I'm currently thinking about the various sources that influence time when it is synchronized with the NTP protocol. One of them is the physical layer (OSI-model).
When sending a bit over a cable, what influences the variance in transmission time?
Please note that I am NOT interested in the latency, but the variance. So what makes the time the transmission needs change?
I guess for metal cables it is mostly differences in temperature. Is there anything for a given glass fiber cable that changes the signals need to be transfered over it? Are there any aging processes that change the time?
What about WLAN signals? On which order of magnitude does the time from sending and receiving differ, depending on conditions of the environment (air temperature, maybe increased error rates when many devices are sending?)