Transcript Team W1 1. Bobby Colyer (W11) 2. Jeffrey Kuo (W12) 3. Myron Kwai

Presentation #1: Rijndael Encryption
Team W1
1. Bobby Colyer (W11)
2. Jeffrey Kuo (W12)
3. Myron Kwai (W13)
4. Shirlene Lim (W14)
~*Team Manager: Rebecca Miller*~
Stage I:
21st January 2004
DESIGN PROPOSAL
Overall Project Objective:
Implement the new AES Rijndael algorithm on chip
18-525 Integrated Circuit Design Project
Status
 Project Chosen
 Alternatives Studied and Eliminated
 Specifications Defined
 Verilog Obtained
 Gate-Level Verilog





Cut down 128-bit design into a manageable size (32? 64?)
Test Benches
Schematic Design
Layout
Simulations
18-525 Integrated Circuit Design Project
Design Decisions
 128-bit encryption too large for the scope of this course
 Plan to step down to 32-bit
 Number of rounds (normally 10) most likely need to be
lessened
 Doing encryption only
 Leaving decryption out due to size restrictions
 Encryption + Decryption for 128-bit is 40,000 gates!
 Need to include the key schedule
 Takes up ~85% of the area
18-525 Integrated Circuit Design Project
Alternative Projects
 2D DCT
 Was considered by Zack’s group and was found to be done
 Had the powerpoint almost done before we found out…
 Other Encryption Algorithms
 Triple DES – outdated and sucks, can be cracked fairly easily now
 Serpent – have code, better than RC6, but Rijndael was chosen for a reason right?
 Histogram Equalizer
 Had MATLAB code but involved floating point statistics (FP Division!)
 FPU – Floating Point Unit
 Does Floating Point Add/Sub/Mult/Div
 Too insane – Rebecca said so
 Edge Detection
 Too simple, already done as sub-component of previous projects
 Simple MIPS Processor
 Too simple, need to add more instructions, marketing?
 Way too many more… my computer is full of them!
18-525 Integrated Circuit Design Project
Why is this the AES?!
 Fastest Encryption / Decryption in general
 No key setup time (hardcoded)
 Serpent, MARS, RC6 all require setup time
 Better than DES, Triple DES (Outdated)
 Triple DES could be cracked in linear time easily with machines from ‘90s
 25x faster on the same general-purpose hardware
 Easily extendable and scaleable (versatile)
 Block length, key length, and number of rounds
 Key and block length can be any multiple of 32 (128-bit, 192, 256…)
 Fast, simple, compact algorithm
 “Rijndael is the best mixture of simplicity, speed, and protection”
18-525 Integrated Circuit Design Project
This is why AES is needed…
“The EFF's $200,000 machine breaks DES in a few days. An aviation website gives
the cost of a B1 bomber as $200,000,000. Spending that much, an intelligence
agency could expect to break DES in an average time of six and a half
minutes.”
18-525 Integrated Circuit Design Project
Applications
 Suited for Smart cards
 Currently adapted to modern processors
 Pentium and RISC processors
 Theoretically unbreakable using modern technology
 Hacking algorithms take ~2^87 – 2^100 operations
 Today’s machines aren’t powerful enough to compute & check
 To protect digital information
 Data, voice, video, and images from attack, impersonation, or eavesdropping
 SSL Security for Internet Browsers
18-525 Integrated Circuit Design Project
Block Diagram
18-525 Integrated Circuit Design Project
Transistor Count
(Assuming 32-bit Implementation)




~256 Registers
XORs
Inverters/Buffers
SBOX
 Registers
 Key Schedule
 XORs
 Shifters (Hardcoded – Just routing wires)
 Muxes
Total:
18-525 Integrated Circuit Design Project
~3500
~1200
~500
~12000
~100
0
~8000
~25300
Problems & Questions
 Cutting code down from 128-bit to 32-bit
 Found different implementations of Rijndael
 Some had divider/multiplier, some just XORs
 Too many transistors – Pushing the limit
 Design too big – may leave out portions of design
 Parallel design = More transistors, but faster
 Need balance
 Group M2 from 2000 should not have done the DCT
18-525 Integrated Circuit Design Project