[4073] in cryptography@c2.net mail archive
Re: Intel announcements at RSA '99
daemon@ATHENA.MIT.EDU (Arnold G. Reinhold)
Wed Jan 27 10:49:48 1999
In-Reply-To: <199901262059.NAA29068@nyx10.nyx.net>
Date: Tue, 26 Jan 1999 17:17:06 -0500
To: Colin Plumb <colin@nyx.net>
From: "Arnold G. Reinhold" <reinhold@world.std.com>
Cc: cryptography@c2.net
I agree strongly with your position that the raw bits should be available
from any on-chip randomness generator and that whitening can be done in
software. I would further suggest that the underlying physics of the random
process be evident in some fashion. For example, it might actually be
better if the 0/1 ratio were slightly temperature dependent. This would
provide an additional audit mechanism.
I do not agree, however, that 1 bit per second would be fast enough. I am
not sure what you mean by " Anything more can be generated by cryptographic
means," but I want a reliable source of entropy that is independent of the
many unproven mathematical assumptions underlying cryptography. Also we
should not assume that we forsee all possible uses for this source of
randomness. For example, people doing analysis using Monte Carlo methods
might have use for a fast source of near-random bits that could be easily
whitened in software.
Arnold Reinhold
Consultant and Author
reinhold@world.std.com
