The History of One-Time Pads and the Origins of SIGABA
Blog post from Steve Bellovin:
It is vital that the keystream values (a) be truly random and (b) never be reused. The Soviets got that wrong in the 1940s; as a result, the U.S. Army’s Signal Intelligence Service was able to read their spies’ traffic in the Venona program. The randomness requirement means that the values cannot be generated by any algorithm; they really have to be random, and created by a physical process, not a mathematical one.
A consequence of these requirements is that the key stream must be as long as the data to be encrypted. If you want to encrypt a 1 megabyte file, you need 1 megabyte of key stream that you somehow have to share securely with the recipient. The recipient, in turn, has to store this data securely. Furthermore, both the sender and the recipient must ensure that they never, ever reuse the key stream. The net result is that, as I’ve often commented, “one-time pads are theoretically unbreakable, but practically very weak. By contrast, conventional ciphers are theoretically breakable, but practically strong.” They’re useful for things like communicating with high-value spies. The Moscow-Washington hotline used them, too. For ordinary computer usage, they’re not particularly practical.
I wrote about one-time pads, and their practical insecurity, in 2002:
What a one-time pad system does is take a difficult message security problem — that’s why you need encryption in the first place — and turn it into a just-as-difficult key distribution problem. It’s a “solution” that doesn’t scale well, doesn’t lend itself to mass-market distribution, is singularly ill-suited to computer networks, and just plain doesn’t work.
One-time pads may be theoretically secure, but they are not secure in a practical sense. They replace a cryptographic problem that we know a lot about solving — how to design secure algorithms — with an implementation problem we have very little hope of solving.