Skip to main content

Capstone Thread — A Defensible Reliability Decision

Carry this module into your capstone by writing a short decision record that connects three claims: detector precision, retry reliability, and measured performance. Use the supplied synthetic data first; production use requires new evidence for its assumptions.

Deliverable​

Attach your Lab A count table and stopped-retry derivation, Lab D workload comparison, and experiment-design memo. Define the target population, event, stopping rule, and cost unit for each number. Show how a shared outage changes the value of retries, and how a changing workload can reverse an aggregate comparison.

Review questions​

  • What do the calculations establish exactly under the model?
  • Which assumptions were stipulated, measured, or left unknown?
  • Which additional observation would most affect the decision?
  • What result would cause you to stop, revise the model, or collect more data?

Completion​

Meet PS2, PS3, PS7, and PS8 in the assessment rubric, retain a reproducible report, and defend one changed assumption without reading the solution. This is a foundation for later reliability and benchmarking work; it does not replace operating or reviewing a real system.