Application of Neural Networks to Power Analysis

Abstract

The purpose of this work is to research the possibility of a side-channel attack, more precisely power consumption attack (recovering the encryption key according to the board’s power consumption schedule) on the AES-128 algorithm implemented in hardware. Basically, various methods can be used to make an attack, including SPA (Simple Power Consumption Attack) and DPA (Differential Power Consumption Attack). SPA methods involve a simple visual analysis of energy consumption graphs, while DPA involves the use of statistical methods to recover the encryption key. One way to make side-channel attacks more effective is to implement machine learning methods for the described purposes.

Keywords

• cryptography
• AES
• side-channel attacks
• power analysis attack
• neural networks