Storage layer decisions are effectively permanent — migrating a high-traffic DynamoDB table or an RDS cluster mid-flight is expensive and risky. Understanding each service's consistency model, scaling behavior, and cost profile upfront prevents both architectural regret and outage-inducing migrations down the line.
What This Pillar Covers
- Amazon S3 — buckets, storage classes, lifecycle rules, versioning, and S3 Select
- RDS and Aurora — Multi-AZ failover, read replicas, parameter groups, and automated backups
- DynamoDB — partition key design, GSIs, on-demand vs. provisioned capacity, DynamoDB Streams
- ElastiCache — Redis vs. Memcached, cluster mode, eviction policies, and the cache-aside pattern
- FSx and Storage Gateway — managed file systems and hybrid cloud storage integration
Who This Is For
Backend engineers and DBAs choosing between AWS storage and database services, and platform teams who need to reason about consistency and scaling trade-offs before committing to a data layer.