One-time pad

The one-time pad is the only known encryption system that offers perfect secrecy — information-theoretic security that cannot be broken by any adversary, regardless of computational power. Proven by Claude Shannon in 1949, its requirement is simple and brutal: the key must be truly random, as long as the message, never reused, and kept secret. These conditions make the one-time pad impractical for most real-world communication, but its existence establishes an absolute benchmark against which all other ciphers are measured.

The security of the one-time pad rests on the key's entropy, not on the algorithm's complexity. Because every possible plaintext is equally consistent with any given ciphertext, cryptanalysis is impossible in principle. The system does not resist attack; it makes attack meaningless. This is not a matter of degree — it is a categorical difference between the one-time pad and every computationally secure cipher.

The practical failures of one-time pad systems have historically stemmed from key reuse, not algorithmic weakness. When the same key encrypts two messages, the XOR of the two ciphertexts reveals the XOR of the plaintexts, often enabling recovery of both messages. The Venona project decrypted Soviet communications precisely because their one-time pad keys were reused under wartime pressure. This failure mode is structural: the system's security collapses entirely when its key management assumptions are violated.

The one-time pad's impracticality has led some to dismiss it as a theoretical curiosity. This is a mistake. The one-time pad is not an obsolete technology; it is a proof that the boundary between possible and impossible security is known and exact. Every claim that a practical cipher "approaches one-time pad security" is a category error. Computational security is not an approximation of perfect secrecy; it is a different kind of guarantee altogether, and conflating the two has produced dangerous overconfidence in deployed systems.