A company needs to use Amazon SageMaker to train a model on more than 300 GB of data. The training data is composed of files that are 200 MB in size. The data is stored in Amazon S3 Standard storage…

Question

Options

• AAmazon Elastic File System (Amazon EFS) file system
• BAmazon FSx for Lustre file system
• CAmazon S3 in fast file mode while using S3 Express One Zone
• DAmazon S3 in fast file mode without using S3 Express One Zone

• A
  4% (1)
• B
  21% (5)
• C
  8% (2)
• D
  67% (16)

Explanation

Option D is correct because SageMaker's fast file mode streams data directly from S3 without requiring a full dataset copy before training begins, and since the data already resides in S3 Standard (which also serves a dashboard), there is no need to migrate or duplicate it to a costlier storage class - making this the lowest-overhead, most cost-effective choice.

Option A (EFS) is wrong because EFS requires copying the 300+ GB dataset from S3 into the file system first, adding both data-transfer costs and higher per-GB EFS storage costs versus S3 Standard.

Option B (FSx for Lustre) is wrong for the same data-duplication reason, and FSx for Lustre carries a significant hourly provisioned-capacity cost that makes it more appropriate when extreme throughput is needed, not merely cost-effectiveness.

Option C (fast file mode + S3 Express One Zone) is wrong because S3 Express One Zone storage costs roughly 7× more per GB than S3 Standard; since the data already lives in Standard and must remain accessible to the dashboard, moving or duplicating it to Express One Zone adds unnecessary expense.

Memory tip: "Fast file = stream it where it stands." If data already lives in S3 Standard, fast file mode lets SageMaker read it in place - no migration, no duplication, no extra tier. Any option that adds a managed file system (EFS, FSx) or a premium storage class (Express One Zone) introduces cost that the question is asking you to avoid.

No community discussion yet for this question.