From: "Uhh...this is Joe [Randall Farmer]" <rfarmer@HiWAAY.net>
Date: Thu, 29 Jan 1998 06:06:35 +0800
Subject: Predicting cipher life / NSA rigged DES? / Destroying encrypted data (Tangent to Re: Burning papers)
In-Reply-To: <199801252007.OAA19838@manifold.algebra.com>
Message-ID: <Pine.OSF.3.96.980125142747.30243A-100000@fly.HiWAAY.net>
MIME-Version: 1.0
Content-Type: text/plain



[These messages were postponed for trillions of years I finally sent them;
apologies if something is grossly outdated.]

Been thinking, most applications for ciphers assume solely based on cipher x's
keysize that data will be secure for a certain length of time. It'd be nice if
we had some way to estimate how long we can hope the cipher to last. Of course,
there's no way to predict anything for sure, but you could make an estimate.
I'm wondering if there's any way to make a more accurate prediction of how much
more analysis it will survive with fancy statistics or something. My idea --
which I know wouldn't work very well, which is why I'm asking if there's a way
to actually make a good guess -- is averaging the remaining lifetimes in
analysis-hours of broken ciphers which survived as many person-hours of attack
as the one in question.

===============================================================================

Am I just going crazy, or is it kind of obvious that NSA knew the s-boxes they
provided for DES weren't secure? I mean, they pretty much had to know about the
attacks outside cryptographers are just now discovering -- they have more than
ten years of cryptographers' time every day, and they certainly knew about
differential cryptanalysis. Let's hope they don't meddle similarly in AES... 

===============================================================================

> > Of course, if your documents are on floppy disks, any shredder that 
> > won't jam on them does a pretty good job :-)  
>
> I burned a couple of floppies, too. Actually I am not sure how good job 
> would shredding of floppies do. I assume that bits and pieces of data 
> can still be recovered... But hopefully no one would care enough. 

One fairly simple feature for disk encryptors that came up during one of the
#ElectronicFrontiers (sp?) chats was that of using random numbers with the key
so you can demolish an encrypted volume in a split-second. Works like this:
there's one 192-bit (or whatever your keylength is) value which is a hash of
your passphrase. There's another value, this one a cryptographically random one
of the same size, stored on a fixed physical place on the disk. If you wish to
destroy the data on powerdown, there can be a third value stored in memory,
which is written to disk at authorized shutdown and read+wiped from disk at
startup. 

Anyhow, these two (or three) values are XORed together to form the key used to
encrypt the volume. When your adversaries, armed with their trusty rubber
hoses, come knocking at and/or down your door, you hit a hotkey to start
destroying those 24 bytes on disk, which can be done faster and more
effectively than a wipe of every sector in the volume. The folks with the
rubber hoses are now, assuming this is their first peek at your disk, screwed; 
even with your passphrase, they don't know a thing about your data. 

> 
> 	- Igor.

---------------------------------------------------------------------------
Randall Farmer
    rfarmer@hiwaay.net
    http://hiwaay.net/~rfarmer