Cryptanalysis
This domain tests how cryptosystems fail: attack models, brute force and keyspace, classical frequency analysis, meet-in-the-middle, hash attacks and implementation side channels. Questions ask you to identify the weakest point or the correct complexity.
Attack models
- Ciphertext-only: attacker has only ciphertext; the weakest position.
- Known-plaintext: attacker has matching plaintext/ciphertext pairs.
- Chosen-plaintext/ciphertext: attacker can obtain encryptions or decryptions; stronger.
- Related-key: attacker exploits relationships between keys.
- Side-channel: timing, power, cache or fault behaviour leaks secrets without breaking the mathematics.
Classical and algorithmic attacks
Frequency analysis exploits uneven letter distributions in monoalphabetic ciphers. Vigenère yields to Kasiski examination and index of coincidence. Differential cryptanalysis studies how input differences propagate; linear cryptanalysis exploits linear approximations. For 2DES, meet-in-the-middle reduces the naive 2^112 search to about 2^56 work and memory, which is why 2DES was never adopted.
Hash and password attacks
A brute-force preimage attack on an n-bit hash costs about 2^n; a collision attack costs about 2^(n/2) by the birthday bound. Rainbow tables are precomputed chains defeated by unique salts. Password cracking is usually dictionary or rule-based before brute force. HMACs resist length-extension attacks that plain hash constructions can suffer.
Worked example
Suppose a system uses two-key 2DES: C=E_K2(E_K1(P)). A naive attacker tries all 2^112 key pairs. Instead, encrypt P under every possible K1 (2^56 operations) and store the results. Then decrypt C under every possible K2 and look for a match in the table. The match reveals K1 and K2. The attack needs about 2^56 time and 2^56 memory, so doubling DES did not double the effective security.
Traps
- Saying brute force always takes the full keyspace; on average it succeeds halfway through.
- Confusing collision resistance (2^(n/2)) with preimage resistance (2^n).
- Claiming a salt prevents brute force; it prevents precomputation and hides identical passwords.
- Treating a side-channel attack as a mathematical break of the cipher.
- Forgetting that chosen-plaintext is stronger than known-plaintext.