Are there any benefits to storing data in DynamoDB vs S3 for use with Redshift?

Question

My particular scenario: Expecting to amass TBs or even PBs of JSON data entries which track price history for many items. New data will be written to the data store hundreds or even thousands of times per a day. This data will be analyzed by Redshift and possibly AWS ML. I don't expect to query outside of Redshift or ML.

Question: How do I decide if I should store my data in S3 or DynamoDB? I am having trouble deciding because I know that both stores are supported with redshift, but I did notice Redshift Spectrum exists specifically for S3 data.

Answer 1

Firstly DynamoDB is far more expensive than S3. S3 is only a storage solution; while DynamoDB is a full-fledge NoSQL database.

If you want to query using Redshift; you have to load data into Redshift. Redshift is again an independent full-fledge database ( warehousing solution ).

You can use Athena to query data directly from S3.