Week 11: How does the system fail under pressure?

The Office of Everything's worst day, and identifying which specific design decision from earlier weeks is most responsible for its collapse.

3 May 2027

Every policy examined so far has been tested against ordinary demand, or against one outlier inside an otherwise ordinary day. This lecture asks what happens to the same system on a day where demand surges and something else also goes wrong at the same time.

Outline#

• congestion is a different regime, not just more of the same: a policy that behaves reasonably at normal volume can behave very differently once the arrival rate outpaces the service rate for a sustained period
• compounding failure: the Office’s worst day combines a demand surge (enrolment-variation deadline) with an unrelated system failure (the ticket dispenser jamming, from week 8) at the same time
• tracing the collapse back to a decision, not an event: this lecture argues that the failure was allowed to compound by a choice made earlier in the semester, and asks the Queue Lab to identify which one
• why “the system failed” is not itself an analysis: this course has built the vocabulary (measured wait, chosen policy, honest promise, stuck-versus- slow evidence, audited priority) to say specifically what failed and why
• this is the last lecture before the final redesign: everything examined this semester is now material to defend or replace

Office of Everything file — synthetic case material. Incident summary, the Thursday before the enrolment-variation deadline: ticket volume tripled against a normal day; the dispenser jammed for its usual forty minutes, but this time behind a queue three times longer; the Priority Lane, still specified as in week 6, saw declared urgency rise disproportionately as the queue grew.

The Queue Lab asks you to trace this compounding failure back to one specific decision from an earlier week, and argue why that decision — not the surge, not the jam — is where the design should have been different.