GNN4IP: Graph Neural Network for Hardware Intellectual Property Piracy Detection

TL;DR

GNN4IP employs graph neural networks to effectively detect hardware IP piracy by modeling circuit designs as graphs, achieving high accuracy in identifying copied or obfuscated IPs.

Contribution

This work introduces a novel GNN-based approach for hardware IP piracy detection, surpassing traditional watermarking methods in accuracy and robustness.

Findings

96% accuracy in piracy detection

100% accuracy in recognizing obfuscated IPs

Effective modeling of hardware as graphs

Abstract

Aggressive time-to-market constraints and enormous hardware design and fabrication costs have pushed the semiconductor industry toward hardware Intellectual Properties (IP) core design. However, the globalization of the integrated circuits (IC) supply chain exposes IP providers to theft and illegal redistribution of IPs. Watermarking and fingerprinting are proposed to detect IP piracy. Nevertheless, they come with additional hardware overhead and cannot guarantee IP security as advanced attacks are reported to remove the watermark, forge, or bypass it. In this work, we propose a novel methodology, GNN4IP, to assess similarities between circuits and detect IP piracy. We model the hardware design as a graph and construct a graph neural network model to learn its behavior using the comprehensive dataset of register transfer level codes and gate-level netlists that we have gathered. GNN4IP detects IP piracy with 96% accuracy in our dataset and recognizes the original IP in its obfuscated version with 100% accuracy.