AWS Certified Solutions Architect Professional SAP-C02 Practice Question

Caching the hot, read-only query results in an in-memory data store eliminates most round-trips to disk-based storage and delivers microsecond response times. Amazon ElastiCache for Redis natively provides sub-millisecond latency and can serve millions of requests per second, so placing a short-TTL cache in front of the database immediately cuts CPU utilization and meets the < 5 ms p95 requirement with only minor code changes.

Simply scaling the RDS instance (larger class and more IOPS) improves throughput but still involves disk I/O and network hops that rarely reach micro-second latency. Migrating to Aurora and enabling parallel query benefits analytical scans, not small OLTP reads, and read replicas still introduce replica-lag risk. Deploying CloudFront in front of the ALB targets Internet edge-to-origin latency-not the service-to-database path-and cannot cache dynamic API responses reliably, so it does not address the primary performance bottleneck.