IT lexicon › Security › Square attack

Square attack

Security På svenska → Updated: 2026-10-13

An integral attack on byte-oriented block ciphers, named after the Square cipher it was designed for.

The attack targets Square, a block cipher presented by Joan Daemen, Lars Knudsen and Vincent Rijmen in 1997. Square works on bytes rather than bits. The trick is to feed in 256 plaintexts that differ in only one byte. After a few rounds, the XOR of certain intermediate values is zero, and that property can be exploited to guess parts of the key.

The attack mattered because it applied directly to Rijndael, the candidate that became AES. It only broke versions with a reduced number of rounds, though. AES has ten to fourteen rounds depending on key length, far beyond what the attack reached. A cipher being "broken" usually means a simplified version of it has been broken, and the distance between the two can be uncomfortably small.

← Back to the lexicon