OSDev.org

I was reading this article, and was wondering if anyone thought quantum cryptography is unbreakable.

200 years ago many types of ciphers were thought to be unbreakable, and those were broken. Moreover, scientists have yet to even fully understand the quantum world. How then, can one say that quantum cryptography is perfectly secure? If anything, doesn't science show that nothing is perfect?

(Note: By perfectly unbreakable, I mean unbreakable forever by means of some natural law, not unbreakable now due to technical inefficiencies.)

It would probably take me several months (if not longer) to research into the subject to be able to understand the subject well enough to start making any sort of educated guesses as to whether it would indeed be breakable or not, and it's not in my field of interest, so personally I'd just trust the general consensus of the field in question on this one.

On the other hand, it would be fun to know what sort of qualifiers (on unbreakable) were removed by the editor of that article or the marketing division of the company when preparing the announcement for the general public.

No. In my opinion, it's not unbreakable. Yes, the article says it is, but I don't think so.....

Eventually a Quantum machine will be able to break it.

Over the years, I read quite some papers on the subject, due to a general interest in physics and quantum mechanics. While by no means competent in the area, I think I can give a quick summary. (Not sure if this technique is what the article is about as I have only skimemd it, but all approaches on QC I've read about so far worked along the same lines.) I will not really explain the individual points, but merely point out how they work together.

I trust you know the concept behind Schrodinger's cat - that at the quantum level there can be metastable states where a particle is neither this or that (binary speaking, 0 or 1), but in an "uncertainty cloud" covering both possibilities. The actual orientation is decided only at the moment it is measured.

It is possible to generate multiple particles, which will assume the same orientation when one of them is measured - regardless of the distance that might be between them. (One of the really hard-to-understand concepts of quantum physics, as the "information" of these particle's orientation is "travelling" instantaneously, i.e. faster than light. It does work, though.)

The idea, now, is for Alice to generate two metastable particles linked in this way, and send one to Bob. Bob measures his particle, which assumes an orientation (0 or 1). Bob then sends to Alice that he did receive a reading (not what that reading was). Alice reads her particle (which has the same orientation as Bob's), and they have one bit of truly random, truly private key space.

Mallory could not intercept the particle, as Bob wouldn't recieve it then and wouldn't acknowledge he had received one. Mallory also couldn't copy the particle, as copying a single metastable particle is impossible.

Now, theoretically, Alice and Bob could repeat this process until they have one bit of key space for every bit of message - a one-time pad, which is proven to be unbreakable without the key.

That's the theory of it, essentially. The problems are in the technical implementation. It is awfully hard to select one particle and send it over a distance. (If you send two, Mallory could intercept one, and Bob still receive the other.) You cannot do this by radio, you need line-of-sight or a glass fiber connection. You need pretty sophisticated equipment on both sides of the communication.

I'd say, yes, theoretically quantum cryptography is unbreakable. But it is also impractical in most everyday uses. You won't see it employed e.g. for home banking, private e-mail, or squad-level military use.

Hi,

Solar wrote:The idea, now, is for Alice to generate two metastable particles linked in this way, and send one to Bob. Bob measures his particle, which assumes an orientation (0 or 1). Bob then sends to Alice that he did receive a reading (not what that reading was). Alice reads her particle (which has the same orientation as Bob's), and they have one bit of truly random, truly private key space.

Mallory could not intercept the particle, as Bob wouldn't recieve it then and wouldn't acknowledge he had received one. Mallory also couldn't copy the particle, as copying a single metastable particle is impossible.

Mallory could intercept Alice's particle, and generate 2 linked particles of her own (and send one to Bob). Then Mallory can intercept information from Alice, decode it with Alice's particle and then re-encode it with her own particle before re-sending the information to Bob.

Note: I know everything that can be known about quantum physics, but only if I don't need to remember it....


Cheers,

Brendan

I think every form of cryptography is breakable. Even if it does take more than the entire life span of the universe to break it. If an algorithm exists to decode a message, then it is possible, even if only by perseverance over 100,000,000 years with a bit luck added, that someone else will be able to write an identical algorithm to break it.

As we get further into the future, we will have access to faster computers, and more mature AI systems that would judge human behaviour to find patterns in the sorts of algorithms people create. Artificial Creativity would aso play a major part by generating the initial random algorithms.

Brendan wrote:Mallory could intercept Alice's particle, and generate 2 linked particles of her own (and send one to Bob). Then Mallory can intercept information from Alice, decode it with Alice's particle and then re-encode it with her own particle before re-sending the information to Bob.

Nice. That's where the protocol steps in - how does Alice know the acknowledgement came from Bob, not Mallory?

I know they had it covered somehow, but that's the point where I would have to start researching. (Can't be bothered to do so ATM. )

@ MessiahAndrw:

A message encoded with a random one-time pad is undecodable without the key, even with a machine with unlimited computing power. You simply have no way of knowing when you've hit the correct plaintext and when you've simply hit another plausible sequence of letters of the same length.

(See Wikipedia ]on the security of one-time pads.)

The only problem with the one-time pad is that you need a key that is of the same length as the message, completely random, and available only to Alice and Bob. Quantum cryptography aims at providing a means for this.