Question

A media company is building a new product catalog API for a global streaming service. The backend uses Amazon Aurora PostgreSQL, and product metadata is read extremely frequently but updated only when editors publish changes (roughly a few hundred times per day). The team adds an Amazon ElastiCache for Redis cluster to reduce database load. They require that reads served from the cache always reflect the most recent published metadata immediately after an editor saves a change, while minimizing unnecessary cache writes for products that are rarely read. Which caching strategy best meets these requirements?

Accepted Answer

Use a write-through strategy that updates the cache synchronously on every publish, combined with lazy loading on read misses and a TTL to bound staleness.

Answer

Use lazy loading (cache-aside) alone, populating the cache only on read misses and relying on a long TTL to eventually refresh entries.

Answer

Use write-through caching only, writing every product to the cache on publish and disabling TTL so entries never expire.

Answer

Disable ElastiCache and instead enable Aurora read replicas with a reader endpoint to absorb all catalog read traffic.