C-SAR: SAT Attack Resistant Logic Locking for RSFQ Circuits

Junyao Zhang, Paul Bogdan, Shahin NazarianarXiv 2023 · DOI · arXiv

Introduction

As the semiconductor industry pushes toward exascale computing, traditional CMOS technology faces significant power and speed limitations. Superconducting Electronics (SCE), particularly Rapid Single Flux Quantum (RSFQ) circuits, have emerged as a promising alternative, capable of operating at speeds exceeding 100 GHz with minimal power consumption. However, the unique physical characteristics of RSFQ circuits introduce new security vulnerabilities. This paper presents C-SAR (Camouflaged SAT Attack Resist), a specialized logic locking technique designed to protect the intellectual property (IP) of superconducting circuits against advanced hardware hacking methods.

The Threat: SAT Attacks on Superconducting Logic

The globalization of the semiconductor supply chain has increased the risk of IP theft and the insertion of malicious hardware trojans. Logic locking is a common defense where a circuit's functionality is "locked" behind a secret key. However, the Boolean Satisfiability (SAT) attack has become the "gold standard" for breaking these locks. By comparing a locked netlist with a functional chip, SAT solvers can efficiently prune the key space and recover the secret key in a matter of minutes.

While logic locking has been extensively studied for CMOS, RSFQ circuits present a different landscape. Their gate-level pipelined nature—where every logic gate is synchronized by a clock—makes them potentially easier to reverse-engineer but also offers unique opportunities for defense that traditional combinational logic lacks.

Proposed Solution: The C-SAR Architecture

To counter the SAT attack, the authors propose C-SAR, a multi-layered defense mechanism tailored for RSFQ. The architecture consists of two primary components: the RSAT block and camouflaged Delay Flip-Flop (DFF) arrays.

The C-SAR architecture, integrating camouflaged DFFs and dual comparators to thwart SAT attacks.

The RSAT block is designed to neutralize the efficiency of SAT solvers. In a typical attack, a single input pattern can rule out large portions of the incorrect key space. The RSAT block ensures that each input pattern can only eliminate one incorrect key at a time, forcing the attacker into a brute-force search that scales exponentially with the key size.

Complementing this is the use of Camouflaged DFFs. In RSFQ logic, DFFs are used for synchronization. By creating DFFs that are physically indistinguishable from standard wires (Josephson Transmission Lines or JTLs) to an attacker, the designers can obfuscate the circuit's timing and structure.

Exploiting Temporal Obfuscation

The core innovation of C-SAR lies in its use of "temporal obfuscation." Unlike silicon-based logic, which is largely combinational, RSFQ is inherently "gate-level pipelined."

Gate-level pipelining in RSFQ circuits, where logic is synchronized across multiple clock cycles.

By intertwining camouflaged DFFs with the RSAT block, C-SAR breaks the fundamental assumption of standard SAT attacks: that there is an instantaneous relationship between input and output. An attacker must now provide the correct input sequence over multiple clock cycles to observe the intended result. This shift from combinational to sequential complexity significantly increases the computational burden on the SAT solver.

Experimental Results

The researchers evaluated C-SAR using the ISCAS’85 and 74X-series benchmarks, simulated using JSIM (Josephson Simulator). The results demonstrate robust security:

  • Attack Immunity: Standard SAT attacks failed to break the C-SAR protected benchmarks.
  • Resilience to Specialized Attacks: Even against a "Specially designed SAT" (S-SAT) attack tailored to handle sequential logic, C-SAR outperformed existing CMOS-based defenses like SARLock. For instance, on the 74283 benchmark, C-SAR required double the clock cycles (1025 vs 512) to break compared to SARLock.
  • Manageable Overhead: While the area overhead can be significant for small circuits (up to 80%), it scales linearly. For the large-scale processors where RSFQ is most applicable, this overhead becomes increasingly manageable.

Conclusion and Future Directions

C-SAR represents a significant step forward in securing the next generation of superconducting processors. By turning the inherent pipelining of RSFQ into a security feature, the authors have created a defense that is both physically robust and mathematically difficult to crack. Future work will focus on testing C-SAR on larger, more complex superconducting architectures and investigating its resilience against physical side-channel attacks.

Summary updated

Abstract

Since the development of semiconductor technologies, exascale computing and its associated applications have required increasing degrees of efficiency. Semiconductor-transistor-based circuits (STbCs) have struggled in increasing the GHz frequency. Emerging as an alternative to STbC, the superconducting electrons (SCE) technology promises higher-speed clock frequencies at ultra-low power consumption. The rapid single flux quantum (RSFQ) circuits have a theoretical potential for three orders of magnitude reduction in power while operating at clock frequencies higher than 100 GHz. Although the security in semiconductor technology has been extensively researched and developed, the security design in the superconducting field requires field demands attention. In this paper, C-SAR is presented that aims to protect the superconducting circuit electronics from Boolean satisfiability (SAT) based attacks. The SAT attack is an attack that can break all the existing combinational logic locking techniques. C-SAR can immunize against SAT attacks by increasing the key search space and prolonging the clock cycles of attack inputs. Even in the worst case of C-SAR, in face of S-SAT a specially designed SAT attack, C-SAR can also soar the attack cost exponentially with key bits first, then linearly with the length of camouflaged DFF array. We have shown in this work that the cost of C-SAR is manageable as it only linearly increases as a function of key bits.
BibTeX
@misc{zhang2023c,
  author = {Junyao Zhang and Paul Bogdan and Shahin Nazarian},
  title = {{C-SAR: SAT Attack Resistant Logic Locking for RSFQ Circuits}},
  year = {2023},
  eprint = {2301.10216},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2301.10216},
}
Figure