IT lexicon › Security › Known-plaintext attack

Known-plaintext attack

Security På svenska → Updated: 2026-10-11

The attacker holds both plaintext and ciphertext for some messages and uses them to work out the key.

It sounds like an unreasonable advantage, but it is routine: letters open with greetings, file formats with fixed headers, protocols with predictable fields. At Bletchley Park such guessed fragments were called cribs. German weather reports nearly always contained Wetterbericht, and the bombes against Enigma depended entirely on placing a crib correctly. The RAF occasionally laid mines purely to provoke messages whose content was known in advance.

Modern ciphers are expected to withstand even vast amounts of known plaintext. Older designs do not: the Hill cipher falls to a handful of blocks, and in 1994 Eli Biham and Paul Kocher showed that PKZIP's legacy encryption breaks with around a dozen known bytes. Linear cryptanalysis of DES is also a known-plaintext attack, albeit one needing 243 pairs.

← Back to the lexicon