Australian Company Claims Laser-Based Quantum Crypto is "Unbreakable" (Video)

The QuintessenceLabs website doesn't mince words when it comes to self-promotion. It boasts that they are "The world’s first company to harness the quantum properties of lasers to herald a new generation of data security." InvestCanberra says, "the defense and security policy and procurement centre of Australia is the natural location for large conglomerate defense and security corporations and specialist cyber security, advanced communications and radar, ICT and surveillance businesses alike," and goes on to list QuintessenceLabs as one of several "locally headquartered companies that have grown into internationally successful organizations."

Here's another statement taken from the company's website: "QuintessenceLabs is the first in the world to exploit a new generation of quantum cryptographic technology which enables unbreakable, secure storage and communication of sensitive information through the generation of an ultra-secure cryptographic key." Unbreakable? That's a strong boast. Is it true? And even if it's only partly true, your upper management may call on you to explain (and possibly implement) laser-based quantum security, so you need to know what it is and how it works -- and whether it's something your company (or your client companies) need.

Comments

So...?

[fahrbot-bot]

Laser-Based Quantum Crypto is "Unbreakable"

Sharks: 1
NSA: 0

Re:So...?

[lgw]

Well, "quantum crypto" should really be called "quantum key distribution". It's the key distribution part that's "unbreakable" - the rest is just AES or whatever. However, key management is the interesting part of cryptography for attackers: it's easier to somehow find the key than to attack the math.

And the quantum aspect doesn't actually prevent an attacker from snooping during key distribution, but it does provably let you discover that snooping has happened, and act accordingly, which is a valuable thing. (Tapping fiber-optic lines is really easy - not sure why people think it's hard: you just bend the line tight enough for some light to escape, or connect a tap with the same impedance as the line, or whatever.)

Re:So...?

[lgw]

Not it doesn't. Alice can't differentiate between Bob reading the message and an attacker performing a MITM attack.

If that were all it did, it would still be valuable, as it's harder to hide a device that can do that than it is to hide a simple tap. But that's not true - that's the "quantum" part.

Alice sends Bob a photon polarized along a basis randomly chosen by Alice. Bob measures the polarization along a randomly chosen basis. If the bases were the same, one bit was successfully sent, if not it's noise (polarization as measured "up-down" gives no information at all about what you would have measured "left-right", and you can only ever measure one of them, you can't possibly measure both.) After the message is sent, Alice tells Bob (non-secretly) what the basis for every photon was, and the discard all the bits where they didn't randomly line up. The bits where they did randomly agree are the message - the key in this case.

Eve, our eavesdropper, can measure the polarization along some basis, sure, but can't know what basis Alice chose, so the best Eve can do is emit a photon polarized the same say as measured with Eve's basis. If Eve receives all of Alice's photons, and sends photons to Bob, she can only get the transmission right half the time - the chance of guessing right (and thus being undetected) is the same as the chance of guessing the key in the first place.

Re:So...?

[Geoffrey.landis]

but it does provably let you discover that snooping has happened

Not it doesn't. Alice can't differentiate between Bob reading the message and an attacker performing a MITM attack

.This depends on your definition of "quantum crypto".
  End-to-end quantum cryptography allows you to know whether your communication has been intercepted. It IS theoretically immune to MITM. Simple quantum key generation or exchange however, may not be.

Right. Alice can differentiate between Bob reading the message and a Man in the Middle interception, because to intercept the message, the man in the middle has to read the quantum bit -- that is, in the photon case, measure the polarization-- and reading it destroys it. The Man in the Middle can't then send on a copy of the quantum bit, because of the quantum no-cloning theorem. She can send a photon polarized in a way identical to the way she read the photon-- but that only works if the receiver, Bob, happens to choose the same polarization to measure that the Man in the Middle chose to measure.

There is, however, a footnote to this. Eve actually can clone the photon (cloning photons is what a laser does). But she can't clone it perfectly-- or, more specifically, she can't be sure that the "cloned" photon is actually a copy of the original, or a spurious ("spontaneous emission") photon that happens to be in the same place. So, if she clones the photon, and measures one copy, and sends the other copy on, this shows up as noise in the signal-- exactly the same as stray light. So, Eve can read some of the message, if she's sufficiently clever, but how much is limited by how much noise the people communicating will accept in the signal without realizing that they are tapped.

feh

[Black Parrot]

"laser based" is irrelevant, except as a way to get the quantum effects.

And even if those effects are real, I'm guessing that 'quantum' is not able to provide stronger encryption, only to make it easier to *break* encryption.

Tell your boss to spend the money on a new yacht instead.

Re:feh

[geekoid]

Tell you boss to hire a quantum encryption consultant.

Don't mention the consultant happens to be your friend. Charge 225 an hour, split the difference.

By yourself a yacht.

Re:stupid claim

[RobotSpider]

Doesn't have to be unbreakable indefinitely. Only has to be unbreakable until the VC checks clear.

Old news

[Hentes]

This is not a new technology, and have been under lab testing for a while now. The problem is that what's theoretically unbreakable isn't that secure in practice. Turns out it's quite hard to distinguish between eavesdropping and noise.

"Partly true"? What?

[wonkey_monkey]

Unbreakable? That's a strong boast. Is it true? And even if it's only partly true...

...then it's false.

Re:*NOTHING* is unbreakable

[LordLimecat]

You cannot break the key in a properly implemented OTP. You have no way knowing which of the 8 zillion possible valid plaintexts was actually sent.

The weaknesses are only:
  * If the OTP repeats-- that is, the key is not the same length as the message. For an unbreakable 2KB OTP message, you need a 16000bit key (2KB).
  * If the OTP is generated deterministically-- it is not random.
  * Key distribution is vulnerable. No matter what method you use, unless it is face to face, the OTP can be "broken" by intercepting the key.
  * Key storage. If anyone captures your OTP booklet or file, you have no security whatsoever.

If you figure those out, its "perfectly" secure-- but as mentioned it basically requires face-to-face before hand OTP distribution and storing the OTP keys in a physical, airgapped vault.