Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!ucsd!ucbvax!SAYSHELL.UMD.EDU!louie
From: louie@SAYSHELL.UMD.EDU (Louis A. Mamakos)
Newsgroups: comp.protocols.tcp-ip
Subject: Re: Network Temperature Protocol
Message-ID: <9007210143.AA09015@sayshell.umd.edu>
Date: 21 Jul 90 01:43:15 GMT
References: <9007210040.AA28109@ucbvax.Berkeley.EDU>
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The University of Maryland, College Park
Lines: 21

In article <9007210040.AA28109@ucbvax.Berkeley.EDU> you write:

>As part of a distributed computing experiment, we are considering
>setting up a Sun workstation, with a bi-metallic strip and small coil
>tempearture device, and providing a network wide reading, combined
>with time of day service and cartesian location data.

Actually, the facilities for doing most of this are already available on
the internet.  If you run NTP (network TIME protocol) on a host, you can
certainly provide the time of day service.  To determine the temperature,
you need only look at the drift rate of the clock relative to the a 
reference clock which corrolates rather nicely to the temperature of the
crystal in the clock of your computer.  You'll find that its trivial to
see day/night variations of the clock and hot vs. cold days.  All you need
do now is calibrate the drift rate to the temperature.

I'm sure Dave Mill can describe the use of NTP as an earthquake predictor
in another message.  This would be another important factor in picking
a location for a holiday.

louie