[ntp:questions] Query about NTP accuracy

Andy Yates andyy1234 at gmail.com
Fri May 22 23:57:53 UTC 2009

John Hasler wrote:
> Andy Yates writes:
>> Its up to us to specify what we think the SLA should be - the guide is
>> "as accurate as possible"!
> It would be "possible" to provide each client with a cesium clock, but I
> doubt that your budget extends to that.
> I suggest that you determine the maximum accuracy that the clients can
> utilize and work back from that.
Thanks John
I should have said as "accurate possible using NTP" i.e. a well designed

That's what we intend (testing the NTP not a cesium clock!)- I was
hoping somebody else had done some tests before hand so I can get a
handle on the figures and manage expectations before we get the kit and
possibly influence the design. I've got most of what I wanted from you
guys - last time I did any meaningful benchmarking on NTP was about 8
years ago so I thought it was worth asking. Information from here has
been very useful.


More information about the questions mailing list