Well, that was a long post, Marsh!
I think it is a good perspective. And it occurs to me that if this is a
real problem there might be a real solution.
I suggest going to NIST and asking them to run a design competition for
a hardware cell that produces good entropy. Hardware designs aka cells
aka asics aka idk what they call them are often standardised products
these days. You pull one from a library, lay it in a corner, connect up
the lines on your CAD tool and you're done.
Our problem is what to design, what to layout, and how to make it good?
NIST have done well with the competition technique. AES was a good
thing, it brought in 30 designs and the world's cryptographers in one
goal to find the best of the best.
Either way ... where the expertise is unclear and the problem is real
and definable and also of widespread interest, a competition for a
design might get the grey matter churning. EEs get to play this time!
NIST recently produced a new standard for PRNGs that kicked out the
entire entropy question. The goal is a deterministic PRNG, testable and
repeatable. It took me a while to figure it out, but this separation
from the old "all-in-one" thinking over to entropy source plus
deterministic mixer is quite inspired. Point being, they solved half
the problem; they'll be open to the other half?
iang
On 23/02/12 08:55 AM, Marsh Ray wrote:
On 02/22/2012 09:32 AM, Thierry Moreau wrote:
While commenting about...
_______________________________________________
cryptography mailing list
[email protected]
http://lists.randombit.net/mailman/listinfo/cryptography
