[kevkrom]

If every single sector on the drive gets forcibly written over, then data will be gone from that disk.

Negative. It still leaves a magnetic resonance than can be detected. Military standard (at leafs from 20 years ago, when I knew the current standards) require every sector of the disk to be overwritten three times, using three different data patterns. Even then I bet it's not a complete wipe but just "good enough" for operational purposes. The only true way to make a drive unrecoverable is to physically destroy it.

[coon2000]

Negative. It still leaves a magnetic resonance than can be detected. Military standard (at leafs from 20 years ago, when I knew the current standards) require every sector of the disk to be overwritten three times, using three different data patterns. Even then I bet it's not a complete wipe but just "good enough" for operational purposes. The only true way to make a drive unrecoverable is to physically destroy it.

Peter Gutmann theorized that using scanning tunneling microscopy you might be able to pull some data from a hdd, like a letter out of the entire encyclopedia. On top of that, that was from 20 years ago when drives stored far lass data, stored the data horizontally and the heads were far less precise. Todays far denser drives with vertical storage, extremely precise heads is a different matter. A single pass of writing over every sector, as long as it is truly random, is more than enough to delete all data on the drive.

Gutmann's old method for data deletion was 35 passes of random sector writes, which is 28 eight more than US govt top secret wipes. Today, he states that one pass of his method is all that is required.

Produce one instance where anyone has recovered data from a disk that has random 0's and 1's written to every sector, you will not find one.

[coon2000]

Follow up discussion on Gutmann's paper with comments from Gutmann concerning data deletion here:

http://seclists.org/bugtraq/2005/Jul/464