"vcd2df" -- Leveraging Data Science Insights for Hardware Security Research

TL;DR

This paper introduces tools to convert hardware design trace data into data science formats, enabling security analysis using high-level languages and exploring scalable methods for CPU vulnerability research.

Contribution

It presents novel libraries for converting VCD files into data frames and demonstrates how data science techniques can be applied to hardware security analysis.

Findings

Converted VCD files into data frames for analysis

Applied data science tools to RTL trace data

Explored scalable analysis of CPU vulnerabilities using Spark

Abstract

In this work, we hope to expand the universe of security practitioners of open-source hardware by creating a bridge from hardware design languages (HDLs) to data science languages like Python and R through novel libraries that convert VCD (value change dump) files into data frames, the expected input type of the modern data science tools. We show how insights can be derived in high-level languages from register transfer level (RTL) trace data. Additionally, we show a promising future direction in hardware security research leveraging the parallelism of Spark to study transient execution CPU vulnerabilities, and provide reproducibility researchers via GitHub and Colab.