Optical boffins cut the cost of quantum cryptography

Woosh...

sound of this level of sci/tech going over the head.

To quote Ian Dury "There ain't half been some clever b**tards!"

But will it be able to cope with a photon torpedo, that's the real question?

Re: Something is bothering me...

It is because you can only make one observation, and hence only determine a single component of the photon's attributes; thus, replicating the photon is impossible. Check out e.g. http://www.cs.dartmouth.edu/~jford/crypto.html

i know nothing about this

but I think the extra time delay this causes alerts the receiver to the interception

Denial of service

It seems to me that an enemy, or someone just wanting to disrupt the key generation process only has to 'continually observe' the key set-up stream and the two parties will never be able to establish secure communications.

But...

if you change the photon (rotating it 45 degrees) without ruining the encryption then shirley the gook(s) can do the same and quietly intercept said photons.

To Steven Hunter

You're pretty much on the money there. There have been successful lab-based attacks of this sort. Essentially, they can be mitigated by properly designed authentication between the two parties, though- veryifying that Alice really is talking to Bob, rather than Eve*. Care over this sort of mechanism is the mainstay of traditional properly-designed secure protocols, but in some cases was initially missed in discipline gap between physics and crypto/information theory.

There's a really clever way around it using the quantum channel and some basis flipping of randomly generated numbers to generate a key, but it's a bit technical for a friday afternoon, if you have a tiny brayne like wot I have.

However, be aware that a debate has taken place about man-in-the-middle attacks, and solutions are out there. Is it time to go home yet?

- Ancipital

*Traditionally, when talking about such things, Alice and Bob are the parties communicating, Eve is the name given to the EVUL N SKERRY attacker.

Just don't look at it funny

@Steven:

Wikipedia to the rescue:

Quantum cryptography > Quantum Key Exchange > Polarized Photons

http://en.wikipedia.org/wiki/Quantum_cryptography#Polarized_photons_-_Charles_H._Bennett_and_Gilles_Brassard_.281984.29

What happens

...when there are only six people on the planet who know how things work - and they decide not to tell?

Can someone help the ignorant?

This stuff goes way over my head, and I'm not afraid to admit it. But from my ignorant understanding from this article, one device sends these photons to a second device. If a third device intercepts these photons, they are effectively "trashed", and the receiving device knows something is amiss.

It's probably just because I'm ignorant with regard to this field, but what are the practical uses for this technology? If a man-in-the-middle device cannot exactly replicate the photons, then the two communicating devices must be physically linked (put another way, a switch or router would not be able to replicate the photons, either). I have no doubt this technology can be used to secure keys for important things (perhaps securing/validating the controllers in a nuclear power plant?), but I can't think of many uses where two devices would have a physical connection to each other. Can someone point out some examples that I'm overlooking?

Basics as I understand

Quantum cryptography is basically useful for security across a network. Almost all current securities work on the integer factorisation problem. That is, it's hard to find which 2 prime numbers multiplied together to make a number that gets transferred across the internet. Say for example, the number 15... well we know its made up by 3 and 5, with a 200 digit number though, that's another matter.

However, with the advent of quantum computers, they can break this problem in a polynomial time rather than exponential. Basically, rather than say 2^N it solves it in N^2. If N was 1024, thats 308 digits compared to only 7 for the latter. So much much faster, basically rendering current key exchange protocols useless.

So how to protect the key that gets used to encrypt bank details going across the internet after quantum computers are around?

To give you an idea of one way I like to think about it: Imagine binary and decimal.

If I told you "One zero" (10), that could be 10 in decimal or 2 in binary, depending on the basis that you read it in. You have no way of knowing which one it is. If you tried to tell someone else about it in either binary or decimal, you only have a 50/50 chance of getting the right one. Using this, you can effectively exchange a key without actually telling anyone about it.

(Note: 10 and 11 make nice comparisons for the binary/decimal analogy, but they just use 0 and 1 really. 0 and 1 mean the same in binary/decimal so I used 10/11 instead). So If someone sent 11 in binary then 10 in decimal then 10 in binary. The other guy would guess they are using decimal then decimal then binary, he reads the data to see if they are either 11 or 10. He then asks the sender about each one weather he was right or not. He finds out which are wrong and he will ignore that one. So the secret key would be 10 10. Using this, you shared information and the people in between can only guess.

Now if you did this with 20 parts the person guessing it in the middle would have a million to 1 chance of getting it right. So rather than just factorising a number, which will eventually be a simple math problem... you will have to play a lottery. If you had 72 parts to it, you would only have a 4277 million MILLION to one chance to guess it right; basically making it impossible to find the key.

This article is just about (Per my analogy) rather than needing a detector for binary and decimal, you divert one of them on a longer path and change it to the other basis. Now you know the one that arrives later is the one that got changed, despite being both in the same basis.

It's good to see some advancements in the field.

P.s. I'm doing a final year module at university in quantum computing. It's very hard for me to wrap my head around a lot of it as well. I have over £100 worth of books sitting here about it to read as well =/ (I'm lazy).Quantum entanglement is a whole new level of screwing with your brain. If ever you could call something "magic", this would be it.