Question

A developer is building a product catalog service backed by an Amazon RDS MySQL database. Read traffic is very high and heavily skewed toward a small set of popular products, causing elevated database CPU. The team adds an Amazon ElastiCache for Redis cluster and wants a caching strategy where data is loaded into the cache only when it is first requested, minimizing memory used for products that are never read. How should the developer implement the read path?

Accepted Answer

On each read, query the cache first; on a cache miss, read from RDS, write the item into the cache with a TTL, then return it (cache-aside/lazy loading).

Answer

Write every product to the cache at the same time it is written to RDS so the cache always contains the full catalog (write-through).

Answer

Point the application directly at ElastiCache as the primary data store and periodically flush writes to RDS in batches.

Answer

Enable a DynamoDB Accelerator (DAX) cluster in front of RDS to transparently cache read queries.