Which approach will meet these requirements?
Use an Amazon RDS DB instance. Shut down the instance once the data has been read.
Use Amazon Aurora Serverless. Allow the service to spin resources up and down, as needed.
Use Amazon DynamoDB in on-demand capacity mode.
Use Amazon S3 and load the data from flat files.
Explanations:
Amazon RDS is not cost-effective for static datasets that are only read at startup. The startup time would be slower due to database overhead.
Aurora Serverless incurs higher costs due to its scaling mechanism, which is unnecessary for a static dataset. It also introduces a delay in startup.
DynamoDB is designed for highly scalable, low-latency applications but is overkill for a small static dataset. On-demand capacity mode incurs unnecessary costs.
Storing the data in Amazon S3 as flat files is the most cost-effective and fastest solution for loading static data into cache at startup.
I appraise that the answer is:
Use Amazon S3 and load the data from flat files.