Differential Power Analysis (DPA) is a powerful technique that has proven devastatingly effective against various encryption algorithms such as AES, Triple-DES and RSA. DPA has been able to successfully extract the secret key from these algorithms during the process of decrypting cyphertext information. Even in implementations with counter-measures, DPA can still be effective with additional traces, however at the cost of a corresponding increase in processing times. Hundreds of thousands to millions of traces may be necessary to defeat counter-measures; even high-end workstations can take days to process such large data sets. Graphics Processing Units (GPUs), available in most desktops as graphics cards on the other hand, have been effectively used in many parallel processing tasks. We use the GPU to accelerate DPA analysis and find a speedup of more than 200 times over workstation processing time. Using this method, we reduce computation time for a particular attack from 41 minutes to 12.24 seconds including the time for data transfer between the GPU and CPU.
[1]
Paul C. Kocher,et al.
Differential Power Analysis
,
1999,
CRYPTO.
[2]
Stefan Mangard,et al.
Power analysis attacks - revealing the secrets of smart cards
,
2007
.
[3]
장훈,et al.
[서평]「Computer Organization and Design, The Hardware/Software Interface」
,
1997
.
[4]
P. J. Narayanan,et al.
Accelerating Large Graph Algorithms on the GPU Using CUDA
,
2007,
HiPC.
[5]
V. Barnett,et al.
Applied Linear Statistical Models
,
1975
.
[6]
Hiren Patel.
GPU accelerated real time polarimetric image processing through the use of CUDA
,
2010,
Proceedings of the IEEE 2010 National Aerospace & Electronics Conference.
[7]
Paulius Micikevicius,et al.
3D finite difference computation on GPUs using CUDA
,
2009,
GPGPU-2.
[8]
Jie Cheng,et al.
CUDA by Example: An Introduction to General-Purpose GPU Programming
,
2010,
Scalable Comput. Pract. Exp..