Design Cost-Optimized Architectures

AWS billing has four main levers, and most cost questions are really asking which one fits the workload. There's capacity commitment (Reserved Instances, Savings Plans, Reserved Nodes) for steady usage; interruption tolerance (Spot) for fault-tolerant batch; access-frequency tiering (the S3, EFS, and FSx storage classes) for data that cools off over time; and right-sizing (Compute Optimizer, Trusted Advisor) for resources that are over-provisioned. What the exam is really testing is whether you can pick the right lever from the workload's predictability and tolerance — not whether you can recite discount percentages.

Cost optimization is only the goal when nothing else outranks it. If the scenario carries a hard requirement — availability, latency, durability, or compliance — the correct answer satisfies that first and then optimizes cost within it. The cheapest option that violates the stated constraint is always a distractor.

The cheapest resource is the one you switch off. Right-size over-provisioned instances and volumes, delete unattached EBS volumes and idle load balancers, and schedule non-production environments to shut down outside business hours — then, and only then, layer commitment discounts on the steady-state baseline that's left. Commit before you clean up and you've locked in your waste for one to three years.