S3 storing JSON vs DynamoDB

Question

Taking into consideration that DynamoDB is quite more expensive than S3...

Why not to store JSON files in S3 instead using DynamoDB as a store?

One disadvantage of this approach could be querying, filtering or even paging. But let's say that the system is very simple and it only queries by id. The id could be the name (or key) or the file.

Another point could be concurrency. But let's say that users only access/write their on data.

Is there any other scenario or fact which will make S3 a really bad choice?

Answer 1

I agree if the below features provided by DynamoDB are ruled out ie.

1. Concurrency
2. Indexing (for faster access)
3. Other features like secondary indexes

S3 can be used, as it eventually stores objects as key-value pairs