Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

'Advantages' of Quantum Encryption

  1. Oct 5, 2012 #1
    I'm finding it absurdly difficult to get detailed information about the exact methods of transmitting quantum encrypted messages. I have a working understanding of quantum mechanics but I'm not a computer scientist. The most I can find is descriptions of QKD which use a quantum channel as well as a public/classical channel to generate a shared key. But... then what? Do the sender/receiver communicate through a classical channel like normal just using the newly generated key? What advantage does this have over a randomly generated predetermined key? That is to say, what use is the QKD in the case that both parties can agree upon a true random key beforehand?

    The main point hinges upon the inherent inability to transmit new information without the use of a classical channel. The best I can imagine is a scenario where two parties are essentially two halves of a delayed choice quantum eraser and then communicate through a classical channel to establish coincidence. After which point the coincidence data would have to be encoded.

    Can someone help me out here? I'm not seeing what the big fuss is about
     
  2. jcsd
  3. Oct 5, 2012 #2

    jedishrfu

    Staff: Mentor

  4. Oct 5, 2012 #3
    I've read that article thoroughly, and as I mentioned before I understand quantum key distribution (which is all that article talks about). But look at the "Man-In-The-Middle" Attack section to see what I mean. The only way to actually be secure is to use a shared key like an OTP to authenticate communication, at which point the quantum encryption is only as strong as the initial shared key... which essentially makes the quantum encryption pointless. Eavesdropping can only be found out when its done to the quantum channel, but for QKD both parties need to communicate through a classical channel... which can be eaves dropped on freely. And both parties need to communicate literally all of their information through the classical channel which seems to then completely defeat the purpose of having 'secure' quantum channels at all...
     
  5. Oct 5, 2012 #4

    jedishrfu

    Staff: Mentor

  6. Oct 5, 2012 #5
    In QKD it only distributes a key so that you can communicate by a classical channel.

    As long as the classical channel is public I don't see how man in the middle systems can break it. But you do need some kind of authentication system for the classical channel. That should not be hard with a public channel.

    You don't need an initial key in order to set up QKD but even if you did it would still be as safe as one time pad. You would just have a much shorter pad that lasted forever.

    Even with pure quantum encryption you would still need a public classical channel.
     
  7. Oct 5, 2012 #6
    Thanks Jedi that was really informative.

    And thanks ppnl that verifies what I thought.

    The point I'm making is that the quantum encryption, at the end of the day, is only as effective as the encryption for the classical channel. In order for Alice and Bob to compare their quantum analyses they have to openly share their results with one another through a classical channel. If this is intercepted then the jig is up: Eve would have both measurements and would be able to generate the same key that Alice and Bob do after they do their comparison. And since a classical channel doesn't adhere to indeterminacy, eve could intercept and then push forward any data passed through the classical channel. Therefore the classical channel needs an encryption all it's own, the breaking of which folds the system. So... why not use this intial encryption and the classical channel to begin with... instead of using this intial encryption to generate a new encryption?
     
  8. Oct 5, 2012 #7
    You are right that "all" quantum cryptography does is to create a random secret shared key. But that is also all that is needed for unbreakable encryption!

    Note that when generating the shared key, the only information Alice and Bob communicate over the classical channel (assumed public and unencrypted) is their choice of basis for their measurement. They don't communicate their actual results (0 or 1). When they happen to have chosen the same basis (50% of the time) they keep their secret result as part of their common key. They can also sacrifice a fraction of their generated key to make arbitrarily sure that no one else have eavesdropped on the actual photon exchange.

    In the end, what this does is to give Alice and Bob a truly random key that they can know for certain no one else has. They can also continue to generate more and more bits giving them a key of arbitrary length. Then, to do the actual communication, they use this key as a one-time pad for the message they want to encrypt (using the same length of the key as the message). This is actually an unbreakable encryption, and the encrypted message can be sent over any public channel.

    You can certainly achieve the same unbreakable encryption by agreeing in forehand of a long, secret, random key. That is just not as easy as it sounds in practice! You will always have the tricky logistic problem of how to distribute that key safely to the other part (and store it safely too). QKD "just" solves the problem of generating and distributing the key in an effective and fully secure manner.
     
  9. Oct 5, 2012 #8
    Thanks Dali,

    For some reason I was missing the fact that the shared information is only about the chosen basis of each party, not the resulting 1/0 associated. So that definitely clears that up thank you!
     
  10. Oct 6, 2012 #9
    When quantum encryption is widely used, encryption methods used today will be flawed because of the ability of quantum computers to easily factor very large numbers. Therefore, in that sense, an advantage will be that quantum encryption is the only option! :)
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook