Slashdot Mirror
BBN Announces Functional Quantum Encrypted Network
anzha writes "BBN Technologies has announced that under DARPA's Quantum Network Project to have built in conjunction with Harvard University the world's first functional quantum encrypted network. This is probably funded under DARPA's Quantum Information Science and Technology Program."
Harvard should link to http://www.harvard.edu/ or http://www.fas.harvard.edu/ if you want the Faculty of Arts and Sciences (the undergraduate institution that everybody loves) not "hardvard.edu".
While technologically do-able, I'm uncertain as to whether this will succeed in the commercial world.
I think we'll all just have to wait and see.
Visit CryptoGnome in his home.
Check the Wikipedia my dear.
Or alternatively, see this (goatse-free) image.
when has cryptography ever stopped someone reading a message? Do you meant that it is unbreakable?
I spent ages trying to think of sig, but never did
The traffic on conventional networks can be passively or actively observed and recorded, often without the communicating parties noticing. This apparently cannot happen with quantum communication because if a photon is observed in transit, its state changes and hence the eavesdropping is obvious to the communicating parties.
"Quis custodiet ipsos custodes?"
This scheme might be subject to denial-of-service attacks by eavesdroppers, but I'm sure they've thought of that in their network design. Probably they can send the keys via alternate routes in case of interruption of a link.
Have you read my blog lately?
Not all the laws, just P != NP.
None, except for one-time pads (which have other problems). For example, many schemes depend on the fact that it's impractical to factor large numbers. The truth is that nobody knows how to factor large numbers today, but it's also true that nobody knows how hard this problem really is. Perhaps someone clever will figure out how to do it tomorrow... and in the meanwhile, someone already has figured out how to factor large numbers using quantum computing. Nobody has built a quantum computer large enough to run the algorithm, but once they do, you can kiss all these schemes goodbye.
Am I part of the core demographic for Swedish Fish?
But RSA didn't take off until the patents expired a couple of years ago!
Yes, just like RSA, and Diffie-Hellman key exchange, SHA-1...
None of those are in patent. RSA was patented but that patent expired a few years back. SHA-1 was never patented nor was Diffie-Helman.
we don't have secure communications, what we have are communications that nobody knows how to break yet
Well, not exactly we have the One time pad but that aside: What makes physics different to mathematics? You can't prove a physical theory is true like you can a theorem. There is a small chance quantum mechanics is wrong and there is an alternate theory that describes the photons in a deterministic way.
Yes, it's a small chance.. but don't forget that there's also a small chance that you can find a quick algorithm to solve AES. Changing the laws of physics tends to happen once a century - Theorems on the other hand last forever.
Simon.
No, you still have the problem of working out how to exchange the public key and know that it has not been interfered or tampered with during transit. Quantum Encryption can be more correctly thought of as Quantum Key Exchange. It provides a means of transfering keys together with the knowledge of whether that key has been intercepted in transit or mot.
Answers to lots of your questions at quantum.bbn.com, which is the actual document repository used by the development team. I think it's pretty cool that they make so much material publically available. There's also an overview linked from the BBN homepage.
Details, details, details.
Quantum cryptography, at least in this application, only uses the quantum network to exchange KEYS to conventional symmetric crypto.
The same crypto algorithms are used, this is just a "secure" method of key exchange. PKI was invented because of the problem of exchanging keys securely -- this is just a fancy way of doing the same thing.
Learning HOW to think is more important than learning WHAT to think.
Diffie-Hellman -- US patent 4200770
The SHA-1 algorithm is not patented, but many uses of the algorithm are.
Do your reading.
Am I part of the core demographic for Swedish Fish?
The point of quantum cryptography is not to make the crypto unbreakable, but to make attempts to eavesdrop on it detectable.
The network consists of fibre optic cables over which SINGLE PHOTONS are transmitted back and forth between "Alice" and "Bob". If anyone is trying to spy on you -- poof, your bits disappear, and you notice.
The actual crypto that's used on the network is fairly normal. The quantum part protects the key exchange.
"if you simply pass along exactly what you found"
In order to find it you must look at it. In order to look at a photon you must make it collide with the back of your eye thus destroying the photon. No two people can look at the same photon.
Likewise, no two machines can look at the same photon because in order for a machine to look at a photon it must measure it. In order to measure it, the photon must collide with a sensor which destroys the photon.
So sure, a man in the middle can intercept a photon but will not understand it's meaning and destroy it in the process preventing it from reaching it's original destination.
P != NP is not a law; it is only suspected to be true and remains an open problem, not yet proven.
...as observing the data in transit actually changes it
Quantum mechanical systems, unlike classical systems, can exist in a superposition of states. A classical bit for example, can only be either 0 or 1, while a quantum bit, or qubit, can exist as both 0 and 1 at the same time with some probability. Hence, when you 'observe' a quantum system, the system is forced to be (I won't use the word collapse here!) in a new state consistent with the apparatus or observable you used to observe it. That's an oversimplified explanation. Go to the tutorials section at the Cambridge Quantum Computing website for more tutorials and simple reading on how this stuff works, including some very cool articles by Artur Ekert, who independantly discovered quantum crypto
To answer several questions at once, the short answer about how it works is a consequence of the uncertainty principle: when you observe a photon (or any particle, for that matter), you have to interact with it in some way. When you do that, you change some of its properties.
"Observing the entangled photon(s) would not change the originals..."
Not exactly true. Look into the EPR experiment and what's known as "spooky action." It turns out acting on one entangled photon instantaneously (faster than light) affects its partner. For what you're saying, though, this doesn't really matter, as no information can be transmitted this way (luckily). However, entangling photons requires letting them interact, which will disrupt the original.
For instance, if you measure the polarization of a photon, which was previously in a superposition of polarization states, in some sense you have created the new polarization of the photon, you have made it be what you measured it to be. So if I send you a diagonally polarized photon, and you measure it straight up and down, after it passes through your measuring device it will be purely straight up or straight down, whichever you measured it to be. So if somebody taps the line, we will be able to tell, because they will change the polarization of the photons I send you and you will get gibberish.
This is of course a bit simplistic, but that's the heart of the matter. Objectivity is dead. You are part of the system. If you observe it, you will inevitably have an effect on it. It's kind of cool.
The neat thing about this is that, assuming QM is correct, there is no way to circumvent it with new technology or more powerful computers or anything else. No matter how cool your tech is, you can't observe a system without changing it.
My site: Free Nature Pictures
Please try to get headlines right.
This is not quantum encryption. Photon entanglement simply allows the recipient to detect if someone was listening. It's much like a signature, only stronger (signatures only go bad if someone tries to modify the data; quantum state of entangled photons changes if anyone even looks at the data).
You don't want to send critical information over such a link. You use that link to send a symmetric encryption key. Then you use crypto.
Eve, a passive MITM (WITM), can prevent you from ever using crypto by keeping the link tapped. You keep sending crypto keys across, but each time you realize they've been compromised. You cannot get anywhere in that situation unless you use public key crypto, at which point the quantum-entangled nature of the link gets you no extra security.