SAP-C02 Practice Question 19: NAS File Data Migration
AWS Certified Solutions Architect – Professional (SAP-C02) · Test 2 · question 9 of 10
A research institute is moving an analytics platform to AWS. Its 350 TB of research data (about 400 million files) lives on an on-premises NFS NAS, and the target is Amazon EFS, which the rehosted compute will mount. The data centre has a 10 Gbps AWS Direct Connect connection shared with production systems.
The migration plan has these constraints. The transfer must not use more than 3 Gbps of the link at any time. Researchers keep writing to the NAS for the six weeks of the migration, so repeated passes must copy only what has changed since the previous pass, keeping the final cutover pass small. The team must be able to prove that every file arrived intact, with source and destination checksums compared. And the team doesn't want to write or operate copy scripts or build its own transfer servers.
Which approach should the migration team use?
Show answer and explanation
Answer:
AWS DataSync is the managed, online service for moving file data from on-premises storage such as NFS into AWS storage services including Amazon EFS. Each constraint maps to a built-in task setting. A bandwidth limit caps the throughput each execution uses, and it can be changed on a running execution, so the 3 Gbps ceiling on the shared link is enforced by the service. With the transfer mode set to transfer only data that has changed, DataSync copies everything once and then, on later runs, only the data and metadata that differ between source and destination. Repeated passes during the six weeks keep shrinking the delta, and the final pass at cutover is small. DataSync runs checksum verification during the transfer and can also verify at the end, comparing source and destination checksums to prove integrity. Its parallel, multi-threaded transfer engine handles hundreds of millions of files with no scripts to maintain. For 400 million files, the team would likely split the namespace across several tasks.
Option A solves a different problem. S3 File Gateway gives on-premises clients hybrid NFS access to S3; it is not a NAS-to-EFS migration tool. Pointing the workloads at S3 instead of the EFS target changes the target architecture, and nothing in it enforces a bandwidth cap or provides checksum reports for the migration.
Option B can technically work, but it is exactly the self-built transfer fleet and script maintenance the team ruled out. Single-threaded rsync also crawls over hundreds of millions of small files, and nothing centrally enforces the 3 Gbps ceiling across parallel jobs.
Option C uses a protocol endpoint meant for partner file exchange as a bulk migration engine. The team would need to write the change tracking, throttling, and verification themselves, the scripts they didn't want to own.
References: Configuring how to handle files, objects, and metadata (transfer mode); Setting bandwidth limits for your AWS DataSync task