2019 Fiscal Year Annual Research Report
General and Efficient Masking Strategy for Fixed Secret Value Against Side-Channel Attacks
Project/Area Number |
19K21529
|
Allocation Type | Multi-year Fund |
Research Institution | The University of Electro-Communications |
Principal Investigator |
李 陽 電気通信大学, 大学院情報理工学研究科, 准教授 (20821812)
|
Project Period (FY) |
2019-04-01 – 2020-03-31
|
Keywords | AES / Side-Channel Attacks / Mask |
Outline of Annual Research Achievements |
In this research, we largely improved the key recovery efficiency for the simple power analysis on the masked expanded AES round keys. Compared to existing work, we improved the key recovery speed up to 95%. In a certain scenario, the key recovery success rate is improved from 47% to 96%. These improvements come from several techniques as (1) new data structure in the key recovery, (2) improved information extraction from Hamming weights, (3) full usage of all leakage traces, and (4) optimized recovery sequence of the key bytes. We further extended the improved key recovery architecture considering more variations of the leakage model and the noise model. We designed a masking scheme to achieve the highest key recovery complexity using the limited randomness. The key concept is to maximize the computational distance between the key bytes using the same mask. We summarized the relationship between the security evaluation result and the masking schemes based on the established key recovery architecture. These relationships will be formalized to a few design principles of the masking schemes. The related research result is under submission. Besides, inspired by this project, 1. we identified the unexpected information leakage of the alarm signal sent by a laser sensor that detects irregular current; 2. we investigated the possibility and the noise tolerance for the physical attack using the leakage from 5th AES round; and 3. we improved the key recovery efficiency of the template attack for recently proposed long-distance side-channel attack named screaming channel attack.
|
Research Products
(4 results)