The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
Os ataques de canal lateral de aprendizagem profunda (DL-SCAs) têm sido ativamente estudados nos últimos anos. Nos DL-SCAs, redes neurais profundas (DNNs) são treinadas para prever os estados internos da operação criptográfica a partir das informações do canal lateral, como traços de energia. É importante selecionar rótulos de saída DNN adequados que expressem estados internos para DL-SCAs bem-sucedidos. Nós nos concentramos no método multi-rótulo proposto por Zhang et al. para o padrão de criptografia avançada implementado por hardware (AES). Eles usaram os traços de energia fornecidos pelo conjunto de dados públicos AES-HD e relataram revelar um único byte de chave em condições nas quais a chave alvo era a mesma usada para treinamento DNN (chave de perfil). Neste artigo, discutimos uma melhoria para revelar todos os 16 bytes da chave em condições práticas nas quais a chave de destino é diferente da chave de perfil. Preparamos AES implementado em hardware sem contramedidas SCA em ASIC para o ambiente experimental. Primeiro, nossos resultados experimentais mostram que a DNN usando multi-rótulo não aprende suficientemente o vazamento do canal lateral a partir dos traços de potência adquiridos com apenas uma chave. Em segundo lugar, relatamos que DNN usando multi-label aprende ao máximo o vazamento de canal lateral usando três tipos de chaves de perfil, e todos os 16 bytes da chave alvo são revelados com sucesso, mesmo que a chave alvo seja diferente das chaves de perfil. Finalmente, aplicamos o método proposto, DL-SCA usando multi-rótulo e três chaves de perfil contra AES implementado em hardware com contramedidas de mascaramento de caixas S rotativas (RSM). O resultado experimental mostra que todos os 16 bytes principais são revelados com sucesso usando apenas 2,000 rastreamentos de ataque. Também estudamos as razões para o alto desempenho do método proposto contra contramedidas RSM e descobrimos que as informações dos bits fracos são efetivamente exploradas.
Yuta FUKUDA
Ritsumeikan University
Kota YOSHIDA
Ritsumeikan University
Hisashi HASHIMOTO
Ritsumeikan University
Kunihiro KURODA
Ritsumeikan University
Takeshi FUJINO
Ritsumeikan University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copiar
Yuta FUKUDA, Kota YOSHIDA, Hisashi HASHIMOTO, Kunihiro KURODA, Takeshi FUJINO, "Profiling Deep Learning Side-Channel Attacks Using Multi-Label against AES Circuits with RSM Countermeasure" in IEICE TRANSACTIONS on Fundamentals,
vol. E106-A, no. 3, pp. 294-305, March 2023, doi: 10.1587/transfun.2022CIP0015.
Abstract: Deep learning side-channel attacks (DL-SCAs) have been actively studied in recent years. In the DL-SCAs, deep neural networks (DNNs) are trained to predict the internal states of the cryptographic operation from the side-channel information such as power traces. It is important to select suitable DNN output labels expressing an internal states for successful DL-SCAs. We focus on the multi-label method proposed by Zhang et al. for the hardware-implemented advanced encryption standard (AES). They used the power traces supplied from the AES-HD public dataset, and reported to reveal a single key byte on conditions in which the target key was the same as the key used for DNN training (profiling key). In this paper, we discuss an improvement for revealing all the 16 key bytes in practical conditions in which the target key is different from the profiling key. We prepare hardware-implemented AES without SCA countermeasures on ASIC for the experimental environment. First, our experimental results show that the DNN using multi-label does not learn side-channel leakage sufficiently from the power traces acquired with only one key. Second, we report that DNN using multi-label learns the most of side-channel leakage by using three kinds of profiling keys, and all the 16 target key bytes are successfully revealed even if the target key is different from the profiling keys. Finally, we applied the proposed method, DL-SCA using multi-label and three profiling keys against hardware-implemented AES with rotating S-boxes masking (RSM) countermeasures. The experimental result shows that all the 16 key bytes are successfully revealed by using only 2,000 attack traces. We also studied the reasons for the high performance of the proposed method against RSM countermeasures and found that the information from the weak bits is effectively exploited.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.2022CIP0015/_p
Copiar
@ARTICLE{e106-a_3_294,
author={Yuta FUKUDA, Kota YOSHIDA, Hisashi HASHIMOTO, Kunihiro KURODA, Takeshi FUJINO, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Profiling Deep Learning Side-Channel Attacks Using Multi-Label against AES Circuits with RSM Countermeasure},
year={2023},
volume={E106-A},
number={3},
pages={294-305},
abstract={Deep learning side-channel attacks (DL-SCAs) have been actively studied in recent years. In the DL-SCAs, deep neural networks (DNNs) are trained to predict the internal states of the cryptographic operation from the side-channel information such as power traces. It is important to select suitable DNN output labels expressing an internal states for successful DL-SCAs. We focus on the multi-label method proposed by Zhang et al. for the hardware-implemented advanced encryption standard (AES). They used the power traces supplied from the AES-HD public dataset, and reported to reveal a single key byte on conditions in which the target key was the same as the key used for DNN training (profiling key). In this paper, we discuss an improvement for revealing all the 16 key bytes in practical conditions in which the target key is different from the profiling key. We prepare hardware-implemented AES without SCA countermeasures on ASIC for the experimental environment. First, our experimental results show that the DNN using multi-label does not learn side-channel leakage sufficiently from the power traces acquired with only one key. Second, we report that DNN using multi-label learns the most of side-channel leakage by using three kinds of profiling keys, and all the 16 target key bytes are successfully revealed even if the target key is different from the profiling keys. Finally, we applied the proposed method, DL-SCA using multi-label and three profiling keys against hardware-implemented AES with rotating S-boxes masking (RSM) countermeasures. The experimental result shows that all the 16 key bytes are successfully revealed by using only 2,000 attack traces. We also studied the reasons for the high performance of the proposed method against RSM countermeasures and found that the information from the weak bits is effectively exploited.},
keywords={},
doi={10.1587/transfun.2022CIP0015},
ISSN={1745-1337},
month={March},}
Copiar
TY - JOUR
TI - Profiling Deep Learning Side-Channel Attacks Using Multi-Label against AES Circuits with RSM Countermeasure
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 294
EP - 305
AU - Yuta FUKUDA
AU - Kota YOSHIDA
AU - Hisashi HASHIMOTO
AU - Kunihiro KURODA
AU - Takeshi FUJINO
PY - 2023
DO - 10.1587/transfun.2022CIP0015
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E106-A
IS - 3
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - March 2023
AB - Deep learning side-channel attacks (DL-SCAs) have been actively studied in recent years. In the DL-SCAs, deep neural networks (DNNs) are trained to predict the internal states of the cryptographic operation from the side-channel information such as power traces. It is important to select suitable DNN output labels expressing an internal states for successful DL-SCAs. We focus on the multi-label method proposed by Zhang et al. for the hardware-implemented advanced encryption standard (AES). They used the power traces supplied from the AES-HD public dataset, and reported to reveal a single key byte on conditions in which the target key was the same as the key used for DNN training (profiling key). In this paper, we discuss an improvement for revealing all the 16 key bytes in practical conditions in which the target key is different from the profiling key. We prepare hardware-implemented AES without SCA countermeasures on ASIC for the experimental environment. First, our experimental results show that the DNN using multi-label does not learn side-channel leakage sufficiently from the power traces acquired with only one key. Second, we report that DNN using multi-label learns the most of side-channel leakage by using three kinds of profiling keys, and all the 16 target key bytes are successfully revealed even if the target key is different from the profiling keys. Finally, we applied the proposed method, DL-SCA using multi-label and three profiling keys against hardware-implemented AES with rotating S-boxes masking (RSM) countermeasures. The experimental result shows that all the 16 key bytes are successfully revealed by using only 2,000 attack traces. We also studied the reasons for the high performance of the proposed method against RSM countermeasures and found that the information from the weak bits is effectively exploited.
ER -