Persistent File Storage for Modern Applications

Amazon EFS attachments are configured in the application metadata,
such as the Amazon Elastic Container Service (Amazon ECS) task definition or the Kubernetes persistent
volume, including connectivity, so developers can focus on their applications,
not infrastructure.

Amazon ECS, Amazon Elastic Kubernetes Service (Amazon EKS), AWS Fargate, and Amazon EFS are fully elastic, scale up and down rapidly based on demand. 

Amazon ECS, Amazon EKS, AWS Fargate, AWS Lambda, and Amazon EFS are regional services. You can build applications that span multiple availability zones, with automatic failover. 

AWS provides a secure cloud computing environment. Access to Amazon EFS can be controlled based on the IAM role of the Amazon ECS task.

Pay only for the storage and compute you use. Amazon EFS scales on demand from zero to petabytes with no disruptions, growing and shrinking automatically as you add and remove files, and Amazon ECS, Amazon EKS, and AWS Fargate Cluster Auto Scaling enables capacity to grow and shrink to meet demand.

Web serving and content management systems require shared data access across multiple containerized application instances, data persistence, and data durability. Examples include applications like WordPress and Drupal, that benefit from scaling out to multiple instances for performance and redundancy, and need to share uploads, plugins, and templates.

A stateful microservice is part of a loosely coupled application that needs to remember some things about its state each time it runs, as in the way a weather application remembers your home city. For these modern applications, Amazon EFS is a data foundation, operating in concert with containers and serverless technologies to reliably and consistently deploy to AWS, allowing data to persist application state.

Containers are a way to deploy machine learning (ML) training jobs, inference endpoints, and tools that run consistently in multiple environments. Using containers can solve problems such as consistency, portability, and dependency management that complicate matters for data scientists and developers whose focus should really be on their code, datasets, and training models, not on infrastructure. ML tools like Amazon SageMaker Notebooks, as well as open source tools like Jupyter use Amazon EFS to manage data scientist home directories. Amazon FSx for Lustre can be used for your most compute-intensive ML training jobs where the highest throughput is required.