Quantum Random Numbers For Download

PSUdaemon writes "The University of Geneva has produced a website that allows you to download truly random numbers generated from an Optical quantum random number generator. They will also be releasing a client API that you can use directly in your codes to download random numbers."

Re:Interesting, but not that useful

[zulux]

If you felt paranoid enough to need quantumly generated random numbers, would you really get the numbers over the internet from an untrusted source?

Even if this source of randomness is compromised, adding it to your already existing sources of randomness coulden't hurt. It's best to layer sources of randomness on top of each other - so if one source or two isen't random, the whole stack of randomness isen't compromised.

Re:This is SOOOOOO Bad

[JabberWokky]

I would imagine that it's suffering from overuse right now. Top story on Slashdot and all.

However, yes... you can trust this to be random, and no, you can't trust it to "correctly destroy all of the information between here and there".

I don't believe that the intent of this is to do realworld crypto nor games (which is what other people are claiming the other "major" use of random numbers are). A set of purely random numbers is really only useful to people testing mathematic theories or other high math science work. For crypto, decent pseudo-random sequences (or the old "pull from an analog source" trick) is perfectly fine. This is overkill for realworld crypto (not to mention broadcast via the internet), which means that this is primarily useful - to math scientists.

--
Evan

Re:Single point of failure

[tolan-b]

They will also be releasing a client API that you can use directly in your codes to download random numbers.

I suspect that will be encrypted and involve public key swapping to avoid man in the middle.

Re:Not bad, not bad...

[image]

> Not too terribly bad of a distribution to my eye.

It's hard to tell.

Here's a simple perl script to demonstrate:

#!/usr/bin/perl -w

my ( $count, $low, $high ) = @ARGV;

defined $count and defined $low and defined $high or
die( "USAGE: rand.pl count low (inclusive) high (exclusive)\n" );

my $x = 0.5;
my $r = 3.6;

for ( $i = 0; $i < 256; $i++ )
{
$x = $r * $x * ( 1 - $x );
}

for ( $i = 0; $i < $count; $i++ )
{
$x = $r * $x * ( 1 - $x );
my $bits = ( $x * 1000 ) - ( int ( $x * 1000 ) );
print int( ( $bits * ( $high - $low ) + $low ) ) . "\n";
}

Try saving that as rand.pl and running it like this:

$ ./rand.pl 100 0 10

Seems like a very random distribution, doesn't it?

However, in reality, it's merely chaotic. It is based on the logistic function over a known chaotic region (discarding the first few significant bits to adjust for the range of x). The point being that it is very difficult to know when something is truly random or not.

You can read more about the logistic function here [wolfram.com] and random numbers here [wolfram.com].

Re:How do they know ?

[astroboscope]

How do they really really know for certain, that these numbers are truly random?

There are statistical tests (see Knuth), like spectral flatness and incompressibility, but complete "certainty" has to rest on the theoretical underpinnings of QM, with testing by Bell's inequality (discussed elsewhere on this page).

And what kinds of applications might they be used for?

Secure communications, ignoring for now the problem of distributing the random bits.

Why does it need to be a quantum random number generator? How come you cant use an aerial and pick up white noise?

That "white noise" is contaminated by a jumble of deterministic TV and radio signals, that potential attackers could also detect or predict. It would be better to detach the aerial and amplify the output from a warm resistor, which is I think what the VIA motherboards do. Conceivably, though, somebody with far too much time and money on their hands could watch (or have watched) the molecules in your resistor unreasonably closely, and attempt guessing what they'll do. Using smaller particles is better, since observing them perturbs them so that their behavior can't be predicted, but that's QM by definition. You (and VIA) could also argue that warm resistors already include lots of quantum noise.

Re:Truly Random Number ?

[noodler]

"Consider Heisenberg's uncertainty principle: the more precisely the position is known, the less precisely the momentum can be known"

this only exists because we have limited ability of measuring.,
any measurement on that level will influence the thing you are measuring.,
so we invented the theory of quantum mechanics to describe this phenomenon.,
but in fact we don't know for sure what's beyond this because we have no instruments to measure it.,
it's our own inability to not interact with a sample that shapes quantum mechanics! never forget that.,

this whole has led to many many proposterous theories including the collapsing states theorie and multiverse theory.,

heisenberg talks about propabilities but propability is inherently an abstraction of large quantities of things.,
there is always a system underneath propabilities wether we are able to see it or not.,

you also need to remember that all science is an APPROXIMATION of truth., our perception of truth will ALWAYS be limited by the quality of our measurements.,

since our measurements are imperfect so will our understanding of the truth be imperfect.,

i'm not saying that quantum mechanics does not work, just that it cannot describe the whole truth., in fact quantum mechanics 'works' because it limits itself to what we ARE capable of measuring., but there is no reason whatsoever to assume that what me measure is complete.,

the fact is that we are quite uncapable of describing every aspect of the universe we live in.,
live with that., :)