Re: What have we done? (Here's what.)

Brian Murphy (brianm@earthlink.net)
Mon, 23 Jun 1997 10:26:13 -8


> Date: Sat, 21 Jun 1997 12:14:27 -0700
> To: cmcurtin@research.megasoft.com
> Subject: Re: What have we done? (Here's what.)
> From: "Karl J. Runge" <runge@crl.com>
> Cc: runge@jfku.edu, deschall@gatekeeper.megasoft.com

>
> On Sat, 21 Jun 1997, C Matthew Curtin <cmcurtin@research.megasoft.com> wrote:
> > number of | time to crack the message
> > tumblers | (on the average, at DESCHALL Project speed)
> > ----------|-------------
> > 40 | 78 seconds
> > 48 | 5 hours
> > 56 | 59 days
> > 64 | 41 years
> > 72 | 10,696 years
> > 80 | 2,738,199 years
> > 88 | 700,978,948 years
> > 96 | 179,450,610,898 years
> > 128 | 770,734,505,057,572,442,069 years
>
>
> Let me add my twist to the nice calculation Matt has done.
>
> Here are some estimates with respect to a "Quantum Computer(TM)".
>
> In an atom, an electron can orbit around the nucleus roughly 10^15 times
> per second.
>
> If we set the DESCHALL assembly code gurus loose on a quantum computer I'm
> positive ;-) they could refine their code to the point where it checks
> one key every time an electron does a single orbit.
>
> Since I'm trying to construct an UPPER BOUND here, bear with me and let me
> further assume each atom acts as an independent computer.
>
> An estimate to the total number of particles in the Universe is 10^80
> (a great ice-breaking fact to blurt out at parties!)
>
> So... the Ultimate Code Breaking computer would be every atom in the Universe
> trying a key each electron orbit. (I gotta get me one of those!)
>
> How many keys could it check in, say, 10 billion years (the estimated
> age of the universe), which is about 10^10 * 10^7 = 10^17 seconds?
>
> Number of keys = (number of computers)
> * (keys per second for one computer)
> * (number of seconds)
>
> Number of keys = 10^80
> * 10^15
> * 10^17
>
> = 10^112 = 2^372
>
> So a 372 bit key is REALLY safe IMHO against brute force attack!
>
> This is basically just doing Triple-DES twice, which I imagine
> is barely noticeable performance-wise when encrypting/decrypting
> messages.
>

I think 3DES really only has an effective key length of 112 bits. It
encrypts with Key 1, decrypts the message with Key 2 (producing
gibberish), then re-encrypts with Key 1 again. So you only need to
figure out two keys. I'm also not a crypto-guru(tm), but i don't
think that stacking 3DES will necessarily provide you with much more
security. It could very concievably make you more vulnerable to
various forms of cryptanalysis. But i do get yer point. 372 bits it
is..

>
> Karl
>
>
>
>

Brian Murphy
Network Engineer
Earthlink Network, Inc.
brianm@earthlink.net
It's my Internet.