PSN-L Email List Message

Subject: RE: Network time standard
From: "Keith Payea" kpayea@...........
Date: Sat, 9 Apr 2005 13:53:14 -0700


Not all NTP clients are created equal...  There are actually two standards.
True NTP uses four time stamps to attempt to correct for transmission
delays, and also runs a fairly complex filter based on many samples to set
the time in your computer.  Simple NTP (SNTP) uses just one time stamp, and
is usually run once a day or every few hours with no filtering.
 
Try to find an NTP that is based on the University of Delaware version.  Dr.
Mills at UDel is considered the true guru of NTP and his team have put many
years into squeezing the best performance out of it.
 
My experience when working for TrueTime (Now a division of Symmetricom) was
that you could achieve 10 milliseconds or so if the server was not too far
away on the public internet.  If you run a local time service on your own
network, the timing is much better than that.  It usually takes several
messages to get there due to the filtering.  Also, true NTP will perform
much better if you give it more than one server to work with.  The software
makes some attempt to evaluate the stability of each source and choose the
best.
 
Obviously, NTP is not as good as GPS, but it is very comparable to WWV or
WWVB or DCF77 radio performance.
 
Cheers,
 
Keith
 
 

  _____  

From: psn-l-request@.............. [mailto:psn-l-request@............... On
Behalf Of ChrisAtUpw@.......
Sent: Saturday, April 09, 2005 1:06 PM
To: psn-l@..............
Subject: Re: Network time standard


In a message dated 09/04/2005, shammon1@............. writes:

If you don't have a GPS, as I said, the software comes with an extensive
list of network time servers that can be accessed via the Internet to obtain
accurate time. Overall, the experience with this software package has been
very positive and after 30-days of testing I'm now recommending it to other
members of the Public Seismic Network.


Hi Steve,
 
    Does it give any timing accuracies in milli seconds for the various time
servers? 
 
    One of the problems with time services is that while the signal may
leave the server on time, it may be digitally delayed on it's way though the
telecom system. Since there are many services now operating, the inaccuracy
around some timing marks may be very significant. 
 
    What timing errors have you measured, please?
 
About the Public NTP Time Server Lists 

The lists are provided for information purposes only and represent the best
information available at the current date. The operators of the servers
listed do not commit to provide time service other than on a volunteer basis
and with no guarantee of accuracy or availability. Further information of a
technical nature can be obtained from the www.ntp.org 
site, the comp.protocols.time.ntp newsgroup, or the NTP Support
  Web. 

    Regards,
 
    Chris Chapman
 
     





Not all NTP = clients are=20 created equal...  There are actually two standards.  True NTP = uses=20 four time stamps to attempt to correct for transmission delays, and also = runs a=20 fairly complex filter based on many samples to set the time in your=20 computer.  Simple NTP (SNTP) uses just one time stamp, and is = usually run=20 once a day or every few hours with no filtering.
 
Try to find = an NTP that=20 is based on the University of Delaware version.  Dr. Mills at UDel = is=20 considered the true guru of NTP and his team have put many years = into=20 squeezing the best performance out of it.
 
My = experience when=20 working for TrueTime (Now a division of Symmetricom) was that you could = achieve=20 10 milliseconds or so if the server was not too far away on the public=20 internet.  If you run a local time service on your own network, the = timing=20 is much better than that.  It usually takes several messages to get = there=20 due to the filtering.  Also, true NTP will perform much better if = you give=20 it more than one server to work with.  The software makes some = attempt to=20 evaluate the stability of each source and choose the best.
 
Obviously, NTP is=20 not as good as GPS, but it is very comparable to WWV or WWVB or DCF77 = radio=20 performance.
 
Cheers,
 
Keith
 
 


From: psn-l-request@................. [mailto:psn-l-request@............... On Behalf Of=20 ChrisAtUpw@.......
Sent: Saturday, April 09, 2005 1:06=20 PM
To: psn-l@..............
Subject: Re: Network = time=20 standard

In a message dated 09/04/2005, shammon1@............. writes:
If you don't have = a GPS, as I=20 said, the software comes with an extensive
list of network time = servers=20 that can be accessed via the Internet to obtain
accurate time. = Overall, the=20 experience with this software package has been
very positive and = after=20 30-days of testing I'm now recommending it to other
members of the = Public=20 Seismic Network.
Hi Steve,
 
    Does it give any timing accuracies in = milli=20 seconds for the various time servers?
 
    One of the problems with time services is = that=20 while the signal may leave the server on time, it may be digitally = delayed on=20 it's way though the telecom system. Since there are many services now = operating,=20 the inaccuracy around some timing marks may be very significant.
 
    What timing errors have you measured, = please?
 
About the Public NTP Time Server Lists

The lists are provided for information purposes only and represent = the best=20 information available at the current date. The operators of the servers = listed=20 do not commit to provide time service other than on a volunteer basis = and with=20 no guarantee of accuracy or availability. Further information of a = technical=20 nature can be obtained from the www.ntp.org site, the comp.protocols.time.ntp newsgroup, or the NTP Support = Web.=20

    Regards,
 
    Chris Chapman
 
     

[ Top ] [ Back ] [ Home Page ]