ASCADv2

Description

ASCAD database version 2. This database contained the power consumption of a STM32 Cortex M4 microcrontroller (STM32F303RCT7) during 800.000 random AES encryptions. The AES encryptions are protected with shuffling and affine masking, and the implementation is available on https://github.com/ANSSI-FR/SecAESSTM32. The raw dataset is split into 8 files of 100.000 encryptions, and the extracted dataset contained the 800.000 preprocessed traces with additional metadata.

Warning:

Due to an acquisition issue, the last trace of each raw file (ascadv2-stm32-conso-raw-traces{1..8}.h5) and the associated metadata are set to zero. Moreover, traces 99999, 199999, 299999, 399999, and 499999 in the profiling part of the file ascadv2-extracted.h5 are also zero. These traces may introduce a bias if used during the evaluation process.

In standard usage, the attack part of the ascadv2-extracted.h5 file does not contain any null traces and can be safely used for evaluation. However, we highly suggest to remove all the null traces from the entire dataset.

Producteur
Agence Nationale de la Sécurité des Systèmes d'Information

Licence
Licence Ouverte / Open Licence
Dernière mise à jour
24 juillet 2024

Vues

0

Téléchargements

0

Qualité des métadonnées:
À améliorer(44 %)

  • Documentation des fichiers manquante

  • Fréquence de mise à jour non renseignée

  • Couverture temporelle non renseignée

  • Couverture spatiale non renseignée

Fichiers 12Réutilisations et API 0Discussions 2Ressources communautaires Informations
Votre question porte sur autre chose que ce jeu de données ? Visiter notre forum

2 discussions

SHA1 values of files to be downloaded.

Kulkarni Praveen
—
Posté le 9 juillet 2025
In the original github repo there is a mention or SHA1.txt file at bottom which no longer exists. I still have those files so I am posting here. "author_hashes_type": 'sha1', "author_hashes": { ascadv2-stm32-conso-raw-traces1.h5: "4352910b4f4dea2de872f2122bb530582a238e02", ascadv2-stm32-conso-raw-traces2.h5: "b32a239bbdb3d683833d625a109f17a55d36f24a", ascadv2-stm32-conso-raw-traces3.h5: "10c4cdc37578f8ef6bd59608345548ffd6ba3771", ascadv2-stm32-conso-raw-traces4.h5: "f0aacc9587e94532f86a26355db81e5b4ab2848b", ascadv2-stm32-conso-raw-traces5.h5: "4158149ac9a0bd7a8dda758ddb45927e7015b831", ascadv2-stm32-conso-raw-traces6.h5: "5fd51946c8552514f4c56e8730a0f7cb9ca1ac5e", ascadv2-stm32-conso-raw-traces7.h5: "ab5e6fc117b11aae733ce61c6424f7d55a487039", ascadv2-stm32-conso-raw-traces8.h5: "ef72fb2a388099ef41fbdbb0ccaefc287f048290", ascadv2-extracted.h5: "8de295c8c18d033b13c03474189c0976a49ca993", },

Metadata

DELETED DELETED
—
Posté le 9 février 2021
Could you explain the metadata & labels components like Figure 5 in "Study of Deep Learning Techniques for Side-Channel Analysis and Introduction to ASCAD Database"? I do appreciate your help in advance.