Lecture 3: Block Ciphers and the Data Encryption

Standard
Lecture Notes on Computer and Network Security
by Avi Kak (kak@purdue.edu)
January 15, 2016
12:28am
c
2016
Avinash Kak, Purdue University

Goals:
To introduce the notion of a block cipher in the modern context.
To talk about the infeasibility of ideal block ciphers
To introduce the notion of the Feistel Cipher Structure
To go over DES, the Data Encryption Standard

CONTENTS
Section Title

3.1

Page

Ideal Block Cipher

3.1.1

3.2

Size of the Encryption Key for the Ideal Block Cipher

The Feistel Structure for Block Ciphers

6
7

3.2.1

Mathematical Description of Each Round in the

Feistel Structure

10

3.2.2

Decryption in Ciphers Based on the Feistel Structure

12

3.3

DES: The Data Encryption Standard

16

3.3.1

One Round of Processing in DES

18

3.3.2

The S-Box for the Substitution Step in Each Round

22

3.3.3

The Substitution Tables

26

3.3.4

The P-Box Permutation in the Feistel Function

28

3.3.5

The DES Key Schedule: Generating the Round Keys

30

3.3.6

Initial Permutation of the Encryption Key

33

3.3.7

Contraction-Permutation that Generates the 48-Bit

Round Key from the 56-Bit Key

35

3.4

What Makes DES a Strong Cipher (to the

Extent It is a Strong Cipher)

37

3.5

Homework Problems

39

Computer and Network Security by Avi Kak

Lecture 3

3.1: IDEAL BLOCK CIPHER

In a modern block cipher (but still using a classical encryption

method), we replace a block of N bits from the plaintext with a
block of N bits from the ciphertext. This general idea is illustrated
in Figure 1 for the case of N = 4. (In general, though, N is set
to 64 or multiples thereof.)
To understand Figure 1, note that there are 16 different possible
4-bit patterns. We can represent each pattern by an integer between 0 and 15. So the bit pattern 0000 could be represented by
the integer 0, the bit pattern 0001 by integer 1, and so on. The
bit pattern 1111 would be represented by the integer 15.
In an ideal block cipher, the relationship between the input blocks
and the output block is completely random. But it must be
invertible for decryption to work. Therefore, it has to be one-toone, meaning that each input block is mapped to a unique output
block.
The mapping from the input bit blocks to the output bit blocks
can also be construed as a mapping from the integers correspond3

Computer and Network Security by Avi Kak

Lecture 3

ing to the input bit blocks to the integers corresponding to the

output bit blocks.
The encryption key for the ideal block cipher is the codebook
itself, meaning the table that shows the relationship between the
input blocks and the output blocks.
Figure 1 depicts an ideal block cipher that uses blocks of size 4.
Each block of 4 bits in the plaintext is transformed into a block
of 4 ciphertext bits.

Computer and Network Security by Avi Kak

Lecture 3

b
Plaintext bit block:

b1

b2

Convert 4 incoming bits to one of 16 integers

Random 11 mapping of 16 input integers

to 16 output integers

Convert integer to a 4bit pattern

Ciphertext bit block:

Figure 1: The ideal block cipher when the block size equals
4 bits. (This figure is from Lecture 3 of Lecture Notes on Computer and Network Security by
Avi Kak)

Computer and Network Security by Avi Kak

Lecture 3

3.1.1: The Size of the Encryption Key for the Ideal

Block Cipher

Consider the case of 64-bit block encryption.

With a 64-bit block, we can think of each possible input block
as one of 264 integers and for each such integer we can specify an output 64-bit block. We can construct the codebook by
displaying just the output blocks in the order of the integers corresponding to the input blocks. Such a code book will be of size
64 264 1021.
That implies that the encryption key for the ideal block cipher
using 64-bit blocks will be of size 1021.

The size of the encryption key would make the ideal block cipher
an impractical idea. Think of the logistical issues related to the
transmission, storage, and processing of such large keys.

Computer and Network Security by Avi Kak

Lecture 3

3.2: The Feistel Structure for Block Ciphers

The DES algorithm for encryption and decryption, which is the

main theme of this lecture, is based on what is known as the Feistel
Structure. This section and the next two subsections introduce this
structure:

Named after the IBM cryptographer Horst Feistel and first implemented in the Lucifer cipher by Horst Feistel and Don Coppersmith.
A cryptographic system based on Feistel structure uses the same
basic algorithm for both encryption and decryption.
As shown in Figure 2, the Feistel structure consists of multiple
rounds of processing of the plaintext, with each round consisting
of a substitution step followed by a permutation step.
The input block to each round is divided into two halves that we
can denote L and R for the left half and the right half.
7

Computer and Network Security by Avi Kak

Lecture 3

In each round, the right half of the block, R, goes through unchanged. But the left half, L, goes through an operation that
depends on R and the encryption key.
The permutation step at the end of each round consists of swapping the modified L and R. Therefore, the L for the next round
would be R of the current round. And R for the next round
be the output L of the current round.

Computer and Network Security by Avi Kak

Lecture 3

Plaintext block
(Divide into two halves, L and R)

Round Keys
K
1

Round1
F(K,R)

R
K
2

Round

F(K,R)

L
Round n

R
K
n
F(K,R)

Ciphertext block

Figure 2: The Feistel Structure for symmetric key cryptography (This figure is from Lecture 3 of Lecture Notes on Computer and Network Security
by Avi Kak)
9

Computer and Network Security by Avi Kak

Lecture 3

3.2.1: Mathematical Description of Each Round in the

Feistel Structure

Let LEi and REi denote the output half-blocks at the end of the
ith round of processing. The letter E denotes encryption.
In the Feistel structure, the relationship between the output of
the ith round and the output of the previous round, that is, the
(i 1)th round, is given by
LEi
REi

=
=

REi1
LEi1 F (REi1, Ki)

where denotes the bitwise EXCLUSIVE OR operation. The

symbol F denotes the operation that scrambles REi1 of the
previous round with what is shown as the round key Ki in
Figure 2. The round key Ki is derived from the main encryption
key as we will explain later.
F is referred to as the Feistel function, after Horst Feistel naturally.
Assuming 16 rounds of processing (which is typical), the output
of the last round of processing is given by
10

Computer and Network Security by Avi Kak

LE16
RE16

Lecture 3

=
=

RE15
LE15 F (RE15, K16)

11

Computer and Network Security by Avi Kak

Lecture 3

3.2.2: Decryption in Ciphers Based on the Feistel

Structure

As shown in Figure 3, the decryption algorithm is exactly the

same as the encryption algorithm with the only difference that
the round keys are used in the reverse order.
The output of each round during decryption is the
input to the corresponding round during encryption
except for the left-right switch between the two
halves. This property holds true regardless of the
choice of the Feistel function F .
To prove the above claim, let LDi and RDi denote the left half
and the right half of the output of the ith round.
That means that the output of the first decryption round consists of LD1 and RD1. So we can denote the input to the first
decryption round by LD0 and RD0. The relationship between
the two halves that are input to the first decryption round and
what is output by the encryption algorithm is:

12

Computer and Network Security by Avi Kak

Lecture 3

LD0
RD0

=
=

RE16
LE16

We can write the following equations for the output of the first
decryption round
LD1

=
=
=

RD0
LE16
RE15

RD1

=
=
=
=

LD0 F (RD0, K16)

RE16 F (LE16, K16)
[LE15 F (RE15, K16)] F (RE15, K16)
LE15

This shows that, except for the left-right switch, the output of
the first round of decryption is the same as the input to the last
stage of the encryption round since we have LD1 = RE15 and
RD1 = LE15
The following equalities are used in the above derivation. Assume
that A, B, and C are bit arrays.

[A B] C

=
13

A [B C ]

Computer and Network Security by Avi Kak

Lecture 3

A A
A 0

=
=

0
A

The above result is independent of the precise nature

of the Feistel function F . That is, the output of each round
during decryption is the input to the corresponding round during
encryption for every choice of the Feistel function F .

14

Computer and Network Security by Avi Kak

Lecture 3

Encryption

Decryption
Round Keys

Plaintext block
(Divide into two halves, L and R)

Plaintext block
(Divide into two halves, L and R)

K
1
LE

RE

LD

=
16

RE

RD

LE
0

16

F(K,R)

Round1

F(K,R)
LE
1

RE

K
2
LD

RD =
15

RE
1

15

LE

F(K,R)
Round

F(K,R)
LE

RE

15

15

K
16
LD

Round 16

RE

RD
1

15

LE

15

F(K,R)
F(K,R)

LE
16

RE

16

LD

Ciphertext block

RE

RD
16

Ciphertext block

Figure 3: When a Feistel structure is used, decryption

works the same as encryption. (This figure is from Lecture 3 of Lecture
Notes on Computer and Network Security by Avi Kak)

15

LE
16

Computer and Network Security by Avi Kak

Lecture 3

3.3: DES: THE DATA ENCRYPTION

STANDARD

Adopted by NIST in 1977.

Based on a cipher (Lucifer) developed earlier by IBM for Lloyds
of London for cash transfer.
DES uses the Feistel cipher structure with 16 rounds of processing.
DES uses a 56-bit encryption key. (The key size was apparently
dictated by the memory and processing constraints imposed by
a single-chip implementation of the algorithm for DES.) The key
itself is specified with 8 bytes, but one bit of each byte is used as
a parity check.
DES encryption was broken in 1999 by Electronics
Frontiers Foundation (EFF, www.eff.org). This resulted
in NIST issuing a new directive that year that required organizations to use Triple DES, that is, three consecutive applications
16

Computer and Network Security by Avi Kak

Lecture 3

of DES. (That DES was found to be not as strong as originally

believed also prompted NIST to initiate the development of new
standards for data encryption. The result is AES that we will
discuss later.)
Triple DES continues to enjoy wide usage in commercial applications even today. To understand Triple DES, you must first
understand the basic DES encryption.
As mentioned, DES uses the Feistel structure with 16 rounds.
What is specific to DES is the implementation of the F function
in the algorithm and how the round keys are derived from the
main encryption key.
The round keys are generated from the main key by a sequence
of permutations. Each round key is 48 bits in length.

17

Computer and Network Security by Avi Kak

Lecture 3

3.3.1: One Round of Processing in DEA

The algorithmic implementation of DES is known as DEA for

Data Encryption Algorithm.
Figure 4 shows a single round of processing in DEA. The dotted
rectangle constitutes the F function.
The 32-bit right half of the 64-bit input data block is expanded
by into a 48-bit block. This is referred to as the expansion
permutation step, or the E-step.
The above-mentioned E-step entails the following:
first divide the 32-bit block into eight 4-bit words
attach an additional bit on the left to each 4-bit word that is
the last bit of the previous 4-bit word
attach an additional bit to the right of each 4-bit word that is
the beginning bit of the next 4-bit word.
Note that what gets prefixed to the first 4-bit block is the last bit
of the last 4-bit block. By the same token, what gets appended
to the last 4-bit block is the first bit of the first 4-bit block. The
18

Computer and Network Security by Avi Kak

Lecture 3

reason for why we expand each 4-bit block into a 6-bit block in
the manner explained will become clear shortly.
The 56-bit key is divided into two halves, each half shifted separately, and the combined 56-bit key permuted/contracted
to yield a 48-bit round key. How this is done will be explained
later.
The 48 bits of the expanded output produced by the E-step are
XORed with the round key. This is referred to as key mixing.
The output produced by the previous step is broken into eight
six-bit words. Each six-bit word goes through a substitution step;
its replacement is a 4-bit word. The substitution is carried out
with an S-box, as explained in greater detail in Section 3.3.2.
[The name S-Box stands for Substitution Box.]
So after all the substitutions, we again end up with a 32-bit word.
The 32-bits of the previous step then go through a P-box based
permutation, as shown in Figure 4.
What comes out of the P-box is then XORed with the left half
of the 64-bit block that we started out with. The output of this
19

Computer and Network Security by Avi Kak

Lecture 3

XORing operation gives us the right half block for the next round.
Note that the goal of the substitution step implemented by the
S-box is to introduce diffusion in the generation of the output
from the input. Diffusion means that each plaintext bit must
affect as many ciphertext bits as possible.
The strategy used for creating the different round keys from the
main key is meant to introduce confusion into the encryption
process. Confusion in this context means that the relationship between the encryption key and the ciphertext must be
as complex as possible. Another way of describing confusion
would be that each bit of the key must affect as many bits as
possible of the output ciphertext block.
Diffusion and confusion are the two cornerstones of block cipher
design.

20

Computer and Network Security by Avi Kak

LE i1

Lecture 3

RE i1
32 bits

32 bits

Expansion Permutation

The Feistel Function

F( RE i1 , K i )

48 bits
Round Key K
48 bits
Substitution with 8 Sboxes
32 bits
Permutation with PBox

LE

RE

Figure 4: One round of processing in DES. (This figure is from

Lecture 3 of Lecture Notes on Computer and Network Security by Avi Kak)

21

Computer and Network Security by Avi Kak

Lecture 3

3.3.2: The S-Box for the Substitution Step in Each

Round

As shown in Figure 5, the 48-bit input word is divided into eight

6-bit words and each 6-bit word fed into a separate S-box. Each
S-box produces a 4-bit output. Therefore, the 8 S-boxes together
generate a 32-bit output. As you can see, the overall substitution
step takes the 48-bit input back to a 32-bit output.
Each of the eight S-boxes consists of a 4 16 table lookup for an
output 4-bit word. The first and the last bit of the 6-bit input
word are decoded into one of 4 rows and the middle 4 bits decoded
into one of 16 columns for the table lookup.
The goal of the substitution carried out by an S-box is to enhance
diffusion, as mentioned previously. As you will recall from the
E-step described in Section 3.3.1, the expansion-permutation step
(the E-step) expands a 32-bit block into a 48-bit block by attaching a bit at the beginning and a bit at the end of each 4-bit
sub-block, the two bits needed for these attachments belonging
to the adjacent blocks.
Thus, the row lookup for each of the eight S-boxes becomes a
function of the input bits for the previous S-box and the next
22

Computer and Network Security by Avi Kak

Lecture 3

S-box.
In the design of the DES, the S-boxes were tuned to enhance the
resistance of DES to what is known as the differential cryptanalysis attack, or, sometimes more briefly as differential attack. [As will be explained in much greater detail (and also demonstrated) in Section 8.9 of Lecture
8, differential cryptanalysis of block ciphers consists of presenting to the encryption algorithm pairs of
plaintext bit patterns with known differences between them and examining the differences between the
corresponding cyphertext outputs. The outputs are usually recorded at the input to the last round of
the cipher. Lets represent one plaintext bit block by X = [X1 , X2 , ...., Xn ] where Xi denotes the ith bit
in the block, and lets represent the corresponding output bit block by Y = [Y1 , Y2 , ..., Yn ]. By the difference between two plaintext bit blocks X and X we mean X = X X . The difference between
the corresponding outputs Y and Y is given by Y = Y Y . The pair (X, Y ) is known as a
differential. In an ideally randomizing block cipher, the probability of Y being a particular value for
a given X is 1/2n for an n-bit block cipher. What is interesting is that the probabilities of Y taking
on different values for a given X can be shown to be independent of the encryption key because of the
properties of the XOR operator, but these probabilities are strongly dependent on the S-box tables. By
feeding into a cipher several pairs of plaintext blocks with known X and observing the corresponding
Y , it is possible to establish constraints on the round key bits encountered along the different paths
in the encryption processing chain. (By constraints I mean the following: Speaking hypothetically for
the purpose of illustrating a point and focusing on just one round of DES, suppose we can show that
the following condition can be expected to be obeyed with high probability: Xi Yi Ki = 0
for some bit Ki of the encryption key, then it must be the case that Ki = X Y .) Note that
differential cryptanalysis is a chosen plaintext attack, meaning that the attacker will feed known
plaintext bit patterns into the cipher and analyze the corresponding outputs in order to figure out the
encryption key. In a theoretical analysis of an attack based on differential cryptanalysis, it was shown
by Eli Biham and Adi Shamir in 1990 that the DESs encryption key could be figured out provided one

23

Computer and Network Security by Avi Kak

Lecture 3

could feed known 247 plaintext blocks into the cipher. For a tutorial by Howard Heys on differential
cryptanalysis, see http://www.engr.mun.ca/~howard/PAPERS/ldc_tutorial.pdf. The title of the
tutorial is A Tutorial on Linear and Differential Cryptanalysis.

24

Computer and Network Security by Avi Kak

Lecture 3

3.3.3: The Substitution Tables

Shown on the next page are the eight S-boxes, S1 through S8,
each S-box being a 416 substitution table that is used to convert
6 incoming bits into 4 outgoing bits.
As mentioned earlier, each row of a substitution table is indexed
by the two outermost bits of a six-bit block and each column by
the remaining inner 4 bit.

25

Computer and Network Security by Avi Kak

Lecture 3

48 bits produced by XORing the output of the Expansion

Permutation and the Round Key

48 bits

S1

S2

S3

S4

S5

S6

S7

S8

32 bits

Figure 5: The 48 bits coming out of the expansion permutation are first XORed with the round key and then, as
shown, fed into the 8 S-boxes of DES. (This figure is from Lecture 3 of
Lecture Notes on Computer and Network Security by Avi Kak)

26

Computer and Network Security by Avi Kak

14
0
4
15

4 13
15 7
1 14
12 8

15
3
0
13

1 8
13 4
14 7
8 10

10
13
13
1

0 9
7 0
6 4
10 13

7
13
10
3

13 14
8 11
6 9
15 0

2
14
4
11

12 4
11 2
2 1
8 12

12
10
9
4

1 10
15 4
14 15
3 2

4
13
1
6

11 2
0 11
4 11
11 13

13
1
7
2

2 8
15 13
11 4
1 14

Lecture 3

The 4 16 substitution table

1 2 15 11 8 3 10
4 14 2 13 1 10 6
8 13 6 2 11 15 12
2 4 9 1 7 5 11
S-box S2
14 6 11 3 4 9 7
7 15 2 8 14 12 0
11 10 4 13 1 5 8
1 3 15 4 2 11 6
S-box S3
14 6 3 15 5 1 13
9 3 4 6 10 2 8
9 8 15 3 0 11 1
0 6 9 8 7 4 15
S-box S4
3 0 6 9 10 1 2
5 6 15 0 3 4 7
0 12 11 7 13 15 1
6 10 1 13 8 9 4
S-box S5
1 7 10 11 6 8 5
12 4 7 13 1 5 0
11 10 13 7 8 15 9
7 1 14 2 13 6 15
S-box S6
15 9 2 6 8 0 13
2 7 12 9 5 6 1
5 2 8 12 3 7 0
12 9 5 15 10 11 14
S-box S7
14 15 0 8 13 3 12
7 4 9 1 10 14 3
13 12 3 7 14 10 15
8 1 4 10 7 9 5
S-box S8
4 6 15 11 1 10 9
8 10 3 7 4 12 5
1 9 12 14 2 0 6
7 4 10 8 13 15 12

27

for S-box S1
6 12 5 9
12 11 9 5
9 7 3 10
3 14 10 0
0
9
3
5

0 7
3 8
5 0
6 13

2 13
1 10
12 6
7 12

12
6
9
0

12 7
5 14
2 12
14 3

11 4 2 8
12 11 15 1
5 10 14 7
11 5 2 12

8 5
2 12
3 14
5 11

11 12 4 15
1 10 14 9
5 2 8 4
12 7 2 14

3 15
15 10
12 5
0 9

13
3
6
10

3 4
13 14
4 10
1 7

14 7 5 11
0 11 3 8
1 13 11 6
6 0 8 13

9 7
5 12
6 8
0 15
3 14
6 11
10 13
9 0

0
9
3
4

5 10
11 5
2 15
14 9

5 10
2 15
0 5
14 2

14 9
8 6
0 14
5 3

6 1
8 6
9 2
3 12

5 0 12 7
0 14 9 2
15 3 5 8
3 5 6 11

Computer and Network Security by Avi Kak

Lecture 3

3.3.4: The P-Box Permutation in the Feistel Function

The last step in the Feistel function shown in Figure 4 is labeled

Permutation with P-Box. The permutation sequence is shown
below. [It looks like a table, but it is not as explained below]

15
0
1
18

P-Box Permutation
6 19 20 28 11 27 16
14 22 25 4 17 30 9
7 23 13 31 26 2 8
12 29 5 21 10 3 24

This permutation table says that the 0th output bit will be the
15th bit of the input, the 1st output bit the 6th bit of the input,
and so on, for all of the 32 bits of the output that are obtained
from the 32 bits of the input.
Do NOT associate any meaning with the row-organization of the
table except for the following: Each row of the table tells us
how to select the input bits for the output byte corresponding to
the row. For example, for the second output byte, the first entry
in the second row means that the 0th bit of the second output
byte meaning the 8th bit of the output will be the 0th bit
28

Computer and Network Security by Avi Kak

Lecture 3

of the 32-bit input. Note that bit indexing is 0-based as it

would be in your Perl or Python script
Keep in mind the fact that, when using the BitVector module
in Python or the Algorithm::BitVector module in Perl, a permutation such as the one shown above can be carried out with a
one-line command. For example, in Python, the code fragment
would look like:
sboxes_output =

BitVector representation of the

output of the S-Boxes
right_half = sboxes_output.permute( pbox_permutation )

where permute() is a method defined for the BitVector class.

The argument pbox permutation you see above is the sequence
of all the entries in the table on the previous page expressed as
a one-dimensional array.

29

Computer and Network Security by Avi Kak

Lecture 3

3.3.5: The DES Key Schedule: Generating the Round

Keys

The initial 56-bit key may be represented as 8 bytes, with the last
bit (the least significant bit) of each byte used as a parity bit.
The relevant 56 bits are subject to a permutation at the beginning before any round keys are generated. This is referred to as
Permutation Choice 1 that is shown in Section 3.3.6.
At the beginning of each round, we divide the 56 relevant key bits
into two 28 bit halves and circularly shift to the left each half by
one or two bits, depending on the round, as shown in the table
on the next page.
For generating the round key, we join together the two halves and
apply a 56 bit to 48 bit contracting permutation (this is referred
to as Permutation Choice 2, as shown in Section 3.3.7) to the
joined bit pattern. The resulting 48 bits constitute our round
key.
The contraction permutation shown in Permutation Choice 2,
along with the one-bit or two-bit rotation of the two key halves
30

Computer and Network Security by Avi Kak

Lecture 3

prior to each round, is meant to ensure that each bit of the original
encryption key is used in roughly 14 of the 16 rounds.
The two halves of the encryption key generated in each round are
fed as the two halves going into the next round.
The table shown below tells us how many positions to use for
the left circular shift that is applied to the two key halves at the
beginning of each round:

Round Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

Number of left shifts

1
1
2
2
2
2
2
2
1
2
2
2
2
2
2
1

When using the BitVector module for programming in Python,

or the Algorithm::BitVector module for programming in Perl,
the steps described above for splitting the 56-bit key, circularshifting each half separately, and then rejoining the two halves
31

Computer and Network Security by Avi Kak

Lecture 3

can be carried simply by a command sequence that in Python

looks like
[left,right] = key_bv.divide_into_two()
left
<< shifts[i]
right << shifts[i]
rejoined_key_bv = left + right

where key bv is the BitVector representation of the 56-bit key

entering the round and shifts is the array that consists of the
second column entries in the table shown on the previous page.
The method divide into two() is defined for the BitVector
class.

32

Computer and Network Security by Avi Kak

Lecture 3

3.3.6: Initial Permutation of the Encryption Key

Permutation
56 48 40 32
0 57 49 41
9 1 58 50
18 10 2 59
62 54 46 38
6 61 53 45
13 5 60 52
20 12 4 27

Choice 1
24 16 8
33 25 17
42 34 26
51 43 35
30 22 14
37 29 21
44 36 28
19 11 3

The bit indexing is based on using the range 0-63 for addressing
the bit positions in an 8-byte bit pattern in which the last bit of
each byte is used as a parity bit. [Note that each row shown above has has
only 7 positions the positions corresponding to the parity bit are NOT included above.
That is, you will NOT see the positions 7, 15, etc., listed in the permutations shown.
Nevertheless, the bit addressing spans the full 0-63 range.]

The permutations
shown above do not constitute a table, in the sense that the
rows and the columns do NOT carry any special and separate
meanings. The permutation order for the bits is given by reading
the entries shown from the upper left corner to the lower right
corner.
This permutation tells us that the 0th bit of the output will be
33

Computer and Network Security by Avi Kak

Lecture 3

the 56th bit of the input (in a 64 bit representation of the 56-bit
encryption key), the 1st bit of the output the 48th bit of the input,
and so on, until finally we have for the 55th bit of the output the
3rd bit of the input.
When programming in Python using the BitVector module, or in
Perl using the Algorithm::BitVector module, the permutations
shown on the previous page can be carried out trivially by calling the permute() method of the modules. Using Python to
illustrate, you could call
user_key_bv = BitVector( textstring = user-supplied_key )
key_bv = user_key_bv.permute( initial_permutation )

where, as mentioned earlier, permute() is a method defined for

the BitVector class and initial permutation is the permutation shown on the previous slide expressed as a 1-dimensional
array of integers.

34

Computer and Network Security by Avi Kak

Lecture 3

3.3.7: Contraction-Permutation that Generates the

48-Bit Round Key from the 56-Bit Key

13
14
25
40
50
33

Permutation Choice 2
16 10 23 0 4 2
5 20 9 22 18 11
7 15 6 26 19 12
51 30 36 46 54 29
44 32 47 43 48 38
52 45 41 49 35 28

27
3
1
39
55
31

As on the previous page, bit addressing shown above uses the full
0-63 range in an 8-byte pattern. Since the last bit of each byte is
used as a parity bit, you will not see the bit positions 7, 15, 23,
etc., in the permutation shown above.
As with permutation shown on the previous page, what is shown
above is NOT a table, in the sense that the rows and the columns
do not carry any special and separate meanings. The permutation
order for the bits is given by reading the entries shown from the
upper left corner to the lower right corner.
Since there are only six rows and there are 8 positions in each
35

Computer and Network Security by Avi Kak

Lecture 3

row, the output will consist of 48 bits.

When programming in Python using the BitVector class, the
permutations shown on the previous page can be carried out trivially by calling the permute() method of the class, as mentioned
earlier.

36

Computer and Network Security by Avi Kak

Lecture 3

3.4: WHAT MAKES DES A STRONG

CIPHER (TO THE EXTENT IT IS A
STRONG CIPHER)

The substitution step is very effective as far as diffusion is concerned. It has been shown that if you change just one bit of the
64-bit input data block, on the average that alters 34 bits of the
ciphertext block.
The manner in which the round keys are generated from the
encryption key is also very effective as far as confusion is concerned. It has been shown that if you change just one bit of
the encryption key, on the average that changes 35 bits of the
ciphertext.
Both effects mentioned above are referred to as the avalanche
effect.
And, of course, the 56-bit encryption key means a key space of
size 256 7.2 1016.

37

Computer and Network Security by Avi Kak

Lecture 3

Assuming that, on the average, youd need to try half the keys
in a brute-force attack, a machine able to process 1000 keys per
microsecond would need roughly 13 months to break the code.
However, a parallel-processing machine trying 1 million keys simultaneously would need only about 10 hours. (EFF took
three days on a specially architectured machine to
break the code.)
The official document that presents the DES standard can be
found at:
http://www.itl.nist.gov/fipspubs/fip46-2.htm

38

Computer and Network Security by Avi Kak

Lecture 3

3.5: HOMEWORK PROBLEMS

1. A text file named myfile.txt that you created with a run-ofthe-mill editor contains just the following word:
hello
If you examine this file with a command like
hexdump

-C

myfile.txt

you are likely to see the following bytes (in hex) in the file:
68

65

6C

6C

6F

0A

Lets now try to encrypt the contents of this text file with a 4-bit
block cipher whose codebook contains the following entries:
6, 0, 13, 4, 3, 1, 14, 8, 7, 12, 9, 15, 5, 2, 11, 10

Lets say that I write the encrypted output into a different file and
then examine this new file with the hexdump -C command.
What will I see in the encrypted file?
2. In general, in a block cipher, we replace N bits from the plaintext
with N bits of ciphertext. What defines an ideal block cipher?

39

Computer and Network Security by Avi Kak

Lecture 3

3. Whereas it is true that the relationship between the input and

the output is completely random for an ideal block cipher, it must
nevertheless be invertible for decryption to work. That implies
that the mapping between the input blocks and the output blocks
must be one-to-one. If we had to express this mapping in the form
of a table lookup, what will be the size of the table?
4. What would be the encryption key for an ideal block cipher?
5. What makes ideal block ciphers impractical?
6. What do we mean by a Feistel Structure for Block Ciphers?
7. Are there any constraints on the Feistel function F in a Feistel
structure?
8. Explain the concepts of diffusion and confusion as used in DES.
9. If we have all the freedom in the world for choosing the Feistel
function F, how should we specify it?
10. How does the permutation/expansion step in DES enhance diffusion? This is the step in which we expand by permutation and
repetition the 32-bit half-block into a 48-bit half-block
40

Computer and Network Security by Avi Kak

Lecture 3

11. DES encryption was broken in 1999. Why do you think that
happened?
12. Since DES was cracked, does that make this an unimportant
cipher?

13. Programming Assignment 1:

Write a Perl or Python script that implements the full DES. Use
the S-boxes that are specified for the DES standard (See Section
3.3.3). Make sure you implement all of the key generation steps
outlined in Section 3.3.5. For the encryption key, your script
should prompt the user for a keyboard entry that consists of at
least 8 printable ASCII characters. (You may choose to either
use the first seven or the last seven bits of each character byte for
the 56-bit key you need for DES.)
What makes this homework not as difficult as you think is that
once you write the code that carries out one round of processing,
you basically use the same code in a loop for the whole encryption chain and the decryption chain. Obviously, you will have
to reverse the order in which the round keys are used for the
decryption chain.
Although you are free to write your own code from scratch, here
are some recommendations: If using Python, you might want to
start with the Python BitVector class. To help you get started
with the Python implementation, please see the hw2_starter.py
41

Computer and Network Security by Avi Kak

Lecture 3

file. If using Perl, use the Bit::Vector module from www.cpan.org.

It is a popular Perl module for manipulating bit arrays. It is also
well documented. To help you get started with the Perl implementation, please see the hw2_starter.pl file. You can download both these starter files through the code archive for Lecture
3.

14. Programming Assignment 2:

Now modify the implementation you created for the previous
homework by filling the 4 16 tables for the S-boxes with randomly generated integers. Obviously, each randomly generated
entry will have to be between 0 and 15, both ends inclusive. Calculate the avalanche effect for this implementation of DES and
compare it with the same effect for your previous implementation.
(See Section 3.3.1 for the avalanche effect.)

42