Slashdot Mirror
Quantum Security
Triode writes "In this months issue of Physics Today there is a very
interesting read entitled 'From Quantum Cheating to Quantum Security' which delves into encryption. Talks about ads and
disads of popular encryption (keys, public keys, DES etc), the
size of current encryption and why it is not (theoretically) good.
Quantum computers could make breaking our current methods of encryptoin easy, so we need to start now with methods of encrytption that would not be so easy. A pretty basic example of a implementation of the B92 protocol is given using a single photon source over a 48km optical fiber. Worth a read.
Check it out at the AIP website."
This is the best walk-through of quantum encryption I've seen, and one of the few that points out the flaws and unknowns which could plague a completed system in the real world. And depressingly enough, there is a note on the Physics Today main page which reads: "All editorial content from the magazine is available on the Web. In the near future, restrictions will apply." As a selfish site junkie, I hope this only means NYT-style registration, not WSJ-type subscribers-only service.
University of Montreal with Gilles Brassard
or
McGill University (also in Montreal) with Claude Crepeau
Both have fairly well known Theoretical Cryptographs in their CS departments that do research in Quantum Cryptography. However, the Quantum physics part is mainly left up to you; That is: you don't need a College Degree in physics to do Quantum Cryptography (some would say it would help). Quantum Cryptography at its core is still only algorithmics like Classical Cryptography but based on a different set of tools then what you're used to.
Mr. Brassard just finished writing a book on quantum cryptography; I'm not sure but I believe it's out on the market currently.
Your second question was to whether or not its more suited to a Math major. Both of these gentleman will tell you that Maths are a big part of any crypto. Having a strong background in math is definitively a plus; in the last few years, doing a double major Math-CS or Math-Physics has been the typical path for people that work with them.
Your third question was to whether or not there were job openings with such requirements. The answer is: Yes, in academia; More or less in large companies' research labs (i.e. IBM labs, Lucent, MS, NEC, etc.); pretty much No anywhere else (there might be a few expections.
However, doing grad studies in CS can hardly be considered a waste of time and you should have no problem finding a job after. Whether or not you'll still do Quantum Crypto is another question.
For what it's worth, they both also work or have worked on other fields of crypto such as Zero-Knowledge proofs (nothing to do with the company that ripped the name from this field of study) and other VERY theoretical aspects of crypto.
Hope this helps.
It isn't as bad as all that. According to the article, a quantum codebreaking machine will have to perform a computation of order O(sqrt(n)), where n is the number of possible keys, in order to solve the problem. Classical brute-force searches are, of course, O(n)
This is a quantum method of breaking DES encryption. The method for breaking RSA and other schemes based on factoring being difficult offers an improvement from exp(O(n^1/3 (log n)^2/3)) to 0(n^2 log(n) log(log(n)) ) which is gigantic.
"If computers that you build are quantum,
Then spies everywhere will all want 'em.
Out codes will fail,
And they'll read our email,
Til we get crypto that's quantum and daunt 'em."
(Jennifer and Peter Shor)
:wq
The net effect is that a quantum computer in the hands of an eavesdropper halves the effective keylength - a 128-bit key is reduced to 64 bits of effectiveness. 64 bits is, of course, not enough security to defend against government-level surveillance resources, but all that has to be done to solve the problem is to increase the keylength to 256 bits.
One of the requirements for the AES candidates was that the algorithm support 256-bit operation. Rijndael, the heir apparent to DES, does support 256-bit operation modes.
...would be to double major in something like EE or CSE and quantum fizzicks. Has anyone ever done this? Were you successful in getting a job that related to both fields? I know that the two departments at my school (Univ. of Washington) would basically not cooperate to let me do this. Maybe this cryptography application would be better suited to a math major - that might be easier to combine with a physics degree.
It seems that half of the world is trying to develop new methods of encryption, wheras the other half is busily trying to break them.
Wouldn't it save everyone a whole lot of effort if everyone sent everything in clear?
This reminds me of a conversation I had awhile back with a fellow geek. He thought that new quantum computers would make an entirely new class of 'haves' and 'have nots', based on the ability to encrypt your information
In a nutshell, once these computers are actually in production, the government will be the first to have them. No current X86 (or such) system will be able to make an unbreakable cypher anymore. No countries, no indivduals, or such. The only people able to make such will be those with these quantum computers, which will most likely be regulated.
The entire idea behind 'privacy through encyrption' will be a thing of the past. True, most crackers won't have access to this equipment. But the NSA, CIA, etc will, and they will be albe to crack any encryption you can throw at it.
Maybe we DID take the blue pill. You wouldn't remember anyway.
We could start by misspelling everything, thus making our communications harder to understand. Slashdot has employed this encryption method for years.
--
``Life results from the non-random survival of randomly varying replicators.'' -- Richard Dawkins
Current encryption is strong, but not infallible. Because of quantum mechanics, you would be able to write perfect public/private key crypto that is not interceptable. To the best of my understanding, quantum crypto has to do with sending photons with specific polarities across a pipe. It works because anybody who wants the information (photon) would have to bother the photon to get it's polarity. So getting the information messes the information and invalidates it(it would be coupled with message integrity checks and public/private key crypto)
As the linked article points out, the quantum methods mean you can guarantee the transmission is secure, but not a lot else. These cryptographic methods have all the security of other methods and then some. The only weakness (and I really mean only) is that the keys are still subject to theft if you aren't very careful.
Ever since I've been studying cryptography, poor Alice has been trying to talk to Bob without having that bitch Eve eavesdrop. Why can't Eve just let them be, for chrissakes?!? Then, as a side benefit, distributed.net would be able to redirect their efforts to something rather more worthwhile, such as looking for imaginary little green men.
On a side note, ever consider the possibility that Einstein was right all along and quantum magic really is bogus? If the linked-to people, currently disregarded by the scientific community as crackpots and throwbacks, end up proven right, that would be damned funny... "Hello? Yes, this is Mr Scientist Man, who is calling? Ah, the NSF? Yes, I know you've been giving us research grant money for the last 50 years to build huge particle accelerators and develop O(1) code-breaking for the NSA... you want to know why our prototype won't work? Well, it turns out that spooky action-at-a-distance is a measurement error, the Bell inequalities were never violated, and the universe is really fundamentally deterministic... sorry about that. See your money back? Not unless the NSF operates in the Bahamas too..."
It's like they say, nobody ever got fired for believing in Einstein...
One last thing... timothy, learn to close your italics.
To the editors: your English is as bad as your Perl. Please go back to grade school.