.
Integration and Compatibility
Amazon S3 Files can be attached to a wide range of AWS compute resources, enabling data sharing across clusters without the need for data duplication. Supported resources include:
- Amazon Elastic Compute Cloud (Amazon EC2) instances
- Containers running on Amazon Elastic Kubernetes Service (Amazon EKS)
- Containers running on Amazon Elastic Container Service (Amazon ECS), including those on AWS Fargate
- AWS Lambda functions
The system integrates with AWS Identity and Access Management (IAM) for access control and encryption. Permissions can be managed at both the file system and object levels using identity and resource policies. S3 Files utilizes POSIX permissions for files and directories, verifying user IDs (UID) and group IDs (GID) against metadata stored in the S3 bucket.
Security is handled via TLS 1.3 for data in transit. For data at rest, the service uses Amazon S3 managed keys (SSE-S3) or customer-managed keys via AWS Key Management Service (AWS KMS).
Use Cases and Service Positioning
AWS positions S3 Files as a solution for interactive, shared workloads that require frequent data mutation. This includes training machine learning (ML) models, building agentic AI systems that rely on file-based Python libraries and shell scripts, and running production applications that require a file system interface.
The service differs from other AWS storage options based on the specific needs of the workload. While S3 Files is intended for interactive access to S3 data, Amazon FSx is designated for workloads migrating from on-premises NAS environments. Specifically, Amazon FSx for Lustre is aimed at high-performance computing (HPC) and GPU cluster storage, while other variants like FSx for NetApp ONTAP, OpenZFS, or Windows File Server provide specific file system capabilities.
Availability and Pricing
As of April 8, 2026, S3 Files is available in all commercial AWS Regions.
The pricing model is based on three primary factors: the volume of data stored within the S3 file system, charges for small file read and all write operations, and the cost of S3 requests generated during the synchronization process between the file system and the S3 bucket.
Monitoring for the service is integrated into Amazon CloudWatch for drive performance and updates, while Amazon CloudTrail is used for logging management events.