[ntp:questions] how to have offset < 1ms

nemo_outis abc at xyz.com
Tue Apr 13 22:10:48 UTC 2010


Rob <nomail at example.com> wrote in
news:slrnhs9f0a.n5.nomail at xs8.xs4all.nl: 

> unruh <unruh at wormhole.physics.ubc.ca> wrote:
>> Of course. That is the thing about solutions. You provide extra
>> information or hardware that is needed to impliment the solution. Now
>> if you have a better one, I am sure he would welcome it. But I have
>> not seen it. 
> 
> In his position, I would investigate the cause of the delay.  There
> may be something going wrong that affects other network operations as
> well.  This needs to be debugged, and suddenly NTP and the other
> applications will work better.
> 
> I certainly would not suggest adding hardware and input to the system
> that is not in the specification.  As I wrote before, he wants to sync
> to a specified source.  The source may be off from GPS, but that is
> what he wants to sync to, not GPS.  And I already gave a very good
> reason for that.
> 
> When I add a system to our LAN I want to sync to the local NTP
> reference, not to something else.  The NTP reference is synced to a
> DCF radio clock and outside sources, but that is not important.  What
> is important is that all systems keep the same time.
> 
> Also, you always ignore the issue that a GPS receiver will not work
> without an antenna position with sky view.  This is often not
> available. (which actually is why we use a DCF receiver, not a GPS).
> 
> Don't think that your favourite solution to your own problem is
> everyone's solution to every problem.
> 

You are a nimble disputant - once you allow yourself all the assumptions 
you make.

However, ignoring your sophistry, the fact remains that, for a very wide 
range of likely scenarios, a ship will have somewhere between an 
excellent and superb view of the sky, and installing a GPS that provides 
microsecond (or so) accuracy will be a matter of, at most, a hundred 
dollars or thereabouts.  In short, the cost to implement such a scheme 
would be less than the time spent here discussing it, even valuing 
folks' time at < $10/hr.

True, one can chase any number of will-o-the-wisps such as the causes of 
the network latency and asymmetry, but that is not the central problem 
posed by the OP.

In short, Unruh may be boring in his reiteration of his simple GPS 
solution, but that is because it works for a very wide range of problems 
- including this one! Boring and repetitive perhaps, but nonetheless 
effective, simple, robust, and cheap.

Perhaps if you seek excitement and intellectual challenge, the OP's 
problem is not the right vehicle?

Regards,






More information about the questions mailing list