[Cryptography] True RNG: elementary particle noise sensed with surprisingly simple electronics

[David Johnston]

On 9/18/16 12:55 PM, sebastien riou wrote:

> "Shrinking entropy sources comes from (a) better circuits and (b) better
> extractor theory"
>
> Could you point to some literature ?
>
This paper
     http://www.deadhat.com/papers/uRNG.pdf is one we built, based on a 
circuit we developed and a paper I found.
It was and may still be the smallest, most efficient secure RNG in terms 
of joules per bit and bits/s/W. It started with this youtube video:
     https://www.youtube.com/watch?v=ZzsFb-6wvoE (skip to minutes 41-42)
leading to the paper:
     http://www.boazbarak.org/Papers/msamples.pdf
which itself has lots of references worth following.
The circuit was described here: 
http://ftp.ece.villanova.edu/perry/cyber-sec/references/IEEE-VLSI-Circuits-2010-Random.pdf

Since then there have been lots more papers on extractor theory, some of 
which express efficient structures, some don't. Most are unreadable to 
non mathematicians and this has limited the rate of adoption. Here's one 
- I defy you to find the extractor in there in under 5 minutes and then 
describe it in actual operations on bits and bytes: 
http://research.microsoft.com/en-us/um/people/yael/publications/2009-2-source.pdf

This pair of papers made the news in 2015, but I never could work out 
what the actual algorithm was. It reads like they proved a function 
exists, but don't define any specifics. It's on my list of papers to 
decode. We've had good (by my definition of good) 2 source extractors 
since 2006 and I don't know what makes these better:
http://eccc.hpi-web.de/report/2015/119/
https://arxiv.org/pdf/1508.01115.pdf

I'm not asserting that the improved extractors are recently described, 
I'm saying it took a good decade for anyone to notice, decrypt the terse 
descriptions and start building them in silicon.

I suspect the largely useless Von Neumann whitener and the related Yuval 
Perez whitener were so often used and used wrongly (E.G. with serially 
correlated sources) because the algorithms in the papers were 
comprehensible to engineers, but the preconditions were not.

DJ