cryptographic applications kris gaj tarek el-ghazawi & gmu, gwu

Cryptographic applications

Kris GajTarek El-Ghazawi

& GMU, GWU students

Cipher

message

cryptographickey

K bits

ciphertext2

Definition of a cipher

M C

family of encipheringtransformations

family of decipheringtransformations

K key K

message space ciphertext spacekey space

enciphering transformation EK(M)

deciphering transformation DK(C)

K ∈ K 3 DK(EK(M)) = MM ∈ M

Substitution Cipher

Key =a b c d e f g h i j k l m n o p q r s t u v w x y z

f q i s h n c v j t y a u w d r e x l b m z o g k p

TO BE OR NOT TO BE

BD QH DX WDB BD QH

enciphering

decipheringTO BE OR NOT TO BE

Number of keys = 26! ≈ 4 ⋅ 1026

4

Secret-key (Symmetric) Cryptosystems

key of Alice and Bob - KABkey of Alice and Bob - KAB

Network

DecryptionEncryptionBobAlice

5

Key Distribution Problem

6

N - Users N · (N-1)2

Keys

Users Keys

5,0001001000 500,000

Public Key (Asymmetric) Cryptosystems

Public key of Bob - KB Private key of Bob - kB

Alice Bob

Network

Encryption Decryption

7

Key exchange for secret-key ciphers

Network

session key(random secret-key) BobAlice

Bob’s public key Bob’s private key

Session keyencrypted using Bob’s public key

Message encrypted using session key

8

Digital SignatureAlice Bob

SignatureMessageSignatureMessage

9

Hash function

Public keycipher

Hash function

Alice’s public key

Hash value 1

Hash value 2

Hash value

Public key cipher

yes no

Alice’s private key

High-throughput secret-key encryption

10

High-Throughput Encryption. . . .

Mi

Mi+1

Mi+2

K0

11

Encryption

Encryption algorithms:

DES, 3DES, AES, RC5, IDEA, etc.

Ci

Ci+1

Ci+2

Fully Pipelined Architecture

. . . .

. . . .

. . . .

Loop unrolling

Pipeline stages inside of cipher rounds

New input & new output every clock cycle

Round 1

Round 2

. . .. . . .

Round k

12

Secret-key cipher breaking

13

Secret-key cipher breaking

Given:

14

Looked for:

Method:

remaining plaintext

ciphertextguessed fragment of the plaintext

or key

successivekeys

exhaustive key search (brute-force) attack

cipher

Secret-key breaking

Message – Ciphertext pair

C0M0

Negligibly smallinput/output

Huge amountof computations

K1 K2 K3 KN

Cipherbreaker

…

Generated by the cipher breakerCorrect key

15

Public-key cryptography forkey exchange

and digital signatures

16

RSA public-key encryption

C = f(M) = Me mod N C

PUBLIC KEYmessage ciphertext

M

M = f-1(C) = Cd mod N

PRIVATE KEY

P, Q - large prime numbersN = P ⋅ Q17e ⋅ d ≡ 1 mod ((P-1)(Q-1))

RSA keysPUBLIC KEY PRIVATE KEY

{ e, N } { d, P, Q }

P, Q - large prime numbersP, Q:

N: N = P ⋅ Q

e: gcd(e, P-1) = 1 and gcd(e, Q-1) = 1

d:18

e ⋅ d ≡ 1 mod ((P-1)(Q-1))

Why public-key cryptography is a good application for reconfigurable

computers?

• computationally intensive arithmetic operations

• unconventionally long operand sizes (160-2048 bits)

• multiple algorithms, parameters,key sizes, and architectures= need for reconfiguration

19

Elliptic Curve Cryptosystems (ECC)

a family of cryptosystems, rather than a single cryptosystem = added security butneed for reconfigurationpublic key (asymmetric) cryptosystemsused for key agreement and digital signaturesimplementations must be optimized forminimum latency rather than maximumthroughput = limited speed-up fromparallel processing

20

Classes of applications

1. input/output intensive applications• bulk data encryption

(DES, IDEA, and RC5 encryption)

2. computationally intensive applications• secret-key cipher breaking based on

the exhaustive key search (DES, IDEA, RC5 breakers)

• public-key cipher breaking based on factoring

3. latency-critical applications• cipher key agreement and signature

(ECC schemes, RSA)

21

Secret key cipher libraries

1. Secret key cipher encryption and decryption

2. Secret key cipher breaking

• DES• IDEA• RC5

• DES• IDEA• RC5

22

Public key cipher libraries1. Operations in the binary Galois Fields GF(2m)

a. polynomial basisb. normal basis

2. Multiprecision integer arithmetic

3. Elliptic Curve Operations

- addition- doubling- scalar multiplication

23

Basic operations of ECCBasic operations in Galois Field GF(2m)

• addition and subtraction (xor): x+y, x-y (XOR)• multiplication, squaring: x ⋅ y, x2

• inversion: x-1

Basic operations on points of an Elliptic Curve• addition of points: P + Q• doubling a point: 2 P• projective to affine coordinate: P2A

Complex operations on points of an Elliptic Curve

24

• scalar multiplication: k ⋅ P = P + P + …+Pk times

Hierarchy of functions

High level kP

P+Q 2P projective_to_affine(P2A)

MUL

INV

Medium level

Low level 2

ROTXORLow level 1

SRC Program Partitioning

C function for µP

C function for MAP

VHDLmacro

µP system

FPGA system

HLL

HDL

26

Investigated Partitioning Schemes

27

µP Software Only

C function for µP kP

C function for FPGA

VHDLmacro

28

Based on public-domain code by Rosing M., Implementing Elliptic Curve Cryptography, Manning, 1999

0HL1 PartitioningC function for µP

0

MUL4

C functionfor FPGA

VHDL macros

29

ROTROT XOR

H

L1V_ROTVARROT

kPP2A

kP

P+Q 2P

MUL2MUL2 MULMUL

INVINV

P2AP+Q2P

0HL2 PartitioningC function for µP

0

MUL4

C functionfor FPGA

VHDL macros

ROTROT XOR

H

V_ROTINV

kPP2A

kP

P+Q 2P

MUL2MUL2

P2AP+Q2P

L2

30

0HM Partitioning0C function

for µP

31

C function for FPGA

VHDL macros

P+QP+Q 2P2P P2AP2A

kPkP H

M

00H Partitioning (VHDL only)

C function for µP 0

C function for FPGA 0

VHDLmacro kP H

32

Results

33

Timing Measurements

.c file.mc file

MAPAlloc.

MAPFree

DMA

DataOut

DMA

Data In

FPGA

Computation

End-to-End time (SW)

MAPfunction MAP function

FPGAConfigure

End-to-End time (HW)

Configuration time

MAPAllocation

time

MAP Release

Time

Results (Latency)

0HL1 866 37 472 14 394 893 1.460HL2 863 37 469 14 394 895 1.450HM 592 37 201 12 391 1305 1VHDL macro 592 39 201 17 391 1305 1

System Level Architecture

Total Overhead

(µs)

Data Transfer

Out Time (µs)

Speedup vs.

Software

Slow-down vs. VHDL macro

Software

FPGA Compu-tation Time (µs)

Data Transfer

In Time(µs)

End-to-End Time (µs)

772,519

350

100

200

300

400

500

600

700

800

900

us

0HL1 0HL2 0HM 00H (VHDL)

Different Architectures

End to End Time

FPGA ComputationTime

Results (Area)

Software N/A0HL1 99 1.68 57 1.3 68 2.610HL2 92 1.56 52 1.18 62 2.380HM 75 1.27 48 1.09 39 1.500H 59 1 44 1 26 1

System Level Architecture

% of CLB slices

(out of 33792)

CLB increase vs. pure VHDL

% of LUTs(out of 67,584)

LUT increase vs. pure VHDL

% of FFs

(out of 67,584)

FF count increase vs. pure VHDL

360

10

20

30

40

50

60

70

80

90

100

%

0HL1 0HL2 0HM 00HDifferent Architectures

CLB slicesLUTFF

Number of lines of code

Algorithm Partitioning

Scheme

VHDL Macro Wrapper

MAP C Main C

0HL1 1007 260 371

349

185

78

153

153

153

153

1291 230

1744 160

36

0HM

1960

0HL2

VHDL macro

37

Conclusions

38

Conclusions

Assuming focus on:

Timing Resources

Ease of programming

Conclusions – cont.The best implementation approach:

0HL1 partitioning scheme

MUL4

C functionfor MAP

VHDL macros

ROTROT XOR

C function for µP

0

H

L1V_ROTV_ROT

kP

INV

P2AP+Q 2P

kP

INV

P2AP+QP+Q 2P2P

MUL2MUL2 MULMUL

893 speedup vs. software and only 0.46 times slowdown versus pure VHDL with ease of

implementation

Conclusions

• Elliptic Curve Cryptosystem implementation challenging for reconfigurable computers because of

• optimization for latency rather than throughput• limited amount of parallelism

• First publication showing a 1000x speed-up for a reconfigurable computer applicationoptimized for data latency

41

General hierarchy of library files suggested

by SRC Computers Inc.

42

Structure of the macro repository

< top of repository >

< macros >

<lib # 1 > <lib # 2 > <lib # 3 >

common rev_d rev_e

macro1 macro2 macro3

rev_f

DataSheethdlfile DebugCodeFileInfoFile BlkBoxFile43

Files describing the macro

Platform independent– HDL file: macro.v or macro.vh

• Verilog or VHDL code defining the macro

– Debug Code File: macro.c • provides the equivalent C functionality for the macro

– Data sheet file: datasheet• contains the documentation for the macro

Platform dependent– Blk Box File: blackbox.v

• Interface (black box) definition for the macro in Verilog

– Info File: info• Info file entry for this macro

44

Software libraries and their role in the development

of SRC libraries

45

Roles of software libraries

1. source of test vectors for VHDL macros

2. emulation of hardware in the debug mode

3. performance comparison

46

How to approach porting your application to reconfigurable computers?

1. Identify class of applications

2. Identify basic operations required by your applications

3. Determine the existence of the RC library of such operations

4. Determine the existence of the microprocessor library of such operations

5. Determine the right granularity for the requiredlibrary operations

47