No content

  64 bit data block   64 bit key of which 56 bits are used.   Uses 16 rounds   A round consists of   Reordering bits   XORing with the key   Run the Feistel function   Swapping the data block halves 

No content

  Methods   DES::algorithm( .. )   DES::keyschedule( .. )   DES::mixer( .. )   DES::f(..)   Implemented in C++   --std=c99   Uses OOP Design 

  Schedule’s all 16 key’s   Initial Permutation   For each round   Choose key for round   Run Feistel function   Cross over halves of key’s   Inverse Initial Permutation 

  Permutation Choice #1   For each round   Shift bits N times   Permutation Choice #2   Store key for this round 

  Rearrange the order of bit’s based on the given parameters.   General use function, total of 6 different permutations that are performed on the key or data block in DES. /* IP defined on pg: 14 */ InitialPermutation[]= { 58, 50, 42, 34, 26, 18, 10, 2, 60, 52, 44, 36, 28, 20, 12, 4, 62, 54, 46, 38, 30, 22, 14, 6, 64, 56, 48, 40, 32, 24, 16, 8, 57, 49, 41, 33, 25, 17, 9, 1, 59, 51, 43, 35, 27, 19, 11, 3, 61, 53, 45, 37, 29, 21, 13, 5, 63, 55, 47, 39, 31, 23, 15, 7 }; 

  Expansion Permutation on data block   XOR Expanded data with Subkey   For each S-Box   Run selected bits through   Primitive Permutation on combined S-Box output 

Average over 3 runs of 1 Million encryptions: 10.879 seconds 

  Loop unrolling of all permutations   Manual function in lining   Where ever mixer was called we in lined the permutation   While unrolling we discovered that in some places we could combine multiple steps to reduce the amount of times we iterate through the data. 

Average over 3 runs of 1 Million encryptions: 5.958 seconds 

  Before 44 seconds   f() = 22.88%   keyschedule() = 20.2%   algorithm() = 8.2%   mixer() = 40.9   get() = 5.6%   After 14.2 seconds   f() = 42.9%   keyschedule() = 24.4%   algorithm() = 24.2%   N/A   N/A 

  How easy it is to make a critical mistake in cryptography.   There are definitely security flaws in our implementation which would probably make it unusable for actual serious cryptography.   

  A few more loops could be unrolled.   It would be really ugly   It would be very complicated.   It would be possible to inline all functions   Not really a clean implementation, not positive how much would be gained   Change from operations on individual array’s of bits to operations on entire bytes (ex. Skein implementation)   Feistel function may be optimizable, not sure how.