[Cryptography] Is Ron right on randomness

[Tom Mitchell]

On Mon, Nov 28, 2016 at 9:38 AM, ianG <iang at iang.org> wrote:

> On 28/11/2016 11:46, Bill Cox wrote:
>
>> On Sun, Nov 27, 2016 at 11:02 AM, ianG <iang at iang.org
>> <mailto:iang at iang.org>> wrote:
>>
>>     On 26/11/2016 09:38, Salz, Rich wrote:
>>
>>             Absolutely right.  Only TRNGs that make raw data available
>>             should be trusted.
>>
> ....

>         Meanwhile, back in the real world...
>
> ....


> Ian, would you agree that something on the platform needs to first
>> ensure that /dev/random is well seeded before OpenSSL reads from
>> /dev/urandom?
>>
>
> Yes absolutely!  That is a platform responsibility - see the thread with
> John Denker where he says:
>
>
The startup case is the most fragile.
This is the easy to attack -- closest to the butterfly taking wing case.

The number of "true" random bits is interesting.
On the internet there are a lot of devices and as long as the
vast majority have a lot of good bits those that have less will
be invisible as long as they are not identifiable as a class.
Specifying twice what you expect makes sense to compensate
for weaker hardware in the connected hardware population.

The more valuable the data the more quality needs to be specified.
Thus the interface needs to allow the likes of the NSA, Google,
Amazon.... to attach better and unique devices.

Modification of hardware like laptops is getting harder.
USB device have risks that invites epoxy.
Apple is all dongles too easy to swap or misplace.

Portable hardware is IMPORTANT as it is the remote and portable access
for almost all workers today.   (Navy data breach, lost laptop this week).
Phones are used more and more for access.    The critical links
are increasingly portable.

Portable devices are the connection point and need to be as well
secured at the servers.

Exotic packaging for thin and light notebooks and sealed phones
are all difficult to inspect yet special tools and tricks allow it to be
repaired
and modified.  The big boys could modify devices in transit no mater the
maker.

Sandboxes and Virtual machines confound easy hardware solutions
with a layer of abstraction.

BIOS "white lists"  limit the exchange of one WiFi device for another
and thus hobble the ability to design and deploy a WiFi+entropy device.

Sensors,  GPS receivers do see a lot of positional uncertainty.
Compensating data from reference sites is one interesting
regional wiggle.  A GPS receiver sitting on a window can
draw a target that will center on one location but any individual
read will have entropy in the low bits.
In Texas..."high-accuracy base stations referred to as "Regional Reference
Points" (RRPs). TxDOT
is participating with the RRPs in the National Continuously Operating
Reference Stations (CORS)
network operated by the National Geodetic Survey (NGS). The RRPs are based
on the North
American Datum (NAD) 83 (EPOCH 2010.0) coordinates and are used to improve
positional
accuracy through post-processing techniques. "
http://ftp.dot.state.tx.us/pub/txdot-info/isd/gps/











-- 
  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/20161128/7413c53f/attachment.html>