Score:1

AES - SBOX avalanche effect question

dz flag

I am making a program that verifies the avalanche effect for the SBOX of the AES block cipher for a certain number of iterations. This program changed one bit randomly in one of the elements of the input list (state) and then applied the subBytes operation to both the original and modified lists. After that, it counts the number of different bits between these two resulting lists.

sbox = [
    0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5, 0x30, 0x01,   0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76,
    0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0, 0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0,
    0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc, 0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15,
    0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a, 0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75,
    0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0, 0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84,
    0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b, 0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf,
    0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85, 0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8,
    0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5, 0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2,
    0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17, 0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73,
    0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88, 0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb,
    0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c, 0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79,
    0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9, 0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08,
    0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6, 0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a,
    0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e, 0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e,
    0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94, 0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf,
    0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68, 0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16
]
sboxInv = [
    0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38, 0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb,
    0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87, 0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb,
    0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d, 0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e,
    0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2, 0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25,
    0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16, 0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92,
    0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda, 0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84,
    0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a, 0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06,
    0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02, 0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b,
    0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea, 0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73,
    0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85, 0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e,
    0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89, 0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b, 
    0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20, 0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4, 
    0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31, 0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f, 
    0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d, 0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef, 
    0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0, 0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61, 
    0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26, 0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d
]
def subBytes(state):
    result = [None] * 16
    for i in range(len(state)):
        result[i] = sbox[state[i]]
    return result

def subBytesInv(state):
    result = [None] * 16
    for i in range(len(state)):
        result[i] = sboxInv[state[i]]
    return result

#Changes a random bit in one of the elements of the state
def changeBit(state):
    result = state[:]
    result[random.randrange(0,16)] ^= 1 << random.randrange(0,8)
    return result

#Return index of first element that is different in the list
def findDiff(list1, list2):
    for i in range(len(list1)):
        if list1[i] != list2[i]:
            return i
    return -1

#Count number of different bits between two integers
def countDiffBits(a, b):
    return bin(a ^ b).count('1')


#Default array
state=[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16]


average = []
for i in range(10000):
    #Create new state with random bit changed
    changed_state = changeBit(state)

    #Apply bin() to each element and print
    print(state)
    sbox_state = subBytes(state) 
    print(list(map(bin,sbox_state)))

    print(changed_state)
    sbox_changed_state = subBytes(changed_state) 
    print(list(map(bin,sbox_changed_state)))

    index = findDiff(state,changed_state)
    print("Changed bits: " + str(countDiffBits(sbox_state[index],sbox_changed_state[index])))

    average.append(countDiffBits(sbox_state[index],sbox_changed_state[index]))

print("Statistics: " + str(Counter(average)))

And this is the output:

Statistics: Counter({5: 2611, 3: 2323, 4: 2236, 2: 1682, 6: 656, 1: 340, 8: 77, 7: 75})

However, as you can see, for 10000 iterations there are 340 cases where only one bit is changed between the two lists, which means that the avalanche effect isn't happening. My question now is if this is some sort of error in the code or if I'm misunderstanding this concept.

poncho avatar
my flag
BTW: Monte Carlo tests are for the times that the problem space is too large to do exhaustive search (which is most of the time, admittedly). This isn't one of those times - there are only 1024 possible input pairs with a single bit differential - you can exhaustively search them...
Score:2
my flag

However, as you can see, for 10000 iterations there are 340 cases where only one bit is changed between the two lists

That's a little higher than expected. For the AES SBox, there are 24 pairs of inputs (counting 02 and 0a as the same pair as 0a and 02) with a single bit input differential that result in a single bit output differential; from sampling 10000 pairs randomly, one would expect to see 234 or so pairs.

Update: I see the problem: your code sets the first set of the pair to be a random value between 0 and 15 (not 255). It turns out that AES has three such pairs that yield a single bit differential output (02,0a, 04,14 and 08,28); that gives an expected 312 matches (out of 10,000); the difference between that and the 340 you saw can be ascribed to randomness.

I'm misunderstanding this concept.

Actually, AES does not rely on the sbox to avalanche (the combination of MixCollumns and ShiftRows does that nicely); instead, it relies on the sbox to have low probability against linear and differential characteristics (and the presence of a handful of 1 bit -> 1 bit differential pairs does not conflict with that)

I sit in a Tesla and translated this thread with Ai:

mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.