Data Encryption

Standard3
Modern Block Ciphers

 now look at modern block ciphers

 one of the most widely used types of cryptographic
algorithms
 provide secrecy /authentication services
 focus on DES (Data Encryption Standard)
 to illustrate block cipher design principles
Block vs Stream Ciphers

 block ciphers process messages in blocks, each of which

is then en/decrypted
 like a substitution on very big characters
 64-bits or more
 stream ciphers process messages a bit or byte at a time
when en/decrypting
 many current ciphers are block ciphers
 broader range of applications
Block Cipher Principles

 most symmetric block ciphers are based

on a Feistel Cipher Structure
 needed since must be able to decrypt
ciphertext to recover messages efficiently
 block ciphers look like an extremely large
substitution
 would need table of 264 entries for a 64-
bit block
 instead create from smaller building
blocks
 using idea of a product cipher
Ideal Block Cipher
Claude Shannon and Substitution-
Permutation Ciphers

 Claude Shannon introduced idea of

substitution-permutation (S-P) networks
in 1949 paper
 form basis of modern block ciphers
 S-P nets are based on the two primitive
cryptographic operations seen before:
 substitution (S-box)
 permutation (P-box)
 provide confusion & diffusion of message
& key
Confusion and Diffusion

 cipher needs to completely obscure statistical

properties of original message
 a one-time pad does this
 more practically Shannon suggested combining S & P
elements to obtain:
 diffusion – dissipates statistical structure of plaintext
over bulk of ciphertext
 confusion – makes relationship between ciphertext and
key as complex as possible
Feistel Cipher Structure

 Horst Feistel devised the feistel cipher

 based on concept of invertible product cipher
 partitions input block into two halves
 process through multiple rounds which
 perform a substitution on left data half
 based on round function of right half & subkey
 then have permutation swapping halves
 implements Shannon’s S-P net concept
Feistel Cipher Structure
Feistel Cipher Design Elements

 block size
 key size
 number of rounds
 subkey generation algorithm
 round function
 fast software en/decryption
 ease of analysis
Feistel Cipher Decryption
Data Encryption Standard
(DES)
 most widely used block cipher in world
 adopted in 1977 by NBS (now NIST)
 as FIPS PUB 46
 encrypts 64-bit data using 56-bit key
 has widespread use
 has been considerable controversy over its security
DES History

 IBM developed Lucifer cipher

 by team led by Feistel in late 60’s
 used 64-bit data blocks with 128-bit key
 then redeveloped as a commercial cipher with input
from NSA and others
 in 1973 NBS issued request for proposals for a national
cipher standard
 IBM submitted their revised Lucifer which was
eventually accepted as the DES
DES Design Controversy

 although DES standard is public

 was considerable controversy over design
 in choice of 56-bit key (vs Lucifer 128-bit)
 and because design criteria were classified
 subsequent events and public analysis show in fact design was appropriate
 use of DES has flourished
 especially in financial applications
 still standardised for legacy application use
DES Encryption Overview
DES Round Structure

 uses two 32-bit L & R halves

 as for any Feistel cipher can describe as:
Li = Ri–1
Ri = Li–1  F(Ri–1, Ki)
 F takes 32-bit R half and 48-bit subkey:
 expands R to 48-bits using perm E
 adds to subkey using XOR
 passes through 8 S-boxes to get 32-bit result
 finally permutes using 32-bit perm P
DES Key Schedule

 forms subkeys used in each round

 initial permutation of the key (PC1) which selects 56-bits
in two 28-bit halves
 16 stages consisting of:
 rotating each half separately either 1 or 2 places depending
on the key rotation schedule K
 selecting 24-bits from each half & permuting them by PC2 for
use in round function F
 note practical use issues in h/w vs s/w
DES Decryption

 decrypt must unwind steps of data

computation
 with Feistel design, do encryption steps again
using subkeys in reverse order (SK16 … SK1)
 IP undoes final FP step of encryption
 1st round with SK16 undoes 16th encrypt round
 ….
 16th round with SK1 undoes 1st encrypt round
 then final FP undoes initial encryption IP
 thus recovering original data value
Strength of DES – Key Size

 56-bit keys have 256 = 7.2 x 1016 values

 brute force search looks hard
 recent advances have shown is possible
 in 1997 on Internet in a few months
 in 1998 on dedicated h/w (EFF) in a few days
 in 1999 above combined in 22hrs!
 still must be able to recognize plaintext
 must now consider alternatives to DES
Strength of DES – Analytic
Attacks
 now have several analytic attacks on DES
 these utilise some deep structure of the
cipher
 by gathering information about encryptions
 can eventually recover some/all of the sub-key
bits
 if necessary then exhaustively search for the rest
 generally these are statistical attacks
 include
 differential cryptanalysis
 linear cryptanalysis
 related key attacks
Strength of DES – Timing
Attacks
 attacks actual implementation of cipher
 use knowledge of consequences of implementation to
derive information about some/all subkey bits
 specifically use fact that calculations can take varying
times depending on the value of the inputs to it
 particularly problematic on smartcards
Differential Cryptanalysis

 one of the most significant recent (public) advances in

cryptanalysis
 known by NSA in 70's cf DES design
 Murphy, Biham & Shamir published in 90’s
 powerful method to analyse block ciphers
 used to analyse most current block ciphers with varying
degrees of success
 DES reasonably resistant to it, cf Lucifer
Differential Cryptanalysis

 a statistical attack against Feistel ciphers

 uses cipher structure not previously used
 design of S-P networks has output of function f
influenced by both input & key
 hence cannot trace values back through cipher without
knowing value of the key
 differential cryptanalysis compares two related pairs of
encryptions
Differential Cryptanalysis

 have some input difference giving some output

difference with probability p
 if find instances of some higher probability input /
output difference pairs occurring
 can infer subkey that was used in round
 then must iterate process over many rounds (with
decreasing probabilities)
Differential Cryptanalysis
Differential Cryptanalysis

 perform attack by repeatedly encrypting plaintext pairs

with known input XOR until obtain desired output XOR
 when found
 if intermediate rounds match required XOR have a right pair
 if not then have a wrong pair, relative ratio is S/N for attack
 can then deduce keys values for the rounds
 right pairs suggest same key bits
 wrong pairs give random values
 for large numbers of rounds, probability is so low that more
pairs are required than exist with 64-bit inputs
 Biham and Shamir have shown how a 13-round iterated
characteristic can break the full 16-round DES
Linear Cryptanalysis

 another recent development

 also a statistical method
 must be iterated over rounds, with decreasing
probabilities
 developed by Matsui et al in early 90's
 based on finding linear approximations
 can attack DES with 243 known plaintexts, easier but still
in practise infeasible
Linear Cryptanalysis

 find linear approximations with prob p != ½

P[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 bits
 get one key bit using max likelihood alg
 using a large number of trial encryptions
 effectiveness given by: |p–1/2|
DES Design Criteria

 as reported by Coppersmith in [COPP94]

 7 criteria for S-boxes provide for
 non-linearity
 resistance to differential cryptanalysis
 good confusion
 3 criteria for permutation P provide for
 increased diffusion
Summary
 have considered:
 block vs stream ciphers
 Feistel cipher design & structure
 DES
 details
 strength
 Differential & Linear Cryptanalysis
 block cipher design principles
 Advance
Encryption
Standard
Topics

 Origin of AES

 Basic AES

 Inside Algorithm
Origins

 A replacement for DES was needed

 Key size is too small

 Can use Triple-DES – but slow, small block

 US NIST issued call for ciphers in 1997

 15 candidates accepted in Jun 98

 5 were shortlisted in Aug 99

AES Security

 AES was designed after DES.

 Most of the known attacks on DES were already
tested on AES.
 Brute-Force Attack
 AES is definitely more secure than DES due to the
larger-size key.
 Statistical Attacks
 Numerous tests have failed to do statistical analysis
of the ciphertext
 Differential and Linear Attacks
 There are no differential and linear attacks on AES
as yet.
AES Competition
Requirements
 Private key symmetric block cipher

 128-bit data, 128/192/256-bit keys

 Stronger & faster than Triple-DES

 Provide full specification & design details

 Both C & Java implementations

AES Evaluation Criteria

 initial criteria:
 security – effort for practical cryptanalysis
 cost – in terms of computational efficiency
 algorithm & implementation characteristics

 final criteria
 general security
 ease of software & hardware implementation
 implementation attacks
 flexibility (in en/decrypt, keying, other factors)
AES Shortlist
 After testing and evaluation, shortlist in Aug-99
 MARS (IBM) - complex, fast, high security margin
 RC6 (USA) - v. simple, v. fast, low security margin
 Rijndael (Belgium) - clean, fast, good security margin
 Serpent (Euro) - slow, clean, v. high security margin
 Twofish (USA) - complex, v. fast, high security margin

 Found contrast between algorithms with

 few complex rounds versus many simple rounds
 Refined versions of existing ciphers versus new proposals

Rijndae: pronounce “Rain-Dahl”

The AES Cipher - Rijndael
 Rijndael was selected as the AES in Oct-2000
 Designed by Vincent Rijmen and Joan Daemen in
Belgium
 Issued as FIPS PUB 197 standard in Nov-2001

 An iterative rather than Feistel cipher

 processes data as block of 4 columns of 4 bytes (128 V. Rijmen
bits)
 operates on entire data block in every round

 Rijndael design:
 simplicity
 has 128/192/256 bit keys, 128 bits data
 resistant against known attacks
J. Daemen
 speed and code compactness on many CPUs
Topics

 Origin of AES

 Basic AES

 Inside Algorithm

 Final Notes
AES Conceptual Scheme

Plaintext (128 bits)

AES Key (128-256 bits)

Ciphertext (128 bits)

40
Multiple rounds
 Rounds are (almost) identical
 First and last round are a little different

41
High Level Description

No MixColumns
Overall Structure
128-bit values
 Data block viewed as 4-by-4 table of bytes
 Represented as 4 by 4 matrix of 8-bit bytes.
 Key is expanded to array of 32 bits words

1 byte

44
Data Unit
Unit Transformation
Changing Plaintext to State
 SubBytes Operation
 The SubBytes operation involves 16 independent byte-
to-byte transformations. • Interpret the byte as two
hexadecimal digits xy
S1,1 = xy16 • SW implementation, use row (x)
and column (y) as lookup pointer

x’y’16
SubBytes Table
 Implement by Table Lookup
InvSubBytes Table
Sample SubByte
Transformation
 The SubBytes and InvSubBytes transformations are inverses of each other.
ShiftRows
 Shifting, which permutes the bytes.
 A circular byte shift in each each
 1st row is unchanged
 2nd row does 1 byte circular shift to left
 3rd row does 2 byte circular shift to left
 4th row does 3 byte circular shift to left
 In the encryption, the transformation is
called ShiftRows
 In the decryption, the transformation is
called InvShiftRows and the shifting is to
the right
ShiftRows Scheme
ShiftRows and InvShiftRows
MixColumns
 ShiftRows and MixColumns provide diffusion to the cipher
 Each column is processed separately
 Each byte is replaced by a value dependent on all 4 bytes in the column
 Effectively a matrix multiplication in GF(28) using prime poly m(x)
=x8+x4+x3+x+1
MixClumns Scheme

The MixColumns transformation operates at the column level; it

transforms each column of the state to a new column.
MixColumn and InvMixColumn
AddRoundKey

 XOR state with 128-bits of the round key

 AddRoundKey proceeds one column at a time.

 adds a round key word with each state column matrix
 the operation is matrix addition

 Inverse for decryption identical

 since XOR own inverse, with reversed keys

 Designed to be as simple as possible

AddRoundKey Scheme
AES Round
AES Key Scheduling
 takes 128-bits (16-bytes) key and expands into array of 44 32-bit words
Implementation Aspects

 The algorithms used in AES are so simple that

they can be easily implemented using cheap
processors and a minimum amount of memory.

 Very efficient

 Implementation was a key factor in its selection

as the AES cipher

 AES animation:
 http://www.cs.bc.edu/~straubin/cs381-05/blockciphers/
rijndael_ingles2004.swf