Side channel power analysis on a target with an AES-256 encryption

Abstract

Information security often relies on strong cryptographic algorithms like Advanced Encryption Standard (AES). However, cracking AES encryption with side-channel attacks (SCA) has become a growing concern. SCA exploits fluctuations in a device's power consumption during encryption, potentially revealing crucial information about the secret key. This has highlighted the significance of further research on understanding how SCA bypasses the protections offered by AES, ultimately compromising the confidentiality of encrypted data. This project involves the implementation of AES-128, AES-192, and AES-256 on a microcontroller, and the recovery of the secret key with SCA, to further investigate the vulnerabilities of a target device utilising such cryptographic algorithms to such attacks. We implemented AES encryption on a device by importing the ‘Crypto’ library in Arduino IDE. Power traces were collected and saved by importing a specialised library supporting the connection with the oscilloscope and ‘NumPy’ in Python script, respectively. SCA and analysis were conducted using the Async2Secure SCA toolkit. We employed Correlation Power Analysis (CPA), a technique that yields correlation coefficients for identifying correlations between power consumption patterns and specific bytes of secret keys. The report details the experimental setup, data acquisition process, and analysis methods used. The results section presents the success rate of the attack in recovering the secret key. We also discuss factors influencing the attack's effectiveness, such as the number of power traces captured and the change in correlation coefficients.