Question

A retail company runs a 4 TB Microsoft SQL Server transactional database on-premises. The analytics team wants a continuously updated copy of the transactional data landed in Amazon S3 in Parquet format for querying with Amazon Athena, without impacting the production OLTP workload. The initial full load plus ongoing changes (roughly 5,000 transactions per minute) must be replicated with minimal transformation effort and no changes to the source schema. Which approach best meets these requirements?

Accepted Answer

Use AWS DMS with S3 as the target endpoint, configured for full load plus change data capture (CDC), and set the target to write Parquet files.

Answer

Use AWS DataSync to copy the SQL Server data files to S3 on a nightly schedule, then convert them to Parquet with an AWS Glue job.

Answer

Enable SQL Server transactional replication to an Amazon RDS for SQL Server instance, then export snapshots to S3 as CSV.

Answer

Use AWS Snowball Edge to perform the initial 4 TB transfer, then run periodic mysqldump exports uploaded to S3 with a Lambda function.