Re: "Hard drive destruction 'crucial'" (BBC)

["nick hatch" <nicholas.hatch () gmail com>]

On Sat, Jan 10, 2009 at 7:37 AM, der Mouse <mouse () rodents-montreal org>wrote:

> How long does it take to extract gigabytes of data by using an atomic
> force microscope to image each bit's recording area?  I don't know how
> fast they are, but whatever fraction of a second it takes to extract a
> byte, reading a gigabyte of disk will take an equal fraction of a
> gigasecond, which I shall let you work out for yourselves.


Hmmm. Wikipedia says that the nominal maximum size of an AFM scan is 150
micrometers^2 . From my experience with AFM, that sounds about right. Quick
math puts the minimum number of scans for one side of a platter at 315,000
-- and that's an absolute minimum based on surface area.

I don't know the data density or track structure for modern magnetic media,
but I don't think you'd even be able to see the bits at this resolution. I
found an image [1] of a "test platter" (not sure what that means) which is
only a few micrometers in size. Increase the above estimate by many orders
of magnitude.

The bigger problem here is tracking. Anyone who has done microscopy of any
sort knows how difficult it can be to get adjacent images. At the very
least, you're talking high-precision piezo movements. AFM is fun if you're
doing bulk surface analysis ("ahh! that's a clear image, let's stop now"),
and can be amazingly frustrating for defect analysis ("where the hell did
that spot go?").

I'm willing to say that nobody (spooks or otherwise) reads data off a HD
using AFM. It's just not the right tool. If you're worried about it: hit the
case with a hammer hard enough to bend/shatter the platter. A bent platter
would be a nightmare for ANY microscopy technique, even for a theoretical
custom spook machine.

-Nick

[1]
http://www.rhk-tech.com/results/UHV-MFM-image-of-Hard-Drive-Test-Platter_5.php

_______________________________________________
Fun and Misc security discussion for OT posts.
https://linuxbox.org/cgi-bin/mailman/listinfo/funsec
Note: funsec is a public and open mailing list.