[ntp:questions] NTP in LAN with ms accuracy.

Naren P nvap23 at gmail.com
Wed Apr 1 23:28:34 UTC 2015

We are trying to setup NTP in our subsystem in a LAN and facing some
challenging issues:

PC_A and PC_B sync to PC_V which are in the same isolated subnet.
PC_V syncs to PC_E which is outside the subnet but within the same LAN.
PC_A, PC_B, PC_V could restart few times a day.

*NTP Commands used:*
*Sync: *ntpd.exe -g -N -q -c "Sync Config file"
*Drift:* ntpd.exe -g -N -c "Drift Config file"

*NTP Configuration:*
*Sync Config file includes:*
server xxxx iburst burst

*Drift Config file includes:*
server xxxx burst minpoll 3 maxpoll 6 prefer true
tinker step 0
tinker panic 0

*Requirements: *Requirement is to setup NTP on these PCs in such a way that:
1) Sync between PCs at startup happens as soon as possible, lets say within
couple of seconds.
2) Sync between PCs as accurate as possible,  lets say within +/-5ms.
3) If possible, Offset between PCs shall remain as low as possible, lets
say within +/- 5ms.
If not, setup NTP to drift and not step as sudden steps are not acceptable.

1) Sync takes about 8s to complete.
2) The offset could be within +/-1ms, if we are lucky. But sometimes offset
could be within +/- 100ms.
3) The clocks could have an offset of +/- 60ms within few mins after sync.

1) Can we speed up the synching further? Is there a way to reduce this to
few 10s of milliseconds.
2) The issue is that NTP is not deterministic and hence not reliable.
How can we assure that the sync will result in +/- 5ms offset consistently?
3) This seems to defeat the whole purpose of synching if the offset cannot
be maintained for atleast few hours.
Is there a way specify so that the offset remains within +/-5ms for few

As per NTP documentation, we should be able to achieve one millisecond
accuracy in LAN.
This seems to be very inconsistent in our testing and our network is not at
all congested during this testing.

Please help us bring together the pieces we might be missing to be able to
use NTP to our needs.

Thank you all.

More information about the questions mailing list