Re: Quantum cryptography
I'm taking the liberty of forwarding this note here because email to jdblair@rogue.cas.muohio.EDU bounces so I couldn't thank John or ask for permission. It was the only reply I got, but I think it is really interesting so... - Joi
From: jdblair@rogue.cas.muohio.EDU (John Blair) Message-Id: <9401280824.AA08537@ rogue > Subject: Re: Quantum cryptography To: jito@netcom.com (Joichi Ito) Date: Fri, 28 Jan 1994 03:24:41 -0500 (EST) In-Reply-To: <9401280501.AA02095@iikk.inter.net> from "Joichi Ito" at Jan 28, 94 02:00:22 pm X-Comment: The enlightened man is one with the subject. -Mumon X-Mailer: ELM [version 2.4 PL21] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 3259
Does anyone know how quantum cryptography works?
- Joi
-- true name: <Joichi Ito> closest email address: <jito@netcom.com> closest fax number: <+81-3-5454-7218> current physical location: <Tokyo> travel path: <.> mosaic home page: http://iikk.inter.net/ -- finger jito@netcom.com for PGP Public Key, RIPEM Public Key --
Joi,
Chances are you're recieving a lot of answers to this, but here's one anyways.
This comes from: Schneier, Bruce, _Applied Cryptography_, pp 408-410, John Wiley & Sons, Inc, New York. pp 408-410
Quantum cryptography relies on the property that one cannot know the position and the velocity of a particle at the same time to prevent evesdropping. Any evesdropper will neccessarily disturb the message in a detectable way. Experimental quatum cryptography relies on polarized light.
If a pulse of horizontally polarized photons is sent through a horizontally polarized filter, 100% of the light gets through. Pass it through a filter 45 degrees off, and each partical has a 50% chance of making it through, and no chance of making it through a filter 90 degrees off.
1) Alice sends Bob a string of photon pulses. Each of the pulses is randomly polarized in one of four directions: horizontal, vertical, left-diagonal, and right-diagonal.
For example, Alice sends Bob: ||/--\-|-/
2) Bob has a polarization detector. He can set his detector to measure horizontal and vertical polarization, or he can set his detector to measure diagonal polarization. He can't do both; quantum mechanics won't let him. Measuring one destroys any possibility of measuring the other. So, he sets his detectosr at random, for example: x++xxx+x++
When Bob sets the detector correctly, he records the correct answer. The rest of the time, he gets a random result. Say he gets: ||/--\-|-/
3) Bob, over an insecure channel, tells Alice what settings he used.
4) Alice tells Bob which settings were correct. In this example, 2,6,7,9.
5) Alice and Bob keep only those settings that were correct. In this example, they keep: *|***\-*-*
According to a pre-arranged code, such as 1 = horizontal and left-diagonal, and 0 = vertical and right-diagonal, they have sent the bits: 0011
Alice will need to send 2n pulses for every bit transmitted, since Bob will be correct 50% of the time, on the average. These bits can be taken as the key to a conventional algorithm, or enough can be sent to generate a one-time pad.
6) Alice and Bob compare a few bits in their strings. If there are discrepancies, they know they are being bugged. If there are none, they discard the bits they used for comparison and use the rest.
For a good overview (according to Schneier) see:
C. H. Bennet, G. Brassard, and N. D. Mermin, "Quantum Cryptography," Scientific American, v. 68, n. 5, 3 Feb 1992, pp. 557-559
C. Zimmer, "Perfect Gibberish," Discover, v. 13, n. 12, Dec 1992, pp. 92-99.
Schneier says British Telecom has used this system to send bits over a 10km fiber-optic link.
see: W. Brown, "A Quantum Leap in Secret Communications," New Scientist, n. 1585, 30 Jan 1993, p. 21.
Hope this made sense. Its really quite amazing to think about. -john.
-- true name: <Joichi Ito> closest email address: <jito@netcom.com> closest fax number: <+81-3-5454-7218> current physical location: <Tokyo> travel path: <.> mosaic home page: http://iikk.inter.net/ -- finger jito@netcom.com for PGP Public Key, RIPEM Public Key --
participants (1)
-
jitoï¼ netcom.com