February 26th, 2003, 11:17 PM
I saw this (below) on slashdot once as a comment regarding hard drive data recovery and paranoia.
I know it's a joke, but is this actually possible: the pseudorandom data writing on pseudorandom addresses of an encrypted partition? I don't have any data valuable enough to require such extreme measures of removal, but do any *nix gurus here have input?
Assuming that you have at least a few seconds to react when they come knocking then planning takes care of a lot of this. The system in question which I'll I call the Naughty Super Secret System or NSSS for short needs to be specially configured. It should have no swap files or swap partitions of any sort. The /tmp directory or any equivalent should be a ramdisk formatted with an encrypted filesystem. Any permanent datastores should also be on encrypted filesystems. The best part is that the NSSS also has a "panic script" thats triggered with a hotkey combination. There will be no time to actually type a command. The panic script will lock the terminal, unmount any ramdisks, change the filesystem password to a random collection of characters if possible and clobber the control structures of the encrypted filesystems with random data (superblocks, fat tables, etc). This is not a lot of data and won't need more than a few seconds to royally bollix. Actually, random data sprayed across an encrypted filesystem will do far more damage than a conventional filesystem. If the clobber script has enough time to hit those control structures with seven passes it should then spray random bytes across the remainder of the partition as long as it's permitted to run. In any case, the clobber script will run until some quick thinking MIB pulls the power cord. That can be made a pisser as well. Remove any obvious way to quickly power off the machine and make it necessary to spend a few more seconds getting at the power cord or UPS. Hmmmm....how's this? Put the UPS inside the machine and rig the physical power switch well inside the case. The machine can be powered up or down by sticking a screwdriver into a hole to operate the switch. LOL, put lots of extra screws in the case too.... That should buy more than enougn time for the panic script to do it's work.
Thanks in advance!
Have you filled out an ID-10-T or PEBKAK form lately?
February 27th, 2003, 08:01 AM
Yup is totaly possible !!
And (I don't dare say it) unbreakable !!
Say you take a message like: secret words
Then encrypt it (in some way) now say bladiblableh comes out. It would take the CIA / FBI / NSA some time (depending on how powerfull the encryption) to decrypt bladiblableh into secret words
then you hit the panic button..
bladiblableh gets hit with some random (pseudo or not) chars..
beadvblapreh can never be decrypted into secret words
and since the script that ran is also encrypted and (possibly) polluted, there is no way of telling exactly what you did..
I'd call it (nearly) unbreakable..
Big downside is you lose all that sensitive info yourself !!
ASCII stupid question, get a stupid ANSI.
When in Russia, pet a PETSCII.
Get your ass over to SLAYRadio
the best station for C64 Remixes !