Upload
ashley-lester
View
235
Download
2
Embed Size (px)
Citation preview
2
Contents
Block Cipher Principles
The Data Encryption Standard
The Strength of DES
Differential and Linear Cryptananlysis
Block Cipher Design Principles
3
Block Cipher principles
Stream Ciphers and Block Ciphers
Motivation for the Feistel Cipher Structure
The Feistel Cipher
4
Stream Ciphers and Block Ciphers
Stream cipher encrypts one bit or one byte at a time. Vigenère cipher, Verman cipher
Block cipher encrypts a block of plaintext as a whole to produce a ciphertext block of equal length. Typical block size: 64 or 128 bits
5
Motivation for the Feistel Cipher Structure
A block cipher operates on a plaintext block of n bits to produce a ciphertext block of n bits. Each plaintext must produce a unique ciphertext block (fo
r decryption to be possible). Such transformation is called reversible or nonsingular.
Reversible Mapping Irreversible Mapping
PlaintextCipherte
xtPlaintext
Ciphertext
00 11 00 11
01 10 01 10
10 00 10 01
11 01 11 01
6
Motivation for the Feistel Cipher Structure
The logic of a general substitution cipher. (for n = 4)
7
Motivation for the Feistel Cipher Structure
A practical problem with the general substitution cipher If a small block size is used, then the system is equivalent
to a classical substitution cipher. Such systems are vulnerable to a statistical analysis of the
plaintext.
If block size is sufficiently large and an arbitrary reversible substitution is allowed, then statistical analysis is infeasible. This is not practical from a performance point of view. For n-bit block cipher, the key size is n X 2n bits. For n = 4, the key size is 4 x 16 = 64 bits. For n = 64, the key size is 64 x 2n 16 = 64 bits
8
The Feistel Cipher
Feistel proposed the use of a cipher that alternates substitutions and permutations.
In fact, this is a practical application of a proposal by Claude Shannon to develop a product cipher that alternates confusion and diffusion functions.
9
Diffusion and Confusion
Shannon suggests two methods for frustrating statistical cryptanalysis. Diffusion and Confusion
10
Diffusion and Confusion
Diffusion To make the statistical relationship between the plaintext
and ciphertext as complex as possible in order to thwart attempts to discover the key.
Confusion To make the relationship between the statistics of the
ciphertext and the value of the encryption key as complex as possible to thwart attempts to discover the key.
11
Diffusion and Confusion
Diffusion can be achieved by a permutation followed by a function.
Confusion can be achieved by a substitution.
12
Feistel Cipher Structure
Feistel structure Input
Plaintext : 2w bits A Key K
Output Ciphertext : 2w bits
13
Feistel Cipher Structure
The input is divided into two halves L0 and R0 and they pass through n rounds.
Round i Input: Li-1, Ri-1, and Ki (round key)
Output: Li and Ri
A substitution is performed on the left half Li-1.
A permutation is performed by swapping the two halves.
),( 11
1
iiii
ii
KRFLR
RL
),( 11 iii KRFL
14
Feistel Cipher Structure
Design features Block size
The larger it is, the securer the cipher is but the slower the cipher is.
64 or 128 bits
Key size The larger it is, the securer the cipher is but the slower the
cipher is. 64 or 128 bits
Number of rounds The larger it is, the securer the cipher is but the slower the
cipher is. 16 rounds is typical.
15
Feistel Cipher Structure
Design features Subkey generation
The more complex it is, the securer the cipher is but the slower…
Round function The more complex it is, the securer the cipher is but the
slower…
Fast software encryption/decryption
Ease of analysis
16
Feistel Decryption Algorithm
Decryption is the same as the encryption except that the subkeys are used in reverse order.
17
Feistel Cipher Structure
Round i
),( 11
1
iiii
ii
KRFLR
RL
),( 11
1
iiii
ii
KRFRL
LR
),(1
1
iiii
ii
KLFRL
LR
18
The Data Encryption Standard
DES Encryption
Initial Permutation
Details of Single Round
Key Generation
The Avalanche Effect
19
The Data Encryption Standard
The most widely used encryption. Adopted in 1977 by NIST FIPS PUB 46
Data are encrypted in 64-bit blocks using a 56-bit key.
21
Initial Permutation
The permutation X = IP(M)
The inverse permutation Y = IP-1(X) = IP-1(IP(M))
The original ordering is restored
22
Single Round
),( 11
1
iiii
ii
KRFLR
RL
F function R i-1 is expanded to 48-bits using E. The result is XORed with the 48-bit r
ound key. The 48-bit is substituted by a 32-bit. The 32-bit is permuted by P.
25
Single Round
Each S-box is given in page 79. Outer bits 1 & 6 (row bits) select one rows Inner bits 2-5 (col bits) are substituted
Example : Input : 011001 the row is 01 (row 1) the column is 1100 (column 12) Output is 1001
26
Key Generation
A 64-bit key used as input Every 8th bit is ignored. Thus, the key is 56 bits.
PC1 permute 56 bits into two 28-bit halves.
27
Key Generation
In each round, each 28 bits are rotated left
and 24 bits are selected from each
half.
30
DES Decryption
Decryption uses the same algorithm as encryption. Feistel cipher Roundkey schedule is reversed.
31
The Avalanche Effect
A small change of plaintext or key produces a significant change in the ciphertext.
DES exhibits a strong avalanche effect.
32
The Avalanche Effect
Example
Plaintext 1 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
Plaintext 2 10000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
Key 00000001 1001011 0100100 1100010 0011100 0011000 0011100 0110010
33
The Avalanche Effect
Example
plaintext 01101000 10000101 00101111 01111010 00010011 01110110 11101011 10100100
Key 1 1110010 1111011 1101111 0011000 0011101 0000100 0110001 11011100
Key 2 0110010 1111011 1101111 0011000 0011101 0000100 0110001 11011100
35
The Use of 56-bit Keys
If the key length is 56-bit, we have 256 = 7.2 x 1016 keys.
In 1998, Electronic Frontier Foundation (EFF) announced ‘DES cracker’ which can attack DES in 3 days. It was built for less than $250,000.
Alternatives to DES AES (key size is 128 ~ 256 bit) and triple DES (112 ~ 168 bit)
36
Differential and Linear Cryptanalysis
Differential Cryptanalysis History Differential Cryptanalysis Attack
Linear Cryptanalysis
37
Differential Cryptanalysis
One of the most significant advances in cryptanalysis in recent years is differential cryptanalysis.
38
History
Murphy, Biham & Shamir published 1990.
The first published attack that is capable of breaking DES in less than 255 complexity. As reported, can successfully cryptanalyze DES with an
effort on the order of 247, requiring chosen plaintexts.
This is a powerful tool, but it does not do very well against DES Differential cryptanalysis was known to IBM as early as
1974
39
Differential Cryptanalysis Attack
The differential cryptanalysis attack is complex.
Change in notation for DES Original plaintext block : m
Two halves : m0, m1
At each round for DES, only one new 32-bit block is created. The intermediate message halves are related.
mi1 mi 1 f(mi,K i)
i 1,2,...,16
40
Differential Cryptanalysis Attack
Start with two messages m and m’, and consider the difference between the intermediate message halves : With a known XOR difference
Then
mi mi m'i
m m m'
mi1 mi1 m'i1
[mi 1 f (mi,K i)] [m'i 1 f (m'i ,K i)]
mi 1 m'i 1 f (mi,K i) f (m'i ,K i)
mi 1 [ f (mi,K i) f (m'i ,K i)]
41
Differential Cryptanalysis Attack
The Overall strategy is based one these considerations for a single round. The procedure is
to begin with two plaintext message m and m’ with a given difference.
to trace through a probable pattern of differences after each round to yield a probable difference for the ciphertext.
42
Differential Cryptanalysis Attack
Actually, there are two probable differences for the two 32-bit halves.
Next, submit m and m’ for encryption to determine the actual difference under the unknown key.And compare the result to the probable difference.If there is a match,
Then, suspect that all the probable patterns at all the intermediate rounds are correct.
With that assumption, can make some deductions about the key bits.
m17 ||m16
EK (m) EK (m') (m17 ||m16)
43
Linear Cryptanalysis
another recent development also a statistical method must be iterated over rounds, with decreasing probabilitiesdeveloped by Matsui et al in early 90'sbased on finding linear approximationscan attack DES with 247 known plaintexts, still in practise infeasible
44
Linear Cryptanalysis
find linear approximations with prob p != ½P[i1,i2,...,ia](+)C[j1,j2,...,jb] = K[k1,k2,...,kc]
where ia,jb,kc are bit locations in P,C,K
gives linear equation for key bitsget one key bit using max likelihood algusing a large number of trial encryptions effectiveness given by: |p–½|
45
Block Cipher Design Principles
DES Design Criteria
Number of Rounds
Design of Function F Design Criteria for F S-Box Design
Key Schedule Algorithm
46
Block Cipher Design Principles
Although much progress has been made that are cryptographically strong, the basic principles have not changed all.
47
DES Design Criteria
Focused on the design of the S-boxes and on the P function.
The criteria for the S-boxes. No output bit of any S-box should be too close a linear
function of the input bits. Each row of an S-box should include all 16 possible output
bit combinations If two inputs differ in exactly one bit, the outputs must differ
in at least two bits. If two inputs differ in the two middle bits exactly, the
outputs must differ in at least two bits.
48
DES Design Criteria
The criteria for the S-boxes (~ continue) If two inputs differ in their first two bits and are identical
in their last two bits, the two outputs must not be the same.
For any nonzero 6-bit difference between inputs, no more than 8 of the 32 pairs of inputs exhibiting that difference may result in the same output difference.
This is a criterion similar to the previous one, but for the case of three S-boxes.
49
DES Design Criteria
The criteria for the permutation P The four output bits from each S-box at round i are distributed so that two
of them affect “middle bits” of round (i + 1) and the other two affect end
bits. The two middle bits of input to an S-box are not shared with adjacent
S-boxes. The end bits are the two left-hand bits and the two right-hand
bits, which are shared with adjacent S-boxes.
The four output bits from each S-box affect six different S-boxes on the
next round, and no two affect the same S-boxes.
For two S-boxes j, k, if an output bit from Sj affects a middle bit of Sk on the
next round, then an output bit from Sk cannot affect a middle bit of Sj .
These criteria are intended to increase the diffusion of the algorithm.
50
Number of Rounds
The greater the number of rounds, the more difficult it is to perform cryptanalysis, even for a relatively weak F.
This criterion is attractive because it makes it easy to judge the strength of an algorithm and to compare different algorithms.
51
Design of Function F
The heart of a Feistel block cipher is the function F.
The function F provides the element of confusion. One obvious criterion is that F be nonlinear.
The more nonlinear F, the more difficult.
Have good avalanche properties. Strict Avalanche Criterion (SAC)
The bit independence criterion (BIC) States that output bits j and k should change independently
when any single input bit i is inverted, for all i, j, and k.
52
S-Box Design
One of the most intense areas of research.
One obvious characteristic of the S-box is its size. An n m S-box has n input bits and m output bits.
DES has 6 4 S-boxes. Blowfish has 8 32 S-boxes.
Larger S-boxes are more resistant to differential and linear cryptanalysis. For practical reasons, a limit of n equal to about 8 to 10 is
usually imposed.
53
S-Box Design
S-boxes are typically organized in a different manner than used in DES. An n m S-box typically consists of 2n rows of m bits each. Example, in an 8 32 S-box
If the input is 00001001, the output consists of the 32 bits in row 9.
54
S-Box Design
Mister and Adams proposed for S-box design. S-box should satisfy both SAC and BIC. All linear combinations of S-box columns should be bent.
Bent functions A special class of Boolean functions that are highly nonlinear
according to certain mathematical criteria.
Increasing interest in designing and analyzing S-boxes using bent functions.
55
S-Box Design
Heys, H. and Tavares, S. proposed for S-boxes. Guaranteed avalanche (GA) criterion An S-box satisfies GA of order if, at least output bits
change. Conclude that a GA in the range of order 2 to order 5
provides strong diffusion characteristics for the overall encryption algorithm.
56
S-Box Design
Best method of selecting the S-box entries. Nyberg suggests the following approaches.
Random Use some pseudorandom number generation or some table of
random digits to generate the entries in the S-boxes. Random with testing
Choose S-box entries randomly, then test the results against various criteria, and throw away those that do not pass.
Human-made This is a more or less manual approach with only simple
mathematics to support it. This approach is difficult to carry through for large S-boxes.
Math-made Generate S-boxes according to mathematical principles.
57
Key Schedule Algorithm
With any Feistel block cipher, the key is used to
generate one subkey for each round.
We would like to select subkeys to maximize the
difficulty of deducing individual subkeys and the
difficulty of working back to the main key.
No general principles have not been proposed.
Hall suggests that the key schedule should guarantee
key/ciphertext Strict Avalanche Criterion and Bit
Indepence Criterion.