Amazon S3, 2017: one command and a region of services

What you'll be able to do

1. 1Predict which requests still succeed when one workload overloads a shared pool, and when it overloads only its own partition.
2. 2Explain how sorting dependencies into critical and optional lets a page keep its core job when an optional dependency fails, and why waiting on an optional call can still take the page down.
3. 3Predict what users see and which requests still succeed when a given dependency fails, with and without a degraded mode.
4. 4Choose a response for each dependency's failure — hide the feature, serve cached or default data, fail open or fail closed — and avoid fallbacks that are rarely exercised or move the load elsewhere.
5. 5Predict whether the zones left after a failure can carry the load, from how many zones there are and how busy each one runs before the loss.
6. 6Choose how many zones to run, how much spare capacity to keep in each, and whether to keep traffic inside a zone or spread it across zones.

Inside the lesson

steps

7

questions

7

simulations

3

You guess first, then run the system and see if you were right.

Step 1 · The setup