Convert clock drift in ppm into error per poll interval, per day and across a holdover window.
A drift figure in ppm is a fraction: 20 ppm means the clock gains or loses 20 microseconds every second. Multiplying by the poll interval gives the error accumulated between corrections, and by the holdover window gives the error if sync is lost entirely. Log correlation, Kerberos authentication and certificate validation all break on clock skew, and a cheap oscillator at 50 ppm drifts more than four seconds a day once upstream sync is gone.
NTP Drift
drift per interval = ppm × interval ÷ 10^6, so drift in ms = ppm × interval in seconds ÷ 1000.
Poll interval for an accuracy target
maximum poll interval = target accuracy ÷ drift rate
drift per interval = ppm × interval ÷ 10^6, so drift in ms = ppm × interval in seconds ÷ 1000. A drift figure in ppm is a fraction: 20 ppm means the clock gains or loses 20 microseconds every second. Multiplying by the poll interval gives the error accumulated between corrections, and by the holdover window gives the error if sync is lost entirely.
Log correlation, Kerberos authentication and certificate validation all break on clock skew, and a cheap oscillator at 50 ppm drifts more than four seconds a day once upstream sync is gone.
This calculator takes 4 inputs: Oscillator drift, NTP poll interval, Target accuracy, Holdover window without sync. The pre-filled defaults are a realistic starting point — replace them with figures from your own environment for a result you can act on.