A company runs a time-sensitive transaction-processing application in Docker containers on VMs. The containers use shared storage for transaction data. Transaction volume is unpredictable. The company cannot modify the application or continue managing the container host VMs. They need low-latency storage that automatically scales throughput as demand rises and must stop administering the Docker hosting environment. How should they migrate to AWS?

The correct solution uses AWS Fargate for Amazon ECS, which eliminates the need to manage EC2 instances, addressing the requirement to stop administering the Docker hosting environment. Amazon EFS provides a scalable, shared file system that automatically scales throughput, meeting the low-latency and unpredictable transaction volume requirements without modifying the application. Mounting EFS directly to Fargate tasks is a standard pattern for shared storage. Migrating to Amazon EKS with Amazon S3 for shared data is incorrect because S3 is object storage, not a file system, and would require significant application modification for transaction data. Using Amazon EBS with Fargate is incorrect because EBS volumes are block storage attached to a single instance, not shared across multiple Fargate tasks. Launching EC2 instances and installing Docker is incorrect as it contradicts the requirement to stop administering the Docker hosting environment.