Network systems security by mort anvari lecture5

After DES… Network Systems Security Mort Anvari After DES…  More symmetric encryption algorithms   Triple-DES Advanced Encryption Standards 9/9/2004 Triple DES  Clearly a replacement for DES was needed   theoretical attacks that can break it demonstrated exhaustive key search attacks  Use multiple encryptions with DES implementations  Triple-DES is the chosen form 9/9/2004 Why Triple-DES?  Double-DES may suffer from meet-in-the-middle attack      works whenever use a cipher twice assume C = EK2[EK1[P]], so X = EK1[P] = DK2[C] attack by encrypting P with all keys and store then decrypt C with keys and match X value can show attack takes O(256) steps 9/9/2004 Triple-DES with Two Keys  Must use encryptions  would seem to need distinct keys  But can use keys with E-D-E sequence    encrypt & decrypt equivalent in security C = EK1[DK2[EK1[P]]] if K1=K2 then can work with single DES  Standardized in ANSI X9.17 & ISO8732  No current known practical attacks 9/9/2004 Triple-DES with Three Keys  Some proposed attacks on two-key Triple-DES, although none of them practical  Can use Triple-DES with Three-Keys to avoid even these  C = EK3[DK2[EK1[P]]]  Has been adopted by some Internet applications, e.g PGP, S/MIME 9/9/2004 Origins of Advanced Encryption Standard  Triple-DES is slow with small blocks  US NIST issued call for ciphers in 1997  15 candidates accepted in Jun 1998  were shortlisted in Aug 1999  Rijndael was selected as the AES in Oct 2000  Issued as FIPS PUB 197 standard in Nov 2001 9/9/2004 AES Requirements  Private key symmetric block cipher  128-bit data, 128/192/256-bit keys  Stronger and faster than Triple-DES  Active life of 20-30 years (+ archival use)  Provide full specification and design details  Both C and Java implementations  NIST has released all submissions and unclassified analyses 9/9/2004 AES Evaluation Criteria  Initial criteria    security – effort to practically cryptanalyze cost – computational algorithm & implementation characteristics  Final criteria     general security software & hardware implementation ease implementation attacks flexibility (in en/decrypt, keying, other factors) 9/9/2004 AES Shortlist  Shortlist in Aug 99 after testing and evaluation      MARS (IBM) - complex, fast, high security margin RC6 (USA) - very simple, very fast, low security margin Rijndael (Belgium) - clean, fast, good security margin Serpent (Euro) - slow, clean, very high security margin Twofish (USA) - complex, very fast, high security margin  Subject to further analysis and comment  Contrast between algorithms with   few complex rounds verses many simple rounds refined existing ciphers verses new proposals 9/9/2004 10 Polynomial GCD  Can find greatest common divisor for polys c(x) = GCD(a(x), b(x)) if c(x) is the poly of greatest degree which divides both a(x), b(x)  can adapt Euclid’s Algorithm to find it:  EUCLID[a(x), b(x)] A(x) = a(x); B(x) = b(x) if B(x) = return A(x) = gcd[a(x), b(x)] R(x) = A(x) mod B(x) A(x) ← B(x) B(x) ← R(x) goto  9/9/2004 30 Modular Polynomial Arithmetic  Can compute in field GF(2n)    polynomials with coefficients modulo whose degree is less than n hence must reduce modulo an irreducible poly of degree n (for multiplication only)  Form a finite field  Can always find an inverse  can extend Euclid’s Inverse algorithm to find 9/9/2004 31 Arithmetic in GF(23) 9/9/2004 32 Rijndael  Process data as groups of bytes (State)  Has 9/11/13 rounds in which state undergoes:     byte substitution (1 S-box used on every byte) shift rows (permute bytes between groups/columns) mix columns (subs using matrix multiply of groups) add round key (XOR state with key material)  Initial XOR key material & incomplete last round  All operations can be combined into XOR and table lookups, hence very fast and efficient 9/9/2004 33 Rijndael 9/9/2004 34 AES Round 9/9/2004 35 Byte Substitution  A simple substitution of each byte  Uses one table of 16x16 bytes containing a permutation of all 256 8-bit values  Each byte of state is replaced by byte in corresponding row (left bits) and column (right bits)   eg byte {95} is replaced by row col byte, which is {2A} S-box is constructed using a defined transformation of the values in GF(28) 9/9/2004 36 Shift Rows  Circular byte shift in each row     1st row is unchanged 2nd row does byte circular shift to left 3rd row does byte circular shift to left 4th row does byte circular shift to left  Decryption does shifts to right  Since state is processed by columns, this step permutes bytes between the columns 9/9/2004 37 Mix Columns  Each column is processed separately  Each byte is replaced by a value dependent on all bytes in the column  Effectively a matrix multiplication in GF(28) using prime poly m(x) =x8+x4+x3+x+1 9/9/2004 38 Add Round Key  XOR state with 128 bits of the round key  Again processed by column (though effectively a series of byte operations)  Inverse for decryption is identical since XOR is own inverse, just with correct round key  Designed to be as simple as possible 9/9/2004 39 AES Key Expansion  Take 128-bit (16-byte) key and expand into array of 44/52/60 32-bit words  Start by copying key into first words  Then loop creating words that depend on values in previous and places back   in of cases just XOR these together every 4th has S-box + rotate + XOR constant of previous before XOR together  Designed to resist known attacks 9/9/2004 40 AES Decryption  AES decryption is not identical to encryption since steps done in reverse  But can define an equivalent inverse cipher with steps as for encryption   but using inverses of each step with a different key schedule  Works since result is unchanged when   swap byte substitution & shift rows swap mix columns and add (tweaked) round key 9/9/2004 41 Implementation Aspects  Can efficiently implement on 8-bit CPU     byte substitution works on bytes using a table of 256 entries shift rows is simple byte shifting add round key works on byte XORs mix columns requires matrix multiply in GF(28) which works on byte values, can be simplified to use a table lookup 9/9/2004 42 Implementation Aspects  Can efficiently implement on 32-bit CPU     redefine steps to use 32-bit words can precompute tables of 256-words then each column in each round can be computed using table lookups + XORs at a cost of 16Kb to store tables  Designers believe this very efficient implementation was a key factor in its selection as the AES cipher 9/9/2004 43 Next Class  Confidentiality of symmetric encryption  Asymmetric encryption: RSA 9/9/2004 44 ... complex, fast, high security margin RC6 (USA) - very simple, very fast, low security margin Rijndael (Belgium) - clean, fast, good security margin Serpent (Euro) - slow, clean, very high security margin...  Initial criteria    security – effort to practically cryptanalyze cost – computational algorithm & implementation characteristics  Final criteria     general security software & hardware... Winner - Rijndael  Designed by Rijmen-Daemen in Belgium  Has 128/192/256 bit keys, 128 bit data  An iterative rather than feistel cipher   treats data in groups of bytes operates an entire block