Skip to main content
Version: Next

Cryptanalysis

  • Process of decryption of ciphers and encrypted text
  • Identifies vulnerabilities in cryptosystems

Cryptanalytic techniques​

Linear cryptanalysis​

  • Known as plaintext attack
  • Applicable to block ciphers and stream ciphers.
  • Given enough pairs of plaintext and corresponding ciphertext, key can be obtained
  • Discovered by By Matsui and Yamagishi in 1992
  • Attacker identifies the linear relation between some bits of the plaintext, some bits of the ciphertext and some bits of the unknown key.

Differential cryptanalysis​

  • Discovered by Israeli researchers Eli Biham and Adi Shamir in the late 1980s.
  • Applicable primarily to block ciphers, but also to stream ciphers and cryptographic hash functions.
  • Applicable to symmetric key algorithms
  • Comparing differences in the inputs to how each one affects the outcome
  • Working with chosen plaintext originally, also works with known plaintext and ciphertext

Integral cryptanalysis​

  • Used on block ciphers
  • Discovered by Lars Knudsen in 1997
  • Input vs Output comparison same as differential, however, runs multiple computations of the same block size input
  • Attacker analyzes outputs of encrypting sets of plaintexts where some of the content is held constant and some of the content is varied through all possibilities

Code-breaking methodologies​

  • Frequency analysis
    • Study of the frequency of letters or groups of letters in a ciphertext
    • E.g. checking cipher chunks against in languages some letters or combination of letters are used more often
    • Can be used to crack a substitution cipher, like rotation cipher ROT13
  • Trickery and deceit
    • Requires a high level of mathematical and cryptographic skills
    • Using social engineering techniques to trick someone to encrypt and send a known message
  • One-time pad
    • A shared random key that has to be the same length or longer than the cipher text
    • Each individual bit or character of plaintext is encrypted by combining it with the corresponding bit or character from the pad using modular addition
    • Assuming to be unbreakable
    • Drawback
      • Key distribution becomes impracticable for large messages as key length is same as as the messages

Cryptography attacks​

Chosen-key Attack​

  • 📝 Attacker knows keys that are used or can choose the secret key.
  • May allow breaking the larger system which relies on that cipher

Rubber-hose attack​

  • Also known as rubber hose or rubberhose attack.
  • 📝 Extraction of cryptographic secrets (e.g. the password to an encrypted file) from a person by coercion or torture
  • E.g. beating that person with a rubber hose until they give up the encryption key.

Ciphertext-only attack (COA)​

  • Also known as known ciphertext attack
  • 📝 Attacker has only access to cipher texts
  • E.g. using frequency analysis to assume plain text
  • Early ciphers (using pen-and-paper) were cracked this way
  • Modern ciphers have strong protections against it
    • take years to separate statistical departure from random noise

Known-plaintext attack (KPA)​

  • Also known as known-plain-text attack
  • 📝 Attacker has access to parts of plaintext and corresponding ciphertext.
  • Can be used to reveal secret keys, code books.
  • Classical ciphers are typically vulnerable

Meet-in-the-middle attack​

  • Also known as meet in the middle attack.
  • Attack over certain block ciphers by decomposing the problem in two halves and proceeds on each part separately
  • Reduces the effort to perform a brute-force attack
  • 📝 Reason why re-encrypting an ciphertext reduces its security
    • The reason that Triple DES or Double DES is considered weak and are no longer used.
  • E.g. transforming an attack that requires 2exp128 time into one that takes 2exp64 time and 2exp64 space
  • Type of known-plaintext attack

Chosen-plaintext attack (CPA)​

  • 📝 Attacker can choose random plaintexts to be encrypted and obtain the corresponding ciphertexts
  • 📝 Two forms
    • Batch chosen-plaintext attack
      • Cryptanalyst chooses all plaintexts before any of them are encrypted.
      • Not so effective
    • Adaptive chosen-plaintext attack
      • Cryptanalyst makes a series of interactive queries
      • Subsequent plaintexts are chosen based on the information from the previous encryptions.

Chosen-ciphertext attack (CCA)​

  • Also known as chosen ciphertext attack or chosen-cipher-text attack.
  • Attacker gathers information by obtaining the decryptions of chosen ciphertexts.
  • Early versions of RSA padding used in the SSL protocol were vulnerable.
  • Types
    • Adaptive chosen-ciphertext (CCA2)
      • Attacker uses the results from prior decryptions to inform their choices of which ciphertexts to have decrypted
    • Non-adaptive chosen-ciphertext
      • Attacker chooses the ciphertexts to have decrypted without seeing any of the resulting plaintexts
    • Lunchtime attack or midnight attack
      • Attacker can have access to system for only a limited amount of time, can access only few plaintext-ciphertext pairs

Side-channel attacks​

Timing attack​

  • Execution times are measured to learn more about the system
  • Information to find can include e.g. key, CPU used, algorithms, input, implementation details etc.
  • A type of side-channel attack

Brute-force attack​

  • Also known as brute force
  • Trying every possible combination of characters to break the encryption
  • 📝 ❗️ Requires a lot of time and processing power.
  • See also Brute-force attack | Cracking passwords

Birthday Attack​

  • Type of brute-force attack but faster that focuses on collisions
  • Based on collisions where attacker uses own plain texts to match hashes (find collisions)
  • Depends on the higher likelihood of collisions found between random attack attempts and a fixed degree of permutations
  • Exploits birthday problem in probability theory
    • E.g. 23 people in room, chance of two having same birthday is not 23 / 365 = ≈6% but it's 50%. For 75 people it's 99% chance.

Rainbow table attack​

  • 📝 Rainbow table contains precomputed hashes to try and find out passwords
  • Faster than brute-force however the trade-off is that it takes a lot of storage (sometimes Terabytes) to hold the Rainbow Tables themselves.
  • Tools

Dictionary attack​

  • Attacker creates and uses a dictionary of plaintext and its ciphertext.
  • E.g. words in a dictionary
  • E.g. previously used passwords, often from lists obtained from past security breaches
  • Attacker observes the operation of a cipher under several different keys
  • Some relationship connecting the keys is known to attacker while key values are unknown.
  • E.g. attacker knows that last 80 bits of the keys are the same

DUHK Attack (Don't Use Hard-Coded Keys)​

  • Allowing attackers to access keys and read communications in certain VPN implementations
  • Based on vulnerability affecting devices using ANSI X9.31 Random Number Generator (RNG) with a hard-coded seed key

Collision attack​

  • Also known as hash collision attack
  • 📝 Tries to find two inputs resulting in same hash value, i.e. a hash collision.
    • Find two different messages m1 and m2 such that hash(m1) = hash(m2).
  • Extended by chosen-prefix collision attack
    • Given two different prefixes p1, p2
      • The attack finds two appendages m1 and m2 such that hash(p1 ∥ m1) = hash(p2 ∥ m2)
        • where ∥ is the concatenation operation.
    • More powerful
  • 💡 The larger the hash value size, the less likely there are for collisions to occur and therefore the more collision resistant the hash algorithm

Cryptography attack tools​

  • L0phtcrack
    • Password cracking tool
    • Used mainly against Windows SAM files
  • 📝 John the Ripper
  • CrypTool
    • Open-source program for for cryptography and cryptanalysis
    • GUI to experiment with cryptographic procedures and to animate their cascades
  • Cryptobench
    • Encrypt, decrypt, hash using many algorithms
    • Helps in the cryptanalysis process of common cryptographic schemes