[ntp:questions] IEEE 1588 (PTP) at the nanosecond level?

Martin Burnicki martin.burnicki at meinberg.de
Tue Mar 18 09:14:40 UTC 2014

Paul wrote:
> On Mon, Mar 17, 2014 at 8:50 AM, Joe Gwinn <joegwinn at comcast.net> wrote:
>> Yes.  My question is basically a query about the current state of the
>> art.
> Some NTP offsets (output may look funny if formatted) clock1 looking at
> clock2 and clock3 (a Raspberry Pi).  This suggests it can be as good as
> your IRIG system.
> Gig ethernet to Gig ethernet.  ~22 days:
>      N           Min           Max        Median           Avg        Stddev
>   244059  -0.098741335   0.019727433     9.586e-06  4.814598e-06
> 0.00038621792
> Gig to Fast ~10 days
>      N           Min           Max        Median           Avg        Stddev
>   112254  -0.000516264   0.000453913     1.127e-06 6.8736914e-06
> 4.2248166e-05

But without additional measurements you still don't know for sure if 
this is the true time offset, or if there is an additional systematic 
time offset (e.g. to an asymmetric network connection) which can't be 

For example:

I have an NTP node running on my home network connected via an 
asymmetric DSL line, getting the time from the internet.

If I monitor that NTP node from our company network which is connected 
to the i'net via a symmetric line then my home NTP server seems to have 
a low offset, below 1 ms.

However, if I let my home NTP node synchronize to a GPS receiver then 
suddenly the the offset observed from our company network increases to a 
few milliseconds.

The reason is that the way the asymmetry affects the polling my how 
server does towards the internet, and the way the asymmetry affects the 
polling of my home server from our company are in the same range, but 
with reversed sign, so the effect is compensated and from our company's 
point of view my home server is very accurate even though it isn't.

If my home server syncs to GPS then the accuracy isn't compensated 
anymore and if I'm monitoring my my home server from our company there 
now seems to be a time offset, where in fact there isn't one anymore.

Martin Burnicki

Meinberg Funkuhren
Bad Pyrmont

More information about the questions mailing list