Machine Learning Power Side-Channel Attack on SNOW-V

2025-12-25

AI Summary

This work evaluates the side-channel security of SNOW-V—a 5G candidate stream cipher—under power-analysis attacks. Using a STM32 platform, we capture high-resolution power traces with the ChipWhisperer framework and first apply deep learning to SNOW-V leakage assessment: specifically, a fully connected neural network (FCN) for key recovery, alongside traditional correlation power analysis (CPA) and linear discriminant analysis (LDA). Trace variability leakage analysis (TVLA) confirms statistically significant information leakage. Experimental results demonstrate that the FCN achieves efficient secret-key recovery, reducing the minimum number of traces required (MTD) by over fivefold compared to the best-performing CPA+LDA combination. This study constitutes the first empirical evidence that SNOW-V is highly vulnerable to machine learning–based power analysis, thereby establishing a novel methodology and providing concrete evidence for side-channel security evaluation of 5G cryptographic algorithms.

Technology Category

• Machine Learning: Hardware-aware ML
• Computer Vision: Adversarial Attacks & Robustness
• Search and Optimization: Evaluation and Analysis

Application Category

• Security and Privacy: Large-scale security measurements
• Responsible Web: Measurement, analysis, and circumvention of Web censorship
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

This paper demonstrates a power analysis-based Side-Channel Analysis (SCA) attack on the SNOW-V encryption algorithm, which is a 5G mobile communication security standard candidate. Implemented on an STM32 microcontroller, power traces captured with a ChipWhisperer board were analyzed, with Test Vector Leakage Assessment (TVLA) confirming exploitable leakage. Profiling attacks using Linear Discriminant Analysis (LDA) and Fully Connected Neural Networks (FCN) achieved efficient key recovery, with FCN achieving > 5X lower minimum traces to disclosure (MTD) compared to the state-of-the-art Correlational Power Analysis (CPA) assisted with LDA. The results highlight the vulnerability of SNOW-V to machine learning-based SCA and the need for robust countermeasures.

Problem

• Demonstrates power analysis attack on SNOW-V encryption
• Uses machine learning to recover keys efficiently
• Highlights vulnerability needing robust countermeasures

Innovation

• Power analysis attack on SNOW-V using machine learning
• Neural networks reduce traces needed for key recovery
• TVLA confirms exploitable leakage in microcontroller implementation