back to article ANU puts quantum random numbers online

Last year, Oxford university demonstrated the use of quantum fluctuations to generate random numbers. Now, the Australian National University has gone a step further – putting its quantum-generated random numbers online. “Vacuum noise” is one of those “couldn’t possibly be true, except that it is” characteristics of quantum …

COMMENTS

This topic is closed for new posts.
  1. Anonymous Coward
    Paris Hilton

    DOES ANBODY UNDERSTAND THIS

    SERIOUSLY I WAS THE VALEDICTORIAN OF MY HIGH SCHOOL AND THIS ARTICLE TOTALLY CONFUSED ME UNTIL MY EARS BLED AND MY EYES STARTED DANGLING OUT OF THEIR SOCKETS RIGHT NOW IM WATCHING MY FINGERS LEAVE MY LEFT HAND AND CRAWL AWAY

    1. Captain DaFt

      Re: DOES ANBODY UNDERSTAND THIS

      These symptoms are obviously, the result of quantum fluctuations in the vaccum between your ears

    2. DRendar

      Re: DOES ANBODY UNDERSTAND THIS

      WTF is a valedictorian, and why would being one make you any more or less able to understand brand new, practically fringe science concepts?

      I actually found it relatively easy to understand what they've done, actually... I don't understand how they did it, or why it is so significant (not being a crypto nerd), but the idea of measuring the minute fluctuations in the strength of a split laser beam, which was passed through a near perfect vacuum and using those measurements to generate random numbers seems pretty straight forward to me...

      1. MrXavia

        Re: DOES ANBODY UNDERSTAND THIS

        @DRender

        The experiment sounded quite simple to me, I am surprised no one has thought of doing it before, surely its obviously a great way to generate random numbers (and I am no crypto nerd either, I just know more randomness is good for cyphers, the rest I leave up to the experts)

        1. Anonymous Coward
          Anonymous Coward

          Re: I am surprised no one has thought of doing it before

          Thought of it yes. the idea has been around for a while is easy and simple. The measurement is not.

          Recording and calculation has to be done literally at the speed of light. As the article mentions the fluctuations occur about a mean, so if you have too much duration in the digital recording point all you get is the mean itself, not the difference. Kudos goes to these guys on getting their measurement precise enough to be useful.

      2. Kubla Cant
        Megaphone

        Re: DOES ANBODY UNDERSTAND THIS

        @DRendar: "WTF is a valedictorian?"

        Apparently, SOMEBODY WHO SHOUTS ALL THE TIME!

    3. Anonymous Coward
      Anonymous Coward

      Re: DOES ANBODY UNDERSTAND THIS

      I thought it was fairly clear what they were doing...

      I don't fully undersand it, but I get the principles...

      Are you green and live under a bridge? or bad product of the US education system?

    4. Ru
      Unhappy

      Re: DOES ANBODY UNDERSTAND THIS

      If you can't resist troll feeding, please at least concentrate on funny trolls. This guy is just dull.

  2. Destroy All Monsters Silver badge
    Trollface

    How to generate classical random bits from lolcat fluctuations

    1) Take all possible pictures of Schrödinger's lolcats and arrange them in a box.

    2) Assign a selection probability to each picture, exponentially decreasing with the fatness of the lolcat, but so that the sum of the probabilities over all pictures still yields at most one.

    3) Select a picture at random, using the previously assigned selection probabilities.

    4) If the tail is on the left, emit 1. If the tail is on the right, emit 0. Otherwise, emit NULL (and restart the experiment.)

    Now use lasers somehow and there you are.

    1. Notas Badoff
      Facepalm

      Re: How to generate classical random bits from lolcat fluctuations

      Step 3 requires a random number. Thus you are pulling more tails than you have cats. The cats now have lasers and are annoyed. You may not see them before they see you. Good luck Dr. Frankenkatzen.

      1. Destroy All Monsters Silver badge
        Unhappy

        IMPOSSIBRU!!!

        Foiled again!?

    2. amanfromearth

      Re: How to generate classical random bits from lolcat fluctuations

      Too complex.

      The simple solution is to attach a piece of buttered toast to the back of the cat. Throw the cat in the air and see which way it lands.

  3. Anakin
    FAIL

    Are they random or not?

    "Anybody downloading random numbers from the site, the ANU says, will receive a unique sequence different from all other users"

    If they guarantee that, the series of number can't be random

    1. Anonymous Coward
      Thumb Up

      Re: Are they random or not?

      Very very true indeed. Truely random has no garantee at all on the values and as you said they can't make any bold statements like garantee it.

      All I need to know about random numbers is that if I get a lucky dip lottery ticket my numbers will always be close bbut not the actualy winning numbers. This is one area of randomness I await to be proved wrong :).

    2. Chris Miller

      Re: Are they random or not?

      I think you're making a valid linguistic point rather than a practical one. If (as I would guess) they have a long list of previously generated random numbers, what they're saying is that they will only issue them once - similar to the traditional 'one time pad'.

      If you and I both request a 4096-bit string from them, there's a 1 in 2^4096 probability that we'll both end up with the same string - but that's equal to zero for all practical purposes.

      1. MrZoolook
        Thumb Down

        Re: Are they random or not?

        Quote:

        I think you're making a valid linguistic point rather than a practical one. If (as I would guess) they have a long list of previously generated random numbers, what they're saying is that they will only issue them once - similar to the traditional 'one time pad'.

        So that's not random. That's just picked randomly from an ever decreasing range of valid outcomes. Eventually (though not in our lifetime admittedly) that range of valid choices will reach 0, and then either the previously selected outcomes are made available for re-selection (and the numbers are not unique anymore) or the only choice left is invalid or 0 or whatever the last number picked was (and the choices are also therefore not unique or just invalid).

        Either way, randomness by it's nature decrees that a previously picked number has the same equal chance of being picked 100 times in a row as 100 different ones being picked once. Guaranteeing a unique random number is impossible. As soon as you put any kind of rule on what numbers can or cannot be picked, it isn't random!

        1. lomuk

          Re: Are they random or not?

          'Either way, randomness by its nature decrees that a previously picked number has the same equal chance of being picked 100 times in a row as 100 different ones being picked once'

          Not quite.

          More like this:

          If we're generating random numbers in the range 1 - n, the chance of the first being a specific number is 1/n, but of being a 'different' number is 1, since we haven't picked any yet.

          For the second number, the chance of it being the specific number is 1/n again, but of it being different from the first is (n-1)/n.

          And so on (1/n;n-2/n)...

          ...which doesn't result in an equal chance.

          Maybe that should read:

          'If a series of numbers fits the definition of random, then a previously picked number has the same chance of being picked 100 times in a row as 100 specific different ones being picked once'

    3. JDX Gold badge

      Re: Are they random or not?

      It's a unique 'capture' for each user, they don't send the same data to multiple users... clearly different 'captures' can be identical based on random probability.

    4. Unicornpiss
      Pint

      Re: Are they random or not?

      -They are guaranteed. If you find someone else that got the same string of random numbers you have, you get a full refund.

  4. Peter Murphy
    Thumb Down

    Suggested Alternate URL for the ANU Quantum Random Numbers Server

    http://physics0054.anu.edu.au/

    For most people, this would resolve to the same site as http://150.203.48.55/index.php

    I provide this for the unfortunate few who work for firms so anal that they automatically censor web addresses that are expressed in IP form.

    Like mine.

    (Can't say who they are, but they work in the "educational" sector.)

  5. Christian Berger

    Is that even a problem?

    I mean random number generators can be done trivially by using thermal noise, or even lots and lots of easily drifting oscillators or other ways. Both can be integrated into cheap integrated circuits.

    So I don't really see the point in this.

    1. Richard 12 Silver badge

      Re: Is that even a problem?

      For a start, it's cooler.

      Secondly, the oscillator technique is psuedo-random - that's a chaotic system, not a random one.

      Thermal noise can be affected by external sources as it varies by temperature, so not properly random.

      Until now the only really high quality self-contained RNGs are based on radioactive decay events, so produce bits fairly slowly and are hard to buy.

      This apparently makes high quality randomness really quickly.

      1. Christian Berger

        Re: Is that even a problem?

        Not exactly. You can also affect quantum processes by external conditions like the temperature. That's not the point. The point is that there's still enough randomness to work with. And oscillators which randomness essentially boils down to the shot noise, as well as shot noise in diodes is a quantum process.

  6. This post has been deleted by its author

    1. Richard 12 Silver badge

      Oh dear, you confused theory with hypothesis.

      Gravity is only a theory, yet you wouldn't jump out of a high building naked.

      In science, a theory is something that fits all known observations better than anything else.

      Quantum theory fits our observations of really tiny stuff incredibly well.

      1. Anonymous Coward
        Anonymous Coward

        Re: Oh dear, you confused theory with hypothesis.

        I always wear clothes when I jump out of high buildings.

    2. Crisp

      Re: Beavis, they said, “guaranteed by quantum theory." Heh, heh, heh!

      Gravity is a theory that's been proven too. For empirical evidence. Throw a brick.

      It'll hit the floor, guaranteed. (Unless you can throw it faster than 11 Kps)

  7. Arthur the cat Silver badge
    FAIL

    Anyone who really understands cryptography

    would never download supposedly random numbers from a web site. They might be lying about the randomness, they could be mistaken, or they could be completely genuine but the site has been hacked or its network connections compromised. Unless *you* control the random number generator you cannot trust it.

    1. Bronek Kozicki

      Re: Anyone who really understands cryptography

      "Unless *you* control the random number generator you cannot trust it."

      .... for cryptographic purposes. And then how do you really know that you control it, if someone else built it for you?

      Otherwise I see nothing wrong with it. In fact I am thrilled to have access to "properly random" (quotes only because of factors you mention) data in real time as fast as my network connection allows.

      Would be even better to have cheap hardware in hand, to do it without involvement of network connection.

      1. Arthur the cat Silver badge

        Re: Anyone who really understands cryptography

        You're quite right about using this source for Monte Carlo or simulation applications, it's a good resource for that, but the original article specifically said

        "It is, however, very useful for applications like cryptography."

        I just wanted to warn against thinking the ANU site could be safely used for cryptography.

  8. amanfromearth

    Oh-oh

    A memory stick with random numbers on it..

    "Now Sir, we need you to give us the password to decrypt this sequence of numbers on your memory stick"

    "But it's totally random"

    "They all say that Sir. Just put your hands in these nice bracelets"

  9. Christian Berger

    One should make a business of selling random numbers...

    That would be a billion dollar business model. :)

  10. Anonymous Coward
    Anonymous Coward

    Hahaha

    Might work.

    But how would you guarantee that the random numbers were in fact random, and not just the same CD copied over and over with a random wraparound location change?

    I always wondered whether this would work for making a truly unbreakable wi-fi link, have one random number store at the transmit end and a second at the receive end. Have the current location within RNDspace change with a temporal key sequence, so that the message can't be decoded by a third party.

    Overwrite the data in the chip(s) perodically so that there is no way to recover the RND.

    With >64GB memory chips this equates to many hundreds of hours of typical Internet use, but standard wifi points could act as relays for the data stream without being able to decode it

    Combining this with the stratellite (aka balloon wifi) system or simply optical, you could upload your data to SkyNet and store it in the clouds.

    Each balloon could copy the data to its neighbours and eventually all of them in the network would have copies of the data or one section of it.

    With 1TB of SSD costing not much this is very handy indeed for distributing files over large areas without using the existing slow, unreliable CreakyBroadband (tm)

    AC/DC

  11. John Gamble
    Boffin

    Yet Another Site For Randomness

    For most of us though, this just isn't critical. The "truly random" sources are usually slow enough that they're main purpose is to seed our cryptographically secure pseudo-random number generators, not to be our cryptographically secure pseudo-random number generators.

    And even then, they're more used for the "coolness factor" (as Richard 12 points out) than for practical reasons.

    Other sites that have done this: Lavarnd (<http://www.lavarnd.org/>, yes, randomness from a lava lamp), Hotbits (<http://www.fourmilab.ch/hotbits/>, "Genuine random numbers, generated by radioactive decay"), and Random.org (<http://www.random.org/>, which uses atmospheric noise).

This topic is closed for new posts.

Other stories you might like