An on-premises application writes thousands of images nightly to a mounted NFS share. After migration the application will run on EC2 with an Amazon EFS mount. A Direct Connect link exists. Prior to cutover, what is the most operationally efficient way to replicate newly created on-premises images to the EFS file system?

The correct option leverages AWS DataSync, which is purpose-built for high-performance data transfer between on-premises storage and AWS services like EFS. Using a DataSync agent on-premises directly accesses the NFS share. Configuring DataSync to send data over Direct Connect via a private VIF and PrivateLink ensures secure, private, and efficient transfer to the EFS file system. A scheduled DataSync task provides the required periodic replication. Incorrect options: Using aws s3 sync to S3 and then Lambda to EFS introduces unnecessary intermediaries and complexity, potentially impacting performance and increasing operational overhead for thousands of images. AWS Storage Gateway File Gateway is designed for file-based access to S3, not direct replication to EFS, and would require additional steps to move data from S3 to EFS. Sending DataSync data to S3 and then using Lambda to EFS is less efficient than directly transferring to EFS, adding latency and cost with the S3 intermediary.