🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: sci.crypt
Article #2812

Disagreement with cost of DEA brute force crackmachine

#2812
From: Mark Johnson
Date: Fri, 22 Jan 1993 14:48
48 lines
2936 bytes
In article <1jmumhINNe3d@roche.csl.sri.com> boucher@csl.sri.com (Peter K. Boucher) writes:

   > In 1976, given economies of scale, each [custom built DES] Integrated
   > Circuit would cost ~$10 ([therefore] $10M/machine).  Allow a factor of
   > two, for other hardware and you pay $20M/machine.  Amortize costs over
   > 5 years, and you pay $10K/day.
   > [ ... ]
   > These are 1976 figures.  Even if we are optimistic by an order
   > of magnitude, the cost of computation falls in 5 years by a factor 
   > of 10; in 15 years (by now) by 1,000.


But you're talking about hardware design, not computation.


Have a look at a hardware budget for your chip, which holds
one copy of DES circuitry.  A swag at such an accounting is
shown below.  It is wrong (optimistic) because it assumes
the cost of the permutations and shufflings is zero, whereas
on silicon, gates are cheap and wires are expensive.

  a.  56 bit register for key bits                        336 gates
  b.  64 bit input register                               384 gates
  c   16 ranks of   32b  register pairs                  6144 gates
  d.  16 ranks of   (32  2-bit-XOR)                      1024 gates
  e.  16 copies of  (32b scramblers using S boxes)        480 gates
  f.  64 bit output register                              384 gates
                                                     -----------------
               TOTAL                                     8752 gates


Your $10-per-chip-in-1976 figure, divided by your "computation
cost reduction factor" of 1000, gives today's cost per chip of
$0.01.  Balderdash! Not in this universe in the year of our lord 1993.

But in today's practice we'd of course use high levels of integration,
putting (say) 32 DES hardware widgets on a single chip.  The
resultant VLSI device would have 280,000 logic gates (about the
complexity of the Intel 486 microprocessor).  Your cost figure
for this device would be $0.32.  Not in this universe in 1993.

Hire some genius to optimize the holy snot out of the hardware,
assume Allah is on your side and the gate count drops by 4X.
Now you can get 128 DES widgets on an i486 sized chip and
you pay the silicon fab $1.28 for this IC.  Good luck Mr. Phelps.


   --- Mark Johnson
Technical Details
Message-ID: <MJOHNSON.93Jan22064848@netcom2.Netcom.COM>
Path: arc1-ca.pugleaf.net!sparky!uunet!olivea!spool.mu.edu!agate!stanford.edu!kronos.arc.nasa.gov!butch!netcomsv!netcom.com!netcom!mjohnson
References: <1444.204.uupcb@ssr.com> <1993Jan9.163011.23230@csi.uottawa.ca> <1jmumhINNe3d@roche.csl.sri.com>

Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!olivea!spool.mu.edu!agate!stanford.edu!kronos.arc.nasa.gov!butch!netcomsv!netcom.com!netcom!mjohnson
From: mjohnson@netcom2.Netcom.COM (Mark Johnson)
Newsgroups: sci.crypt
Subject: Disagreement with cost of DEA brute force crackmachine
Message-ID: <MJOHNSON.93Jan22064848@netcom2.Netcom.COM>
Date: 22 Jan 93 14:48:48 GMT
References: <1444.204.uupcb@ssr.com> <1993Jan9.163011.23230@csi.uottawa.ca>
 <1jmumhINNe3d@roche.csl.sri.com>
Sender: mjohnson@netcom.com (Mark Johnson)
Organization: Netcom Online Communications Service
Lines: 48
In-Reply-To: boucher@csl.sri.com's message of 21 Jan 1993 19:48:33 GMT