Wipedisk Question

That's cool, but I guess there is a price when you build up a bleeding edge system from scratch. 6 months later the parts are 1/3 the cost and more stable.

That sounds hopelessly optimistic for any modern HDD and really sensitive data (though it would certainly stop the most common kind of snooper). Most of the surface of the disk will be intact and MOST of the data could be recovered. Even an intact chunk a few mm in size could contain a lot of data because the surface density of data is so high.

I think a cement kiln or maybe a hammer mill with a small enough sieve would be more effective and probably safer.

Even media that once contained top secret data is still considered top secret after being 'sanitized'.

Irritating, isn't it? My striped RAID system with dual 10K RPM disks is almost matched in speed by a single WD 'AALS' drive costing $75 (a couple of years later, but

There's a bunch of stuff that most programs miss, because a file grabs

4k of disk space at a time, but maybe only writes a fraction of it (so the remaining 3k of old info can be 'in use' by the file system, but still have the old data on it).

Because the new file is using the first 1k, the relic data is not free for an erase utility to deal with; only if you secure-erase the original file BEFORE reusing the disk space, can you be sure it's gone.

Backup all your useful data using a compression program (this will lose the 'relic' data). Erase the disk with a full-wipe program. Restore from the backup.

I prefer thermite.

It should also be sufficient to heat the media to the curie point (or run it over with a strong degausser). Neither works very well with the case together, but anyone with a torx bit can get inside at the expense of less than a minute's labor.

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms

Once you can no longer read the disk with a read head above a spinning platter, the cost of recovery is such that no-one is going to bother unless they contents are important enough to land on the president's desk.

If the drive is physically functional, an ATA "secure erase" command will wipe it to the degree that it's unrecoverable by any known technology.

That's protecting against the possibility of the drive falling into the hands of a foreign government which stores it for several decades in the hope that a way to recover the data will eventually be discovered.

Right now, a secure erase will protect it against anything short of the NSA sending it to Area 51 for analysis with captured alien technology. Simply overwriting the sectors with "dd if=/dev/zero ..." will almost suffice, except that it won't work for sectors which have been transparently remapped by the drive's firmware.

When it comes to circumventing chip security, some pretty impressive stuff can be done with surplus equipment for < $10K worth of equipment, and even more impressive on ~$1M equipment which can be rented for a few hundred $ an hour.

Here's one possible disk data recovery method:-

Yes, it should.. and probably will.. particularly if it's degaussed afterward.

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

Eraser at least is open source which goes a long ways to addressing that concern.

A couple of days later and the News For Nerds site puts up an article called "The Home-Made Hard Disk Destroyer" indexed here to an item mentioning liquid nitrogen:

...and I REALLY despise Slashdot's newest default layout:

r

on#description

All these guys keep *claiming* they've recovered overwritten data, but (also referenced in the Slashdot comments) this guy says he looked at the "evidence" and it's all hooey:

*-extraordinary-claim+DBAN+*-*-*-*-*-*-*-*-ordina=ry-wipe-disk-programs-*.*-*-*-*-*+sophisticated.microscope+no.reference+err=or.rate+18-minute-*+deserving-*-extraordinary-proof+urandom.*.*-*.*-*+not-*=-sectors+urban.legend+GPL+*-single-write-*-*&strip=3D1

Do note the reference there to Rosemary Woods.

I've collected several articles over the years and they all say the same thing: A single-pass overwrite with random data will do the trick. Beyond that it's sci-fi and/or paranoia.