A developer needs microsecond read latencies for frequently accessed product items, and the application cache must be updated whenever products are created, updated, or deleted to ensure consistency. Which solution satisfies these requirements?

The correct solution is to use Amazon DynamoDB with a DynamoDB Accelerator (DAX) cluster. DAX is a fully managed, highly available, in-memory cache for DynamoDB that delivers microsecond read latencies for frequently accessed items. It is designed to be API-compatible with DynamoDB, simplifying integration. DAX also supports a write-through caching strategy, which automatically updates the cache whenever data is written to DynamoDB, ensuring strong consistency for reads and meeting the requirement for cache updates on product changes. Using Amazon RDS with ElastiCache for Redis and a lazy-loading strategy would introduce eventual consistency, as the cache is only updated on read misses, not immediately on writes. Using DynamoDB with an in-memory application cache and lazy loading also suffers from eventual consistency and requires manual cache invalidation logic. Using Amazon RDS with DAX is incorrect because DAX is specifically designed for DynamoDB, not RDS.