Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 23 Apr 2003 13:16:24 -0400
From:      Bill Moran <wmoran@potentialtech.com>
To:        David Kelly <dkelly@hiwaay.net>
Cc:        Joel Rees <joel@alpsgiken.gr.jp>
Subject:   Re: Accurate time without a network connection?
Message-ID:  <3EA6CA68.1060902@potentialtech.com>
In-Reply-To: <20030423161358.GA24633@grumpy.dyndns.org>
References:  <4B518202-74F8-11D7-BCB7-003065ABFD92@mac.com> <20030422194413.GC13774@grumpy.dyndns.org> <20030423175412.B1F6.JOEL@alpsgiken.gr.jp> <20030423161358.GA24633@grumpy.dyndns.org>

next in thread | previous in thread | raw e-mail | index | archive | help
David Kelly wrote:
> On Wed, Apr 23, 2003 at 05:58:47PM +0900, Joel Rees wrote:
> 
>>>On Tue, Apr 22, 2003 at 03:26:19PM -0400, Charles Swiger wrote:
>>>
>>>As for exactly what time the systems have, it doesn't much matter as
>>>long as they all have the same time.
>>
>>If you're serious about it not mattering exactly what time they have,
>>what's the problem with letting one machine be the time server, letting
>>it tune itself and then free run, and syncing all the rest to the
>>slightly-off-time-server?
> 
> Because the other systems do not have a link to the first.
> 
> I could sync them all to a common source they could *hear* but the
> customer won't allow them to *talk* to anything.

I've recently become interested in the theory behind NTP and timekeeping
in general (because of situations in my work that have gotten me interested)
and have followed this thread with some curiousity.

Mind if I ask some rather off-the-wall questions to possible formulate some
out-of-the box ideas?

What, exactly, is the environment you are doing this in?  If I remember
previous posts correctly: You need time keeping, it doesn't have to be
very accurate, or even "correct" (to the degree that the time could be
outright wrong as long as it's consistent within the company - did I
undestand that right?)  The machine in question is not permitted to access
ANYTHING via a network, even other machines withing the company?

First off, can you explain the rational behind this?  It sounds crazy to
me.  Almost sounds like you're asking for the user to have 100% control
over their computer (root access) while being 100% sure that they'd never
mess anything up ;)

With the requirements you've described ... my first guess would be to tell
you to install some sort of high-accuracy clock in the facility that
broadcasts a radio signal, and set up a receiver on every computer in the
office to sync off it.

However, you never state it precisely, but your insistence on isolation
suggest a security concern (is that the reason?) and that radio signal
could easily be hijacked and (best case) used to foul the time keeping
of the whole office or (worste case: the NTP driver you use has flaws)
compromise the systems using the signal.

With that in mind, I would suggest wiring the machines up and using NTP
over the network to sync off one machine that is kept updated either
manually, or via a high-accuracy clock connected to a com port or
something.  It's a lot easier to keep a wired network secure than any
kind of radio/wireless network.

Again ... I'm hoping you can provide more details about the reason you're
trying to do things the way you are.  It really seems like you're asking
for the impossible and you'd be better off listing out your conflicting
goals and determining which one you can discard.

Hope this is helpful and not confrontational in any way: it's intended
to be helpful.

-- 
Bill Moran
Potential Technologies
http://www.potentialtech.com



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3EA6CA68.1060902>