> This raises some serious questions: 1⃣ How bad is the drift in real-world applications?
Maybe ask yourself how applications that are serious about time deal with sync'ing remote sites. Follow that with how many real world internet applications need those levels of accuracy | correlation over distance.
Two examples at the higher end:
* Multiple Gravity sensors across the globe; these need to filter out passing trucks in some US state that don't also occur at the same time in Australia.
* Square Kilometre Array, with radio astronomy dishes and other sensors on opposite sides of the planet.
Both these projects have additional fun quirks; a signal travelling from (say) the heart of the milky way will encounter sensors on a rotating planet that is sometimes orbiting away from the signal and sometimes towards the signal, depending upon time of year.
Both the projects highlighted have been "in the works" since the 1980s. What are the timing requirements in these? What solutions are in play?
Maybe ask yourself how applications that are serious about time deal with sync'ing remote sites. Follow that with how many real world internet applications need those levels of accuracy | correlation over distance.
Two examples at the higher end:
* Multiple Gravity sensors across the globe; these need to filter out passing trucks in some US state that don't also occur at the same time in Australia.
* Square Kilometre Array, with radio astronomy dishes and other sensors on opposite sides of the planet.
Both these projects have additional fun quirks; a signal travelling from (say) the heart of the milky way will encounter sensors on a rotating planet that is sometimes orbiting away from the signal and sometimes towards the signal, depending upon time of year.
Both the projects highlighted have been "in the works" since the 1980s. What are the timing requirements in these? What solutions are in play?