Theoretical breakthrough for quantum cryptography | Hacker News Reader

Theoretical breakthrough for quantum cryptography

technologyreview.com

8 pointssev7 comments

This has become my go-to response to QC-related issues. All you need to know is in the title:

http://rdist.root.org/2008/10/24/quantum-cryptography-is-use...

So, my understanding was that the questions of quantum cryptography stem from the fact that cryptographic algorithms designed to stump current computers will be a breeze for quantum computers, so new algorithms will need to be created to stump quantum computers. I thought that's what was referred to as "quantum encryption". Is there another name for what I'm thinking of?

But now they're saying that this "quantum encryption" they speak of will be "perfectly" secure? Does this mean that, once established, cryptography will no longer need to be a field of study?

Also this: "This solves the problem by embedding it in an extra abstract dimension, which is independent of space. So as long as both Alice and Bob know the way in which all these abstract dimensions are related, the third provides a reference against which measurements of the other two can be made." blows my mind. The fact that this "other dimension" stuff is no longer just speculation, but will soon be part of our lives.

It's two different things:

- one, quantum computers can theoretically do factoring in polynomial time and therefore theoretically break schemes such as RSA (such schemes are proven secure provided that some computational problem is intractable -- in the quantum world the problem becomes tractable)

- two, you can theoretically use entanglement to achieve "perfect cryptography". The "perfectness" result is an absolute impossibility result in physics, and therefore no computer (classic or quantum) can break it. Details on wikipedia: http://en.wikipedia.org/wiki/Quantum_cryptography

The article refers to number two, although in a rather non-technical manner.

Edit: clarifications.

I would not call it perfect: you still need a classical message authentication code to protect against a man in the middle attack.

"entangled triplets of photons, so-called qutrits"

Actually, a qutrit is just a single tri-state qubit. The term describes superposition, not entanglement.

article implies that QC is being used in limited practice right now. i kinda doubt this. for example, this IARPA solicitation implies that they're doing baby steps with very small data. of course, if they had it practically working, it would be classified.. http://www.iarpa.gov/solicitations_mqco.html

QC is used for key exchange. Private-key ciphers can be more secure than public-key ciphers, the problem is how to distribute keys securely. You just need to send about 1k of data (the key), and you only need to send it once. After that, the two parties use the key to encrypt their messages and send them using classical channels.

Reply on news.ycombinator.com