In the optimized data transfer portion of cost optimization strategies chapter, is it safe to assume that Caching engines also save costs on data transfer? If I employ a Caching engine in front of a 10 TB DynamoDB table, I should be able to save on costly table scans coming from repetitive queries. Is that a safe assumption?
Yes, a caching layer in front of DynamoDB can save on table scans and thus more efficent use of RCU/WCU. That’s exactly the use-case for DAX. But, DAX and other caching engines do have a cost as well, so you have to be sure to do your estimations. There is no fixed solution for cost optimization.