Splitting the Interpose PUF: A Novel Modeling Attack Strategy

Nils Wisiol,Niklas Pirnay,Christopher Mühl,Phuong Ha Nguyen,Marten Van Dijk,Marian Margraf,Jean-Pierre Seifert,Ulrich Rührmair

doi:10.46586/tches.v2020.i3.97-120

Abstract

We demonstrate that the Interpose PUF proposed at CHES 2019, an Arbiter PUF-based design for so-called Strong Physical Unclonable Functions (PUFs), can be modeled by novel machine learning strategies up to very substantial sizes and complexities. Our attacks require in the most difficult cases considerable, but realistic, numbers of CRPs, while consuming only moderate computation times, ranging from few seconds to few days. The attacks build on a new divide-and-conquer approach that allows us to model the two building blocks of the Interpose PUF separately. For non-reliability based Machine Learning (ML) attacks, this eventually leads to attack times on (kup, kdown)-Interpose PUFs that are comparable to the ones against max{kup, kdown}-XOR Arbiter PUFs, refuting the original claim that Interpose PUFs could provide security similar to (kdown + kup/2)-XOR Arbiter PUFs (CHES 2019). On the technical side, our novel divide-and-conquer technique might also be useful in analyzing other designs, where XOR Arbiter PUF challenge bits are unknown to the attacker.

Highlights

1.1 Motivation and OverviewThe Interpose Physical Unclonable Functions (PUFs) is one of the most recent Strong PUF design proposals [NSJ+19]
To reflect the time unsuccessfully spent training a model, we define for chosen security parameters n, kup, kdown, target reliability, training set size N, and employed computing resources the time until first success as the expectation of time spend until a model with prediction accuracy higher than 95%, relative to the PUF’s reliability, is obtained
We point out that different instances of XOR Arbiter PUFs may differ in their resistance to modeling attacks [TB15], and t1 only refers to the average time until success, not ruling out the possibility that some instances of the given size may be harder or easier to model

Summary

Introduction

The Interpose PUF (iPUF) is one of the most recent Strong PUF design proposals [NSJ+19] It promises two distinctive and noteworthy features: Firstly, it builds on the Arbiter PUF design, and inherits the practicality and CMOS-compatibility of the latter. It contains some novel design elements, which successfully thwarted those state-of-the-art modeling attacks that were available at the time of its publication. This turned the iPUF into one of the most promising Strong PUF design proposals to date. Note that the transformed challenge x does not depend on manufacturing imperfections, and can be computed by an attacker from the physical challenge c

Methods

Results

Conclusion