I’ve been researching how online time services sync over HTTPS, and the results are surprising.
Due to network latency & asymmetry, even widely trusted time sources can drift by ±50ms or more.
HTTPS-based synchronization has inherent protocol limitations that prevent true atomic accuracy.
Engineers on NTP forums have confirmed that half of RTT (Round Trip Time) is the firm limit—meaning most public clocks are way less accurate than we assume.
This raises some serious questions:
1⃣ How bad is the drift in real-world applications?
2⃣ Is there a way to make HTTPS time synchronization truly accurate?
3⃣ Are industries (finance, security, cryptography) unknowingly relying on bad time data?
I’m curious if others have tested this or if this is an even bigger problem than we think.
Some engineers are experimenting with ways to push accuracy below 10ms, but there’s no widespread solution yet. Thoughts?
For most civilian purposes, time sync is done with NTP, which, using Marzullo's algorithm, can reduce error to around 0.5% of RTT, nothing like "half of RTT". Did you read "0.5%" and thought it meant "50%"?
Is your complaint about something specific? This problem seems imaginary.
> Engineers on NTP forums have confirmed that half of RTT (Round Trip Time) is the firm limit
Which engineers? That statement is not true.
> Some engineers are experimenting with ways to push accuracy below 10ms
Which engineers? NTP accuracy is typically around 0.5ms in good conditions.
The use of the ’ smart apostrophe and the weird 1⃣2⃣ makes me think this post was written by a very confused AI.