Managed File Systems - High-Performance Shared Storage Achieved with Amazon FSx and EFS
This article explains how to build / operate managed file systems making use of Amazon FSx and Amazon EFS.
The Need for Cloud File Systems and AWS Options
Many enterprise applications require a shared file system that can be accessed simultaneously from multiple servers and containers. File-based storage still plays an important role in content management systems, machine learning training data, big data analytics workloads, sharing of home directories, and more. On premises, you operate NAS appliances from NetApp or Dell EMC, Windows file servers, and the like, but capacity planning, hardware maintenance, backup management, and performance tuning take effort. AWS offers two managed file storage services, Amazon EFS (Elastic File System) and the Amazon FSx family, which reduce this operational burden. EFS is an NFS file system for Linux workloads, and FSx offers four file system types: Windows (FSx for Windows File Server), high-performance computing (FSx for Lustre), NetApp ONTAP (FSx for NetApp ONTAP), and OpenZFS (FSx for OpenZFS).
Features and Use Cases of Amazon EFS
Amazon EFS is a serverless, fully managed NFS file system. Capacity automatically expands / shrinks to match the amount of data, and no advance provisioning is needed. It can be mounted simultaneously from many EC2 instances, ECS tasks, and Lambda functions, and it complies with the NFS v4.1 protocol. EFS storage classes include Standard, Infrequent Access (IA), and others, and you choose the file system type from Regional and One Zone. EFS Intelligent-Tiering automatically moves files between storage classes based on access patterns and keeps storage costs down. EFS is also easy to combine with serverless architectures, and it supports direct mounting from Lambda functions. In Bursting Throughput mode, while you have burst credits, you can raise throughput up to 100 MiB/s per 1 TiB of Standard storage (minimum 100 MiB/s). If you select Provisioned Throughput mode, you can secure the throughput you need regardless of storage capacity.
Making Use of the Amazon FSx Family
Amazon FSx offers four file system types suited to different purposes. FSx for Windows File Server supports the SMB protocol and has Windows-native features such as Active Directory integration, DFS namespaces, and shadow copies. It is a candidate migration target from on-premises Windows file servers, and you can carry over existing group policies and access permissions. FSx for Lustre is a file system for high-performance computing (HPC) and machine learning training, and it is used for processing that requires high throughput and IOPS. Through transparent integration with S3 buckets, you can directly access datasets on S3 as a high-speed file system. The following is a CLI example that creates an FSx for Lustre file system. aws fsx create-file-system --file-system-type LUSTRE --storage-capacity 1200 --subnet-ids subnet-0123456789abcdef0 --lustre-configuration DeploymentType=PERSISTENT_2,PerUnitStorageThroughput=250,DataCompressionType=LZ4 FSx for NetApp ONTAP supports migration from on-premises NetApp environments, and you can use ONTAP features such as SnapMirror, FlexClone, and data deduplication on AWS. FSx for OpenZFS, as a general-purpose file system with snapshots, clones, and data compression, is suited to development environments and backend storage for databases.
Pricing Comparison of FSx and EFS
EFS Standard is about $0.30 per GB per month, and EFS IA is about $0.016 (as of August 2026). FSx for Lustre is about $0.14 per GB per month (SSD), and FSx for Windows is about $0.13 (SSD). FSx for ONTAP is about $0.125 (SSD), and with tiered storage, you can automatically move infrequently accessed data to S3. You select a file system according to workload characteristics (POSIX compatibility, Windows compatibility, high throughput) and keep costs down with lifecycle management.
Key Points for Selecting a File System
The choice of a managed file system is determined by the protocol the application requires and the characteristics of the workload. If you want to access it with a standard sharing method from Linux-family servers and containers and use it without worrying about capacity, a type that scales automatically is an easy-to-handle choice. If a Windows application requires integration with Windows' standard sharing method and authentication infrastructure, you choose a type that supports them. For uses that demand high throughput in large-scale computing, a type for high-performance computing is suitable. Rather than deciding uniformly, you determine what fits the workload based on the connection method / performance requirements / affinity with the existing environment.
Summary - Guidelines for Choosing a Managed File System
You use different services according to the purpose: EFS for Linux workloads, FSx for Windows File Server for Windows environments, FSx for Lustre for HPC and ML, and FSx for NetApp ONTAP for NetApp migration. Automatic movement between storage classes with EFS Intelligent-Tiering, data access through the S3 integration of FSx for Lustre, and Active Directory integration of FSx for Windows File Server are each features that help in their respective use cases. It is important to judge the choice of a file system along three axes: protocol requirements, performance requirements, and compatibility with the existing environment.
References (Official AWS Resources)
The primary sources for this page are the official AWS website and documentation. Check the official pages below for the latest specifications and pricing.
If this page and the official documentation disagree, treat the official documentation as authoritative.