[Cryptography] combining lots of lousy RNGs ... or not

[Tom Mitchell]

On Tue, Nov 22, 2016 at 12:03 PM, Ron Garret <ron at flownet.com> wrote:
>
> On Nov 22, 2016, at 7:13 AM, Phillip Hallam-Baker <phill at hallambaker.com>
> wrote:
> > On Mon, Nov 21, 2016 at 5:53 PM, John Denker <jsd at av8n.com> wrote:
> >
> >> Here are some useful equations:
>


> > ​A better equation is H (random + squish) = random​
>
> Talk about burying the lede.
>
> If you XOR, then
>

A thought -- there are only two parts.
The message and a function that generates a table of bits while
operating on the message.

The function to be of value in cryptography needs
to be near impossible to predict and sufficiently
random that the message is not exposed to other attacks.

The internet today has lots of identical hardware running the
same software so a solution that works for two individuals
must also scale to billions.

Analysis needs data and billions of systems generates masses
of data to inspect and explore.    A couple of individuals communicating
do not generate much data so the analysis of their black box method is
harder.

For two individuals to take a couple common random tools
and mix the results a little or a lot would make them harder
to attack.   For the billions of systems to do this exact same thing
generates masses of data for inspection and analysis.

The analysis that an adversary might employ can be costly.
To look at two individuals is expensive while big bulk samples
may cost the same.   Their value equation is the pot of gold
that a flaw in the method can be widely exploited or deployed
on a target on demand.

In the news, monoclonal hardware...
Microsoft reveals minimum PC specs for Windows 10 VR headsets
Look at other "minimum" specifications starting with silicon and
the boot process.  :(















-- 
  T o m    M i t c h e l l
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.metzdowd.com/pipermail/cryptography/attachments/20161122/bcadad3f/attachment.html>