Amazon S3 Files Complete Guide — Mount S3 Buckets as POSIX File Systems [April 2026 Release]

Amazon S3 Files, announced by AWS on April 7, 2026, is a new feature that lets you mount general-purpose S3 buckets as POSIX-compliant native file systems. You can mount via NFS and perform standard file operations — open, read, write, rename, and lock — directly on S3. It is the first POSIX file system capability for a cloud object store, bringing a paradigm shift to legacy apps and AI/ML pipelines.

The conventional wisdom was that S3 is an object store, not a file system. Applications requiring POSIX compliance had to rely on Amazon EFS, FSx, or FUSE-based tools like s3fs, introducing performance overhead and operational complexity. S3 Files eliminates this constraint entirely. - Legacy app support: Run existing file-system-dependent applications on S3 without code changes - AI/ML pipelines: Read large training datasets directly from S3 as files with native performance - Cost optimization: Replace expensive EFS storage with S3's low-cost object pricing

S3 Files operates through a high-performance cache layer built on Amazon EFS, optimizing behavior based on access patterns. - Small files (under 128 KB): Served from the cache layer in milliseconds — ideal for frequently accessed metadata and config files - Large files (1 MB or more): Streamed directly from S3 — handles sequential reads of training datasets and video files - Writes: Batched and synced to S3 approximately every 60 seconds. Use the flush API when strong consistency is required - Concurrent access: The same data is accessible via both NFS mounts and S3 API simultaneously

S3 Files pricing (reference values for us-east-1, as of April 2026):

Cost optimization tip: For a 1 PB bucket with only 1 TB of active data, the premium cache rate ($0.30/GB) applies only to that 1 TB. The remaining 999 TB is charged at the standard S3 rate ($0.023/GB-month), resulting in substantial savings compared to storing everything on EFS.

Here is how to mount S3 Files on an EC2 instance via NFS.

# 1. Enable S3 Files mount target (AWS Console or CLI)
aws s3 create-mount-target \
  --bucket my-s3-files-bucket \
  --subnet-id subnet-xxxx \
  --security-groups sg-xxxx

# 2. Install NFS client
sudo yum install -y nfs-utils   # Amazon Linux
sudo apt-get install -y nfs-common  # Ubuntu

# 3. Mount
sudo mkdir -p /mnt/s3files
sudo mount -t nfs \
  -o nfsvers=4.1,rsize=1048576,wsize=1048576,hard,timeo=600,retrans=2 \
  <mount-target-dns>:/ /mnt/s3files

# 4. Persist across reboots (/etc/fstab)
echo "<mount-target-dns>:/ /mnt/s3files nfs4 nfsvers=4.1,rsize=1048576,wsize=1048576,hard,timeo=600,retrans=2 0 0" | sudo tee -a /etc/fstab

S3 Files shines in the following scenarios:

S3 Files is available in 34 regions at general availability (GA). This includes the Tokyo (ap-northeast-1) and Osaka (ap-northeast-3) regions. The underlying EFS cache layer is deployed in a Multi-AZ configuration, providing resilience against single-AZ failures.

Key constraints to understand before adopting S3 Files: - Concurrent mount targets: Up to 128 mount targets per bucket (subject to regional limits) - Write consistency: Writes are synced to S3 approximately every 60 seconds by default. Explicit flush is required for immediate propagation - Metadata consistency: Changes via NFS may take up to a few seconds to appear via the S3 API - File locking: Advisory locks are supported. Mandatory locks are not - Max file size: Follows S3's 5 TB per object limit - Lambda: Only accessible from within the same VPC

Q1. Can I use S3 Files with my existing S3 bucket? Yes. Simply create a mount target against your existing general-purpose S3 bucket. No changes to existing data are required. Q2. Can Lambda or ECS containers use S3 Files? Yes. Lambda functions and ECS tasks within the same VPC can mount via NFS using standard configuration. Q3. How does performance compare to EFS? For hot data accessed frequently, performance is comparable to EFS. Cold data streamed directly from S3 may have higher first-access latency. Q4. How can I estimate costs? The key is understanding your active (hot) data volume and access patterns. AWS has added S3 Files to the AWS Pricing Calculator for monthly cost simulation. Q5. How is security managed? Access control is handled via IAM policies, VPC security groups, and S3 bucket policies. In-transit encryption (TLS) and at-rest encryption (SSE-S3 / SSE-KMS) are both supported. Q6. Can I combine S3 Files with S3 Intelligent-Tiering or Glacier? S3 Files works alongside standard S3 lifecycle policies. However, objects moved to Glacier or Deep Archive must be restored before they can be accessed via the mount. Q7. What are the benefits of migrating from s3fs-fuse? As a native AWS integration, S3 Files eliminates FUSE layer overhead, significantly improving performance and stability. It is also a fully managed service, reducing operational burden.

Amazon S3 Files opens powerful new options for cloud infrastructure. However, designing the integration, optimizing costs, and configuring security require specialized expertise. Oflight provides end-to-end support for cloud storage design, deployment, and optimization — including S3 Files. From AI/ML pipeline infrastructure to legacy app cloud migration, we tailor architecture proposals to your specific challenges. Learn more about our cloud infrastructure services