Theory 44/Society, Computation & Engineering /Security engineering
Rotation Follows a Detection Ratio
Rotation is scheduled by policy. The right interval is a ratio of two clocks.
The optimal interval for cryptographic key rotation is set by the ratio of exposure time to breach detection time rather than by a fixed schedule, so a system that detects compromise quickly can safely rotate less often.
01The evidence
Breach dwell times vary enormously between organisations, and fixed rotation schedules are applied regardless of detection capability.
02The tension
Rotation policy is set by compliance calendars, which predict risk from the interval and ignore the detection time that determines how long a compromised key is useful.
03The mechanism
A compromised key is useful until it is rotated or the breach is detected, whichever comes first. The expected loss depends on the minimum of the two, so the ratio sets the benefit of rotating.
That makes the ratio the determinant, and predicts the optimal interval should lengthen as detection improves.
04What this is not
Fixed schedules ignore detection. This sets the interval from the ratio, which varies by organisation.
05Predictions
- P1Optimal interval should scale with detection time.
- P2Fast detection should permit longer intervals.
- P3Loss should track the ratio.
06Refutation
Optimal interval independent of detection time would favour fixed schedules.
07What it changes
It would make rotation policy organisation-specific.