## Quantum Cryptography a reality?

BBC News recently reported that the world’s first robust network based on Quantum cryptography hasgone live in Vienna. Comprising 7 locations and 200km of optical fiber, and hosted by Siemens, the network has the capability to re-route connections in the event of link failure, and handle eavesdropping attacks.

Though it sounds a little Star Trek, Quantum Crypto has been around for over 20 years already – IBM patented the classic Quantum key exchange algorithm in 1984. It’s a novel concept because the key exchange algorithm (BB84) is provably secure – i.e., there’s no hack or attack for it unless our understanding of the nature of the Universe is fundamentally wrong.

Basically, the idea is based around measurements of the state of a single photon. Much like the fabled Heisenberg Uncertainty Compensators used in teleportation, the process relies on the fact that you can’t measure a photon without possibly changing it – yes, I know this makes no sense, measure a pound of sugar, and you still have the same sugar yes? Well, just believe me that if you measure a photon, say its polarization, it’s might not be the same afterwards – measuring the polarization will change it, and you have no idea whether the result will be the same as before, or different.

There’s a full write-up of how it works on Wikipedia, but let me see if I can summarize for you:

The fundamental exchange of information between the classic “Alice” and “Bob” characters, without anyone (“Eve” in this case) eavesdropping on the conversation. Simply, Alice sends some photons to Bob of her choice. For each one she selects a photon with a particular polarization and bias (rectilinear or diagonal) – yes, you can select photons with particular characteristics without “measuring” them. So, these go off to Bob.

Bob now measures these photons – for each one he has to pick a bias to measure it in. Bob doesn’t know the bias that Alice intended for the photon so he picks them at random. He measures the polarization of these photons as they come in.

Now comes the interesting part – if Bob gets the bias wrong, if he uses the wrong one compared to Alice, then the polarization result he derives will be random, it won’t match what Alice intended. This is where the measuring-changes-things comes in. However, IF he gets the bias right, he’ll get the right polarization result. Bob can’t determine the bias in advance.

So, now Bob has a list of polarizations for the photons Alice sent, with the bias he used to measure each one. Each time he picked the right bias, he’s got the polarization Alice intended, where he got it wrong, his result will 50% be the wrong polarization.

So, we still don’t seem to know much, but now comes the interesting part – Bob and Alice can just pick up the phone and discuss the bias of each photon. The transmission has been sent so there’s no risk. Bob simply reads out his choices to Alice, and she tells him where he was right or wrong. Bob disregards the measurements where he used the wrong bias, and the ones left (where he luckily picked the right one) becomes the key used to exchange information.

The best thing, is that if Eve listens in, she will be changing the polarization of the photons as they pass through her random bias tests randomly – so, when Bob and Alice have finished their discussion on bias, if they compare a VERY FEW polarization results, they will soon know if anyone was trying to snoop – if any of their results differ they know that someone else has been listening in.

So, that’s the basis of Quantum Cryptography. It’s intended use is as a key exchange mechanism between two parties and as you can see, it’s fundamentally robust. One thing it doesn’t handle though is the classic “man-in-the-middle” attack, but, that’s a topic for another blog.

Simon.

CTO, McAfee Data Protection

PS – for those who managed to read this far, here’s an example thanks to Wikipedia again.

Alice’s Random Bits 0 1 1 0 1 0 0 1

Alice’s random Bias + + x + x x x +

Alice’s Polarization | - \ | \ / / -

Eve’s random measuring Bias + x + + x + x +

Eve’s measured polarization | / - | \ - / -

Bob’s random measuring bias + x x x + x + +

Bob’s measured polarization | / / \ - / | -

Bob and Alice now discuss the correct Bias and disregard the bits where they differ. This gives them the shared secret key.

Shared secret key 0 x 0 x x 0 x 1

And finally they can discuss a few random bits to see if anyone was listening in – to get a “nine nines” probability that no one is listening they need to compare 72 bits (and of course disregard them from the final key).

Alice Original Key: 0 1 0 1

So, we can tell that Eve was listening in. Obviously Eve could have been lucky in our case and got the bias right for the third bit, but, if there are 72 bits or more, the probability of getting them all right is negligible.

## Comments