Capability
Boulder · New Mexico · United Kingdom
The problem
Why this is hard
Holdover is what a system does after its reference disappears. Every timing product claims some of it; the question is always how long, and how gracefully. A quartz oscillator left on its own drifts quickly, and drift is not a benign failure: the equipment does not stop, it goes on producing answers that are increasingly wrong while continuing to report that it is healthy. Operators handle that today by shortening outage tolerances and adding receivers, which is to say by paying for the same weakness twice.
The answer
What the sensor does
An atomic reference on the board extends holdover from the length of a fault to the length of a campaign, without a rack, a temperature-controlled enclosure or a dedicated power budget. The interesting consequence is architectural rather than numerical: once a node can hold on its own, the receiver stops being a critical component and becomes a convenience, and the system can be designed as though the sky were optional.
Applications
Example applications
Defense
Platforms and installations that must go on operating through a denial event, not merely detect one.
Data centers
Ordering and audit trails that stay coherent across racks and sites when the reference is interrupted.