What a terrible paper title.

For anyone wondering, this has nothing to do with microkernels. From the abstract:

  The clock frequency in modern workstations is stabilized by an uncompensated quartz or surface
  acoustic wave (SAW) resonator, which are sensitive to temperature, power supply and component
  variations. Using NTP and traditional Unix kernels, incidental timing errors with an uncompensated
  clock oscillator is in the order of a few hundred microseconds relative to a precision
  source. Using new kernel software described in this paper, much better performance can be
  achieved. Experiments described in this paper demonstrate that errors with a modern workstation
  and uncompensated clock oscillator are in the order of a microsecond relative to a GPS
  receiver or other precision timing source.

What's the date on this paper?

Pet peeve: research papers with no dates.

November 2000

Yep. Source: https://www.eecis.udel.edu/~mills/database/papers/nano/

Assuming the timestamps are not forged, this only shows that the paper was uploaded November 2000. It might have been older. So we need a second source.

Second source: The citations. The paper cites: "Mogul, [...] Pulse-per-second API for Unix-like operating systems [...] March 2000 [...]"

So the paper is not older than November 2000 and not younger than March 2000.

It was published at the "Precise Time and Time Interval Meeting" in November 2000 (which a footnote in nano.pdf says as well)

Could anyone add "(2000)" to the title, please?