SecIC3: Customizing IC3 for Hardware Security Verification
Qinhan Tan, Akash Gaonkar, Yu-Wei Fan, Aarti Gupta, Sharad Malik · 2026
Read SecIC3: Customizing IC3 for Hardware Security Verification by Qinhan Tan, Akash Gaonkar, Yu-Wei Fan, Aarti Gupta, Sharad Malik (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-21353.txt.
Introduction
Qinhan Tan, Akash Gaonkar, Yu-Wei Fan, Aarti Gupta, and Sharad Malik adapt a classic chip-checking method for security. Formal checks can prove that a chip design keeps secrets from leaking to visible outputs. Both secrecy and integrity come down to whether information flows where it should not.
The underlying method, IC3, is a proven way to check safety properties. The authors tailor it to the flow questions that security checks ask. Off-the-shelf tools often struggle with those questions at real chip scale. A custom version could let designers prove security claims before any chip is built.
Catalog record
| CREATOR | Qinhan Tan, Akash Gaonkar, Yu-Wei Fan, Aarti Gupta, Sharad Malik |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 32 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-21353
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
SecIC3: Customizing IC3 for Hardware Security Verification Qinhan Tan, Akash Gaonkar, Yu-Wei Fan, Aarti Gupta, Sharad Malik Abstract Recent years have seen significant advances in using formal verification to check hardware security properties. Of particular practical interest are checking confidentiality and integrity of secrets, by checking that there is no information flow between the secrets and observable outputs. A standard method for checking information flow is to translate the corresponding non-interference hyperproperty into a safety property on a self-composition of the design, which has two copies of the design composed together. Although prior efforts have aimed to reduce the size of the self-composed design, there are no state-of-the-art model checkers that exploit their special structure for hardware security verification. In this paper, we propose SecIC3, a hardware model checking algorithm based on IC3 that is customized to exploit this self-composition structure. SecIC3 utilizes this structure in two complementary techniques: symmetric state exploration and adding equivalence predicates . We implement SecIC3 on top of two open-source IC3 implementations and evaluate it on a non-interference checking benchmark consisting of 10 designs. The experiment results show that SecIC3 significantly reduces the time for finding security proofs, with up to 49.3x proof speedup compared to baseline implementations. I Introduction Hardware security has received increasing attention in recent years as malicious attackers have used hardware vulnerabilities to compromise secret information. Breaches to confidentiality , where attackers steal secrets, and integrity , where attackers modify them, can result in economic loss [ 18 ] and threaten public safety [ 24 ] . One defensive approach that provides strong hardware security guarantees is formal hardware security verification , which uses formal verification techniques to systematically check a hardware implementation against a given attack model. Formal verification uniquely provides exhaustive security guarantees by constructing a rigorous security proof reasoning over all possible executions of a hardware design, and potentially identifying vulnerabilities that may have been missed by other (simulation-based) approaches, such as fuzzing [ 17 , 26 , 37 ] . However, current work in formal security verification [ 33 , 29 , 38 , 35 ] has been limited to small hardware systems and does not scale to real-world designs. In this paper, we propose to address this scaling gap by customizing the model checking algorithm commonly used for proving security properties, particularly confidentiality and integrity. Fig. 1: Non-Interference Checking via Self-Composition. In the literature [ 14 ] , both confidentiality and integrity specify the absence of information flow: for confidentiality, changing a secret value should not change any public output, and for integrity, changing a public input should not change any protected data. More formally, both are forms of a hyperproperty called non-interference , in which every pair of executions that differ only in the values of given source variables produce the same values of given sink variables. We use the terminology of confidentiality throughout the paper. Seminal work by Barthe et al. [ 4 ] showed that the non-interference hyperproperty can be verified by constructing a self-composition , as depicted in Fig. 1 . Two copies of a system are composed together, such that only the secret inputs are allowed to differ between the two copies. This simulates a pair of executions in parallel, so that one can prove non-interference by showing that corresponding outputs between the two copies are always equal — a safety property over the self-composition. To our knowledge, prior work in formal hardware security verification [ 29 , 28 ] has discharged the non-interference verification condition using generic model checking algorithms [ 5 ] , treating the self-composition as a black-box transition system without taking advantage of its symmetric structure. Beyond security verification, previous work in general model checking (e.g., [ 12 , 30 ] ) has benefited from symmetry reduction. These were based on traditional model checking algorithms, and while their techniques do not translate directly to our context, are inspirational to our work. This paper proposes SecIC3, a customization of the state-of-the-art IC3 model checking algorithm [ 5 , 11 ] to make it more effective at verifying non-interference. As part of SecIC3 we introduce two complementary techniques that improve solver performance by exploiting the special structure of the non-interference self-composition: Exploring symmetric state Since a self-composition consists of two identical copies of the same design, its space of reachable states is highly symmetric. In particular, a copy-relabeling mapping allows us to take any reachable state, swap corresponding variables of the two copies, and produce another reachable state. We augment the model checker with this relabeling mapping and use it to strengthen the blocking step of IC3 by learning additional lemmas via relabeling. Adding equivalence predicates For large designs, existing model checkers may struggle to decompose the proof objective relating corresponding outputs of the two copies into subtasks relating corresponding internal subcomponents/variables as this correspondence is not made explicit. We aid this decomposition by extending the self-composition with auxiliary equivalence predicates that indicate whether corresponding subcomponents between the two copies will output the same value. We modify the IC3 algorithm to better utilize these equivalence predicates, proposing three heuristics for determining which predicates to use during lemma learning, each with their own tradeoffs. We have implemented two prototypes of SecIC3, extending the following two open-source IC3 implementations. • ABC-PDR [ 11 ] : a classic IC3 implementation widely used as a baseline in model checking [ 15 , 13 , 9 ] ; • rIC3 [ 27 ] : the winner of Hardware Model Checking Competition 2025 on both bit-level and word-level tracks. We also create a new benchmark for non-interference checking, consisting of 10 different open-source hardware designs. Our evaluations on the benchmark show that SecIC3 can improve the performance of finding security proofs by up to 49.3x times compared to the baseline implementations. Summary of Our Contributions • We propose SecIC3, an IC3-based model checking algorithm customized for verifying non-interference. In the algorithm, we introduce two new techniques, exploring symmetric state and adding equivalence predicates , that use symmetry within a non-interference verification problem to guide lemma learning within the model checker. • We construct a benchmark suite comprising 10 diverse hardware designs specifically for evaluating non-interference checkers. • We implement and evaluate SecIC3 on two open-source IC3 implementations, demonstrating that SecIC3 achieves up to 49.3x times speedup over baseline implementations in proving non-interference for this benchmark suite. II Background: IC3/PDR IC3 [ 5 ] , also known as Property Directed Reachability (PDR) [ 11 ] is a state-of-the-art hardware model checking algorithm for unbounded proofs of safety properties. We will assume some familiarity with the IC3 algorithm, but briefly review the parts relevant to our work. A more complete exposition can be found in Een et al. [ 11 ] . Fig. 2: The frames of an IC3 inductive trace when checking a property P . IC3 refines frames by blocking bad states: the state is shown to be part of an unreachable cube c , and is thereby excluded from the latest frame F_{n}^{\prime}=F_{n}\wedge\neg c . Given a transition system (Init,Tr) with state variables x , initial states Init(x) , and transition relation Tr(x,x^{\prime}) , suppose we wish to check a safety property P(x) .