News for hackers. Stuff that matters.

A comment on Old-School PC Copy Protection Schemes (2006). Parent comment

Re:Old-School PC Copy Protection Schemes (2006) by jonathankoren on Tuesday July 19, 2016 @03:34AM (#12120360)
This worked because the magnetic domains on the disk were larger than the r/w head on drive, and due to play in the servo mechanism. It's the same reason why the DOD creates the 7-pass rule for securely erasing disks. You could simply read the disk multiple times and get slightly different results with some statistical analysis and some cribs, you could at least partially reconstruction the original data.

Today, magnetic disks have magnetic domains pretty much the same size as the head, and the servos are much more accurate. One pass erasing is sufficient. 27-pass is just a waste of time.

Now for SSDs, you're just screwed.

[ Parent | on HN ]
Re:Old-School PC Copy Protection Schemes (2006) by legodt on Tuesday July 19, 2016 @10:54AM (#12122128)
Could you please elaborate on your SSD comment? I'm interested to hear more.
[ Parent | on HN ]
Re:Old-School PC Copy Protection Schemes (2006) by theandrewbailey on Tuesday July 19, 2016 @11:30AM (#12122299)
SSD wear leveling algorithms ensure that the entire disk is used evenly. Any block on the drive can be literally anywhere in the storage medium, even more so than with hard drives. Some SSDs have significantly more storage media than they report (over-provisioning), again for wear leveling.

In theory, SSD erasing is solved by TCG OPAL functionality. All data written to the disk medium is encrypted, whether or not it is taken advantage of by the user. The key it uses cannot be retrieved or set by the user. Tell the disk to change the key it uses and the entire device is unintelligible noise.

Some articles: http://www.techrepublic.com/article/erasing-ssds-security-is...

http://arstechnica.com/security/2011/03/ask-ars-how-can-i-sa...

[ Parent | on HN ]
Re:Old-School PC Copy Protection Schemes (2006) by flyinghamster on Tuesday July 19, 2016 @12:13PM (#12122574)
Shouldn't the ATA Secure Erase command take care of properly blanking an SSD? (Assuming that its firmware can be trusted, of course.)

In practice, it's a bit more complicated, though - typically the BIOS (or is the operating system?) locks out Secure Erase and you have to hot-remove the drive and reconnect it before the drive will accept it.

[ Parent | on HN ]
Re:Old-School PC Copy Protection Schemes (2006) by bitwize on Tuesday July 19, 2016 @12:24PM (#12122644)
> Shouldn't the ATA Secure Erase command take care of properly blanking an SSD? (Assuming that its firmware can be trusted, of course.)

It does that -- by clobbering the hw encryption key.

[ Parent | on HN ]