Question

A retailer is migrating 40 TB of on-premises data stored on a self-managed NFS array to AWS. The data feeds a nightly analytics batch job. The company has a 500 Mbps dedicated internet link that is otherwise idle overnight, and a hard deadline of 6 weeks to complete the migration and decommission the array. During migration, the source data continues to change daily, and the final cutover must reflect the latest changes with minimal reconciliation effort. Which approach best meets the requirements at the lowest operational overhead?

Accepted Answer

Deploy AWS DataSync with an agent on-premises, schedule recurring incremental transfers over the existing link, and run a final task immediately before cutover.

Answer

Order a Snowball Edge device, copy the 40 TB once, and ship it to AWS; re-copy any changed files manually before cutover.

Answer

Use AWS DMS with a full-load-plus-CDC task to continuously replicate the NFS files to Amazon S3 until cutover.

Answer

Set up rsync cron jobs on the source array to push files to an S3 bucket via the AWS CLI, syncing nightly until the deadline.