Or use what PTP does.
en.m.wikipedia.org/wiki/Precision_Time_Protocol
Which is a more accurate protocol than NTP.
en.m.wikipedia.org/wiki/Network_Time_Protocol
Comment on It's the law!
mmddmm@lemm.ee 1 week agoThere’s classics like measuring how long it takes to send a network packet from one device to another
That one is on your clocks quality, not on physics. People do it all the time.
Probably on equipment that is orders of magnitude more expensive than yours, but the post isn’t about costs.
towerful@programming.dev 1 week ago
koper@feddit.nl 1 week ago
That’s just shifting the problem. There is no known way to reliably sync remote clocks except by sending packets and assuming the round-trip time is symmetrical. This is a known problem in physics: en.wikipedia.org/wiki/One-way_speed_of_light
psud@aussie.zone 1 week ago
Sure, but if we take it as true that light speed is the same in every direction – which is perfectly consistent with everything ever measured – you can measure speed between two endpoints using two atomic clocks and a synchronised experiment, with corrections for the relativistic effect of moving the clocks to the different places
koper@feddit.nl 1 week ago
This is the crucial assumption, that to my knowledge hasn’t been proven or disproven. Because the alternative, light goes faster in one particular direction, is also perfectly consistent with everything. And if you’re moving atomic clocks, correcting for time dilations requires you to make assumptions about the one-way speed of light (which we only know from measuring roundtrip times)