[ntp:questions] Re: ntpd -x

David L. Mills mills at udel.edu
Mon Jan 30 19:41:11 UTC 2006


I don't know where you are going on this issue. I am told the FAA does 
indeed use NTP, but I don't know specifically with what application.

The NTP engineering model provides a number of tinker and tos command 
options with which to control the acceptable performance envelope. The 
NTP model mitigates only its sources and does not pretend in and of 
itself to watch other participants in the application. This is of course 
possible using the ntpq and other tools, but is not a feature of the 
intended design.

My monotonically crashed airplane scenario was intended only to provoke 
serious thought about the policy of forbidding step corrections as an 
absolute edict. This in principle has nothing to do with the NTP design 
or implementation.


Brian T. Brunner wrote:

>>>>"David L. Mills" <mills at udel.edu> 01/27/06 12:09PM >>>
> DaveM> At the very real risk of boring everybody on this list to tears, I have 
> DaveM> a burning agenda to expose this issue to a responsible community.
> DaveM> Your flight dispatcher machines are running just fine and one of them 
> DaveM> suddenly veers off course by one hour. Your application is air traffic 
> DaveM> control. Your choices are:
> This seems to cross the boundary between NTP and the application.
> NTP has no way of telling whether all the machines running a particular 
> distributed application all have the same notion of "the time".
> A significant question is whether the distributed database application  
> (air traffic control in this case) runs on machines that all are NTP co-peers 
> or co-clients of a single time server (set)?
> AIUI, NTP only disciplines the local clock, and passes judgement on the credibility
> of other clocks; it does not judge itself to be the false ticker, so it can't tell
> the application that the time returned (on this machine) is hysterical, and the 
> application should go ask someone else.
> I think it is outside NTP's current scope to implement NDC (Network Distributed Clock)
> or NTOD (Networked Time Of Day)  or NTR (Network Time Reference) to  provide 
> "the time" for all distributed application actions.  Are you (David) suggesting that
> the NTP project scope be broadened to embrace some concept of NTOD?
> Brian Brunner
> brian.t.brunner at gai-tronics.com
> (610)796-5838
> *******************************************************************
> This email and any files transmitted with it are confidential and
> intended solely for the use of the individual or entity to whom they
> are addressed. If you have received this email in error please notify
> the system manager.
> This footnote also confirms that this email message has been swept
> for the presence of computer viruses.
> www.hubbell.com - Hubbell Incorporated
> _______________________________________________
> questions mailing list
> questions at lists.ntp.isc.org
> https://lists.ntp.isc.org/mailman/listinfo/questions

More information about the questions mailing list