Quantum Encryption

Posted on by Chris Warburton

I enjoy watching PBS Spacetime, and their recent video on quantum cryptography had some interesting stuff (especially the way BB84 and entangled particles are used for key sharing; the random choice of measurement basis was the “trick” that I hadn’t realised before).

However, some of the computer science was a little off though, especially some mixing up between symmetric and asymmetric encryption, e.g. talking about “sharing a private key” (“private” means never shared).

Symmetric encryption uses the same key to encrypt and decrypt, like a lock which needs the key to be engaged or released. This requires both parties to have the same key, hence it’s a “shared key” (or “shared secret”). The words “public” and “private” don’t apply here.

Asymmetric encryption uses different keys for encrypting and decrypting. The encrypting key is called “public” since it doesn’t need to be secret; it’s like a padlock that anyone can use to lock a package. The decrypting key is called “private” since we never need to share it with anyone, ever; it’s like the key which opens a padlocked package.

Even though private keys are never shared, asymmetric encryption is still vulnerable to a man-in-the-middle when the public keys are obtained.

As comments on the video have said, prime factoring isn’t particularly special; any “one way function” (easy to do but hard to undo) will work. A bunch of alternative methods have been invented which seem to withstand quantum computing.

Plus, I’d also point out that AFAIK there’s no proof that quantum computers are faster than classical computers at any task (the “quantum supremancy” problem). We know that prime factorisation is easy for quantum computers (using Shor’s algorithm); but we don’t know whether it’s actually hard for a classical computer, or if we just haven’t yet found the “trick” to making it easy. In other words, prime factorising might be broken in many ways; quantum computing is just the only way we know at the moment.

In fact, we might even discover a fast classical algorithm for simulating a quantum computer; in which case quantum and classical computers would be equivalent (although one may be more practical or efficient to build & operate).