[4087] in cryptography@c2.net mail archive
Re: Hardware Random Number Generators
daemon@ATHENA.MIT.EDU (Martin Minow)
Thu Jan 28 15:30:07 1999
In-Reply-To: <199901281606.LAA25185@pollen.cs.concordia.ca>
Date: Thu, 28 Jan 1999 12:02:16 -0800
To: cryptography@c2.net, MCKAY john <mckay@cs.concordia.ca>
From: Martin Minow <minow@pobox.com>
>This I got from computer historian, Simon Lavington.
>
>The (Manchester) Ferranti Mark I had a hardware random number generator.
>This was specified by Alan Turing - (A copy of his original
>Internal Report, dated 1949 I believe, still exists.) ...
For what it's worth, Illiac 1 (also a first-generation computer) had,
unintentionally, a hardware random number generator. Recall that
Illiac memory was based on a Williams Tube. This was an electrostatic
display tube. To write a 1-bit to memory, the computer directed the
electron beam to the bit's x-y coordinate. To read memory, it directed
the beam (at a lower energy?) to that spot. If a bit was previously
set, a pulse could be detected by a copper plate affixed to the front
of the tube. (This is from memory: I apologize for any errors).
Williams tube memory suffered from charge leakage. Imagine a tic-tac-toe
configuration. Set the center bit to zero. Now, if you write 1-bits into
the 8 exterior locations, there was a low probability of incorrectly
reading a 1-bit from the center. On Illiac, about 400 cycles would be
needed before the center bit changed. This was not considered good
programming practice (!) but could be used to generate a few
random bits numbers by recovering the precise number of cycles.
Martin. (walking repository of Illiac lore: it was my first computer)
minow@pobox.com
