Design Cost Optimized Architectures

What this chapter covers

Cost optimization means meeting the requirement at the lowest sensible price, not simply choosing the cheapest service. The exam gives a workload pattern and asks which design wastes the least money while still staying secure and resilient.

The chapter follows four outline topics: storage, compute, databases and networking. A fifth page covers the tools that show where money goes, because those tools support all four topics.

How the topics fit together

Each area follows the same logic: match the pricing model to how the workload really behaves, then remove waste.

• Storage is cheapest when data moves to colder classes as it ages.
• Compute is cheapest when steady load is committed, flexible load uses Spot, and spiky load uses serverless.
• Databases are cheapest when the capacity mode fits the load shape.
• Network cost is mostly data transfer, which endpoints, caching and placement can cut.

How to use these pages

Each topic page has the same shape, so you can skim for what you need.

• What it is: the idea in plain words.
• How it works: the mechanism, with diagrams and examples.
• What to remember for the exam: the facts to memorise.
• Common mistakes: the classic traps.
• Check yourself: three questions to test your understanding.

Focus on the trigger words in each scenario, such as unpredictable, steady, interruptible, rarely accessed and global users. They point straight to the pricing model.