[GRLUG] Disk scrubber

[Michael Mol]

On 7/25/06, Tim Schmidt <timschmidt at gmail.com> wrote:
> On 7/25/06, Michael Mol <mikemol at gmail.com> wrote:
> > Depends on the device.  The ability for a read/write head to read
> > faint and overwritten signals is limited by the drive's circuitry and
> > the drive's firmware.
> >
> > I know it's possible to access reallocated sectors on some drives,
> > with the right instructions sent to the drive.  However, accessing
> > overwritten data using the original read/write head and the drive's
> > built-in circuitry would require a firmware upgrade, if it's possible
> > at all.  The software Tim saw likely had access to the correct ioctls
> > to read reallocated sectors.  It's less likely, but still possible,
> > that it had the ability of changing the drive's firmware.
> >
> > I asked because I was curious as to the nature of the software.
>
> I have no idea guys...  It was expensive (more than $500, license tied
> to number of uses) commercial software purchased specifically for a
> single client.  For all I know the software company producing it could
> have agreements with the drive manufacturers to disclose special
> registers or whatever to access extra drive features.
>
> It didn't work 100%, but I was amazed at the time at just how much
> information was recovered.
>
> That said, modern drives are small embedded computers, with megabytes
> of ram, high-performance DSPs, and plenty of programmable hardware.
> In theory at least, there's nothing stopping WD or Seagate from
> including a backdoor of sorts that allows analogue access to the drive
> platter.  Given the cooperation of color laser printer manufacturers
> with the US Gov, as well as the telecoms, and the banks...  why not
> the drive OEMs?

I try not to think about such things.  But as several people in an old
Slashdot article pointed out, if they could access overwritten data,
wouldn't they try to use it to increase storage volume?

I wonder if they already do that for magnetic tape.
-- 
:wq