[ntp:questions] NTP architecture recommendation
dietmar.baur at gmail.com
Fri Aug 17 06:30:54 UTC 2007
On 16 Aug., 23:24, "Richard B. Gilbert" <rgilber... at comcast.net>
> cray74 wrote:
> > Hi,
> > I need to update the current NTP infrastructure of our company, to
> > have more accurate server internal timing (allowable offset during 24
> > hours: < 3 ms). I would appreciate comments on my proposed solution:
> > Topology:
> > HQ localted in central Europe, plus 3 datacenters spread over the
> > world (US, Asia).
> > Current setup:
> > two timeservers in the HQ, which sync to public NTP-servers on the
> > Inet.
> > Problem:
> > the offset between ALL servers needs to be < 3 ms; can't put GPS-
> > clocks in the datacenters (no GPS-signal); using free public NTP
> > servers is not good enough (low accuracy, high jitter because of a
> > number of different Inet-lines with changing latency)
> > My proposal:
> > 1. set up two reference clocks in two separate locations in Europe
> > (where we got offices), each setup consisting of:
> > a) Meinberg Lantime M300 GPS with OCX LQ or even MQ in the office (ie.
> > where I can get the GPS signal)
> > b) Meinberg Lantime M300 MRS (no GPS antenna, just a precise internal
> > clock) in the server room (this one syncs to the first one)
> > 2. the datacenters in the US/Asia will each contain one Meinberg
> > Lantime M300 MRS, syncing to the main 2 stratum 1 servers
> > Question:
> > 1. Will this setup guarantee me the required accuracy?
> > 2. Can the same be achieved with less investment?
> > Thanks for your feedback!
> > Cheers
> > db
> The quality of the source clocks is usually less important than the
> quality (delays and jitter) of the network path between server and
> client. Many of the stratum 1 internet servers have local atomic
> clocks, GPS receivers, LORAN, etc. They KNOW what time it is to within
> a microsecond or better. It's the network path between you and the
> servers that craps all over the quality of the time received!
> I would look harder at getting GPS receivers to work at the data centers
> or, perhaps, a CDMA based solution. You can get a refclock from
> Symmetricom that can lock onto the reference signal of a CDMA base
> station. This should work anywhere that a CDMA cell phone will.
> Your Meinberg M300 MRS is probably capable of keeping time very well;
> the problem will be how to set it with an error less than 3 milliseconds!
Thanks for the clarification, and the hint with a CDMA-based solution.
Of course, ideally, I would try to have a stratum 1 clock available on
the LAN, avoiding the network to mess up my highly accurate time from
the reference clock. Actually this might be the only solution, as cell
phones might not have coverage in our datacenters (for sure not in our
server room in the HQ, which happens to be nuke-safe - we took it
over, wasn't OUR requirement ;-)), and I can't influence the quality
of the long-distance network links between sites.
More information about the questions