[ntp:hackers] A stop-gap authenticated time service

Poul-Henning Kamp phk at phk.freebsd.dk
Tue Nov 10 18:27:02 UTC 2015

In message <8D2BF679AAC7C346848A489074F9F8BF7A9F067D at sjsrvexchmbx2.microsemi.net>, Greg D
owd writes:
>If you are steering a clock using current NTP, MitM can move it.  And accuracy
>is application dependent so it may be premature to say that what Dave Mills'
>called a thermometer can accurately maintain time for days.  In theory, that's
>why you were running the protocol to begin with.

So people tend to forget that the technological basis for NTP were
PDP/11 computers.

If you took the crystal out of a PDP/11 and tried to multiply it to
a couple of GHz with a PLL, you'd end up all over the microwave band.

If you read one of Intels or AMDs datasheets, you'll likely be surprised
by their harsh specs for clocks and crystals.

I have a lot of respect for Dave Mills, but one of the the things
he never updated was the "rules of thumb" derived from the PDP/11,
and that's why NTPD is far too eager to yank the clock around.

Modern motherboard crystals seldom move more than 1 ppm/K, and 1 ppm
for a day is only 86 milliseconds.

Here is a typical motherboard:


Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.

More information about the hackers mailing list