Since we all know files are recoverable with programs after being deleted from the recycling bin, everyone is told to do secure wipes by putting random data over the files you're deleting on the disk itself. From what I understand with Hard Drives (and CDs, and Floppies, etc), the data is written to either be reflective/raised up (1) or non-reflective/sunken (0). My question is then how can a bad guy recover data that's been overwritten just once, since it would change the reflective or height of the gap?
- See SSD (Flash Memory) security when data is encrypted in place and watch the discussion links on hard drives.
- On older drives, it was likely possible to recover overwritten data. The density of modern drives for the past 5+ (and the plus may be 10, 15, or more) years has been so high that a single pass of random data means whatever was there before is just not going to come back unless the NSA knows something about this that the rest of the general public doesn't.
- Modern file systems are such that the data may have been cached in several places, so unless the entire disk was wiped, the answer is to look for cached, duplicated, versioned, etc. copies of the file that were overlooked during the erasure process.
how can a bad guy recover data that's been overwritten just once, since it would change the reflective or height of the gap?
There are three basic methods: multiple copies, error correction, and enhanced detection.
A hard disk, or CD-ROM for that matter contains a filesystem. A filesystem is an organization of files in a way that makes sense to the computer and allows you to arbitrarily create, modify, and delete files at random.
Some filesystems have performance enhancing features like quick writes or protecting an original file while modifying it. Some features may have copies of pieces of a file or even the entire file in a separate location on the disk.
CD-ROM and DVD-ROM filesystems have error correcting codes so that if your disk is chipped or scratched and some bits of a file are unreadable, it can figure out that bits are missing and recreate them.
Magnetic bits on a hard disk platter are writen and read by sensors called heads on the end of a arm which moves across the disk platter. The bits are written in concentric circles called tracks. The tracks are spaced far enough apart from each other to prevent a magnetic bit on a track from causing the head to read an adjacent track incorrectly. However magnetic fields are not clearly discrete and recording on a track will bleed slightly into the margin between the tracks. About ten years ago the size of a disk track that the heads could write were an order of magnitude bigger than what scientific instruments could detect. In that case it was possible to use scanning tunnling electron microscopes to look for the write patterns in the margin between tracks. However, doing so would have been extremely costly and difficult, and likely never happened in the wild.
The general theory is that some residue of the previous state exists due to the limited accuracy of the rewrite. There may have been a proof of concept demonstration, but I'm not aware of any claims that it's actually been done in earnest. It would be incredibly time consuming and expensive to even attempt.