> http://blog.cr.yp.to/20140205-entropy.html

Thanks for the linkage on the DJB paper, I'll check it out.

(skims)

Interesting regarding controlling a few bits of the hash.

Sounds a lot like a variation on subliminal channels:

http://en.wikipedia.org/wiki/Subliminal_channel

Aside: Oh dear, there's other RNG mailing lists?  Drat.  Need to subscribe.

I am at odds with him over pure determinism though.  I don't think his
attack is as important as the attack where one gets the state once and
defeats security forever, since his requires active interaction over a
long period of time versus a one-time confidentiality loss.

I'm interested in the RDRAND/RDSEED design that details testing
the HWRNG output directly (pre-whitening); is this it?

http://software.intel.com/sites/default/files/m/d/4/1/d/8/441_Intel_R__DRNG_Software_Implementation_Guide_final_Aug7.pdf

On Thu, Mar 20, 2014 at 07:11:35PM -0700, coderman wrote:
> indeed, this is far more common than not. (that was another aspect of
> my grief, we should have had good raw access entropy sources in all of
> our microprocessors for many years now!)
> 
> Marsh Ray conveyed a beautiful picture of why this is not so for
> entirely reasonable reasons, but it is none the less frustrating.

What's the hyperlink/msgid for that?  Would like to see.

Sorry my idea was hasty 10 years ago and I did a quick modification to
it before sending.  As you say, it needs a thorough rewrite.  No
worries on your also-hasty response.

And perhaps the whitening step before broadcast (step 5) really is a
bad idea; by sending the RNG output directly, one could have each
network client verify that it has not suffered a "normal" failure (to
whatever test it chooses to run).  It also means a passive network
monitoring device can monitor the system health.  It implies the
clients must necessarily whiten it before use to defend against
adversaries with LAN access or network sniffers, but that's good
design anyway.  Obviously no client can ascertain whether the
randomness is controlled or predictable by an adversary who hacks the
RNG box but that is the nature of the game.

Yes, one's experience and POV certainly shape the desired solution,
and there is room for many.  I've noticed my solutions change from
specific and hard-edged to more broad and softer over time.  Though I
am both happy and sad that my concerns are now widely understood, and
that I was vindicated by now-public cryptologic attacks exactly like I
described on at least 3 occasions.  I think it would be fascinating to
go back through my mail archives and see who vociferously argued that
those threats were invalid, and what other attempts they might have
made to influence opinion and debate in the crypto community.  I think
that would be a wonderful study in the challenge of finesse.
-- 
http://www.subspacefield.org/~travis/
Remediating... LIKE A BOSS


Attachment: pgpTTK5hthn3Z.pgp
Description: PGP signature

_______________________________________________
RNG mailing list
[email protected]
http://lists.bitrot.info/mailman/listinfo/rng

Reply via email to