AWS Certified Solutions Architect Professional SAP-C02 Practice Question

The correct option is to deploy RDS Read Replicas and modify the application logic. This approach addresses both requirements effectively. General read traffic is offloaded to the read replicas, which alleviates the performance bottleneck on the primary instance. The critical requirement for read-after-write consistency is met by implementing an application-level pattern where, for a limited time after a user performs a write, their subsequent read requests are directed to the primary database instance, guaranteeing they see their own updates.

Incorrect options:

• Migrate to Aurora: While migrating to Amazon Aurora and using the reader endpoint is a powerful performance strategy, it does not guarantee immediate read-after-write consistency. Aurora replicas have very low replication lag (often sub-100ms) due to the shared storage architecture, but replication is still asynchronous. This means there is still a small window where a read could be served stale data from a replica, which violates the strict consistency requirement.
• Implement ElastiCache: Adding a caching layer like ElastiCache is a common pattern for improving read performance. However, a simple write-through strategy does not inherently solve the read-after-write consistency problem for the user. It adds another system to manage and requires complex logic to ensure that a user's reads are not served stale data, especially in a distributed application environment.
• Create a cross-region replica: Using a cross-region read replica with Route 53 latency-based routing is a pattern designed for disaster recovery and for reducing read latency for a globally distributed user base. This solution would increase, not decrease, the replication lag due to network distance between regions, making the eventual consistency problem more pronounced and failing to meet the immediate consistency requirement.