Rob Seaman wrote on 2006-02-06 17:37 UTC:
> 2) NTP can convey TAI as simply as UTC.
> 3) Deploy a small network of NTP servers to keep TAI, not UTC.
> 4) NTP client machines could therefore trivially select between TAI
> and UTC by subscribing to different servers.
NTP Version 3 as defined in RFC 1305 is clearly a protocol for
diseminating UTC. This specification has no provisions for disseminating
anything other than UTC (see Appendix E).
You can, of course, define, publish, implement, and promote a new
version (4?) of NTP that can also diseminate TAI, EOPs, leap-second
tables, and other good things. I'm all for it.
You can also propose a completely new protocol. There are TAI-based
Internet time protocols around and in use, for example
http://cr.yp.to/proto/taiclock.txt
What you seem to propose is just the cheap alternative of simply
bastardising the existing NTP protocol by violating RFC 1305 and abuse
it to disseminate something other than UTC, be it TAI, GPS time, UT1,
UT2, some local time. However, you are then no longer using anything
that deserves to be called "NTP Version 3".
To avoid misunderstandings, I would strongly recommend that anyone
operating such a server call their service and protocol provided
something other than "NTP". How about "TAIP", the International Atomic
Time Protocol?
I personally would very much prefer to see a protocol specification that
clearly indicates on the wire if something other than UTC is provided.
Anything else sounds as dangerous to me as using the same kind of plug
in countries that use 115 V and 230 V power. Naive users exist, and if
things appear to fit together, they will be plugged together by someone.
This is more likely to add to the problem than to the solution. I prefer
if the software involved has at least a fair chance of protecting the
user against doing something stupid. Usability experiences gained in a
testbed run by a small group of knowledgeable enthusiasts do not
necessarily scale into the real world.
Besides, bastardised "NTP" servers that replace UTC with TAI and UT1
have been around for quite some some; for instance Patrick Wallace
(Rutherford Appleton Laboratory) reported at the 2003 Torino meeting
about his "UT1P" server.
Markus
--
Markus Kuhn, Computer Laboratory, University of Cambridge
http://www.cl.cam.ac.uk/~mgk25/ || CB3 0FD, Great Britain
Received on Tue Feb 14 2006 - 11:51:00 PST