[time-nuts] ntp and asymmetric delays

[Hal Murray]

tjv at westwood-tech.com said:
> I was going to post my ntp output and ask for an opinion, then this
> discussion popped up.  It would appear that asymmetric delays are the exact
> explanation for what I am seeing.  Is that a reasonable assumption?  It does
> seem to be rather consistent throughout the day, however.  The reason for
> checking against the net when I have a GPS source is that I want ntp to
> continue if/when there is no PPS.  Is there any way to inform ntp of the
> asymmetry?

There are 3 sources of offset that I know about.

There is the hardware level.  If the upload and download speeds differ, that 
will introduce an offset.  You can do the arithmetic to work out how much.  
(You need to know the length of the packet.  I don't have it handy.)

The outbound and return paths can be different.  That is more likely if the 
server is farther away.  When that happens, the offset will be constant.  You 
can usually see occasional steps in the offset if you watch for several days 
when the routing changes.  Sometimes you have to watch for longer.

There are also queuing delays.  Bufferbloat is the extreme version of this.  
You can usually find a minimum of the offset and the offset beyond that is 
normally noisy.


I don't know of any way to tell ntpd about the offset for network packets.  
(There is a way for local refclocks.)


-- 
These are my opinions.  I hate spam.