What do you think about communications data collection and storage?

[Paul Jakma]

On Tue, 5 May 2009, Roland Perry wrote:

> Only if you can show that the clocks are significantly wrong 
> (therefore by several seconds).

Given that it's hard to find accurate, cheap, electronic clocks and 
that keeping time to within 1s/day is a /good/ clock, then surely the 
burden of proof should lie on those who claim time-stamps are 
accurate?

E.g., here's the data-sheet for one of the hugely popular Dallas 
Semiconductor range of RTCs:

  http://www.maxim-ic.com/quick_view2.cfm/qv_pk/2680/t/do#Data%20Sheet

These things are /claimed/ accurate to only 2s/day at their optimal 
temperature range of 25C. Accuracy varies significantly between 
different crystals. Accuracy becomes worse, by the *square*, with any 
change in temperature, and can also be affected by RF noise and 
circuit design issues. Computer equipment cases, particularly in hot 
data-centres, are usually above 25C. Further, there are much higher 
temperature hot-spots inside, and RTCs sometimes may be unfortunately 
close to them.

My laptop has an error, as measured by NTP, of about 0.9s/day (which 
is really good). Figures of around 3s/day seem to be more typical for 
servers of mine (sample size: 3). Which is a drift of nearly 1.5 
minutes every month.

Dallas Semi seem to have newish RTCs out now that can be accurate to 
1m/annum, but they don't seem to be in wide-spread use.

> And why would they be? This issue hasn't overly troubled 
> investigations in the past, and I see no reason why it's useful to 
> raise it now.

I guess it won't be till someone acting for a defendent asks for 
proof that clocks were synchronised...

regards,
-- 
Paul Jakma	paul@clubi.ie	paul@jakma.org	Key ID: 64A2FF6A
Fortune:
This is your fortune.