A universal circuit (UC) can be thought of as a programmable circuit that can simulate any circuit up to a certain size by specifying its secret configuration bits. UCs have been incorporated into various applications, such as private function evaluation (PFE). Recently, studies have attempted to formalize the concept of semiconductor intellectual property (IP) protection in the context of UCs. This is despite the observations made in theory and practice that, in reality, the adversary may obtain additional information about the secret when executing cryptographic protocols. This paper aims to answer the question of whether UCs leak information unintentionally, which can be leveraged by the adversary to disclose the configuration bits. In this regard, we propose the first photon emission analysis against UCs relying on computer vision-based approaches. We demonstrate that the adversary can utilize a cost-effective solution to take images to be processed by off-the-shelf algorithms to extract configuration bits. We examine the efficacy of our method in two scenarios: (1) the design is small enough to be captured in a single image during the attack phase, and (2) multiple images should be captured to launch the attack by deploying a divide-and-conquer strategy. To evaluate the effectiveness of our attack, we use metrics commonly applied in side-channel analysis, namely rank and success rate. By doing so, we show that our profiled photon emission analysis achieves a success rate of 1 by employing a few templates (concretely, only 18 images were used as templates).

At S$\&$P 2023, a family of secure three-party computing protocols called Bicoptor was proposed by Zhou et al., which is used to compute non-linear functions in privacy preserving machine learning. In these protocols, two parties $P_0, P_1$ respectively hold the corresponding shares of the secret, while a third party $P_2$ acts as an assistant. The authors claimed that neither party in the Bicoptor can independently compromise the confidentiality of the input, intermediate, or output. In this paper, we point out that this claim is incorrect. The assistant $P_2$ can recover the secret in the DReLU protocol, which is the basis of Bicoptor. The restoration of its secret will result in the security of the remaining protocols in Bicoptor being compromised. Specifically, we provide two secret recovery attacks regarding the DReLU protocol. The first attack method belongs to a clever enumeration method, which is mainly due to the derivation of the modular equation about the secret and its share. The key of the second attack lies in solving the small integer root problem of a modular equation, as the lattices involved are only 3 or 4 dimensions, the LLL algorithm can effectively work. For the system settings selected by Bicoptor, our experiment shows that the desired secret in the DReLU protocol can be restored within one second on a personal computer. Therefore, when using cryptographic protocols in the field of privacy preserving machine learning, it is not only important to pay attention to design overhead, but also to be particularly careful of potential security threats.

A common strategy for constructing multivariate encryption schemes is to use a central map that is easy to invert over an extension field, along with a small number of modifications to thwart potential attacks. In this work we study the effectiveness of these modifications, by deriving estimates for the number of degree fall polynomials. After developing the necessary tools, we focus on encryption schemes using the $C^*$ and Dobbertin central maps, with the internal perturbation (ip), and $Q_+$ modifications. For these constructions we are able to accurately predict the number of degree fall polynomials produced in a Gröbner basis attack, up to and including degree five for the Dob encryption scheme and four for $C^*$. The predictions remain accurate even when fixing variables. Based on this new theory we design a novel attack on Dob, which completely recovers the secret key for the parameters suggested by its designers. Due to the generality of the presented techniques, we also believe that they are of interest to the analysis of other big field schemes.

In this paper, we introduce the first fault attack on SQIsign. By injecting a fault into the ideal generator during the commitment phase, we demonstrate a meaningful probability of inducing the generation of order $\mathcal{O}_0$. The probability is bounded by one parameter, the degree of commitment isogeny. We also show that the probability can be reasonably estimated by assuming uniform randomness of a random variable, and provide empirical evidence supporting the validity of this approximation. In addition, we identify a loop-abort vulnerability due to the iterative structure of the isogeny operation. Exploiting these vulnerabilities, we present key recovery fault attack scenarios for two versions of SQIsign---one deterministic and the other randomized. We then analyze the time complexity and the number of queries required for each attack. Finally, we discuss straightforward countermeasures that can be implemented against the attack.

In this paper we address the use of Neural Networks (NN) for the assessment of the quality and hence safety of several Random Number Generators (RNGs), focusing both on the vulnerability of classical Pseudo Random Number Generators (PRNGs), such as Linear Congruential Generators (LCGs) and the RC4 algorithm, and extending our analysis to non-conventional data sources, such as Quantum Random Number Generators (QRNGs) based on Vertical-Cavity Surface- Emitting Laser (VCSEL). Among the results found, we identified a sort of classification of generators under different degrees of susceptibility, underlining the fundamental role of design decisions in enhancing the safety of PRNGs. The influence of network architecture design and associated hyper-parameters variations was also explored, highlighting the effectiveness of longer sequence lengths and convolutional neural networks in enhancing the discrimination of PRNGs against other RNGs. Moreover, in the prediction domain, the proposed model is able to deftly distinguish the raw data of our QRNG from truly random ones, exhibiting a cross-entropy error of 0.52 on the test data-set used. All these findings reveal the potential of NNs to enhance the security of RNGs, while highlighting the robustness of certain QRNGs, in particular the VCSEL-based variants, for high-quality random number generation applications.

In symmetric cryptography, vectorial Boolean functions over finite fields F2n derive strong S-boxes. To this end, the S-box should satisfy a list of tests to resist existing attacks, such as the differential, linear, boomerang, and variants. Several tables are employed to measure an S- box’s resistance, such as the difference distribution table (DDT) and the boomerang connectivity table (BCT). Following the boomerang attacks recently revisited in terms of the boomerang switch effect, with a lustra- tion highlighting the power of this technique, a tool called the Boomerang Difference Table (BDT), an alternative to the classical Boomerang BCT, was introduced. Next, two novel tables have been introduced, namely, the Upper Boomerang Connectivity Table (UBCT) and the Lower Boomerang Connectivity Table (LBCT), which are considered improvements over BCT while allowing systematic evaluation of boomerangs to return over mul- tiple rounds. This paper focuses on the new tools for measuring the revisited version of boomerang attacks and the related tables UBCT, LBCT, as well as the so-called Extended Boomerang Connectivity Table (EBCT). Specifically, we shall study the properties of these novel tools and investigate the corresponding tables. We also study their interconnections, their links to the DDT, and their values for affine equivalent vectorial functions and compositional inverses of permutations of F2n . Moreover, we introduce the concept of the nontrivial boomerang connectivity uniformity and determine the explicit values of all the entries of the EBCT, LBCT, and EBCT for the important cryptographic case of the inverse function.

At Asiacrypt 2021, Baksi et al. introduced DEFAULT, the first block cipher designed to resist differential fault attacks (DFA) at the algorithm level, boasting of a 64-bit DFA security. The cipher initially employed a straightforward key schedule, where a single key was XORed in all rounds, and the key schedule was updated by incorporating round-independent keys in a rotating fashion. However, during Eurocrypt 2022, Nageler et al. presented a DFA attack that exposed vulnerabilities in the claimed DFA security of DEFAULT, reducing it by up to 20 bits in the case of the simple key schedule and even allowing for unique key recovery in the presence of rotating keys. In this work, we have significantly improved upon the existing differential fault attack (DFA) on the DEFAULT cipher. Our enhanced attack allows us to effectively recover the encryption key with minimal faults. We have accomplished this by computing deterministic differential trails for up to five rounds, injecting around 5 faults into the simple key schedule for key recovery, recovering equivalent keys with just 36 faults in the DEFAULT-LAYER, and introducing a generic DFA approach suitable for round-independent keys within the DEFAULT cipher. These results represent the most efficient key recovery achieved for the DEFAULT cipher under DFA attacks. Additionally, we have introduced a novel fault attack called the Statistical-Differential Fault Attack (SDFA), specifically tailored for linear-structured SBOX-based ciphers like DEFAULT. This novel technique has been successfully applied to BAKSHEESH, resulting in a nearly unique key recovery. Our findings emphasize the vulnerabilities present in linear-structured SBOX-based ciphers, including both DEFAULT and BAKSHEESH, and underscore the challenges in establishing robust DFA protection for such cipher designs. In summary, our research highlights the significant risks associated with designing linear-structured SBOX-based block ciphers with the aim of achieving cipher-level DFA protection.

Structured Encryption (StE) enables a client to securely store and query data stored on an untrusted server. Recent constructions of StE have moved beyond basic queries, and now support large subsets of SQL. However, the security of these constructions is poorly understood, and no systematic analysis has been performed. We address this by providing the first leakage-abuse attacks against StE for SQL schemes. Our attacks can be run by a passive adversary on a server with access to some information about the distribution of underlying data, a common model in prior work. They achieve partial query recovery against select operations and partial plaintext recovery against join operations. We prove the optimality and near-optimality of two new attacks, in a Bayesian inference framework. We complement our theoretical results with an empirical investigation testing the performance of our attacks against real-world data and show they can successfully recover a substantial proportion of queries and plaintexts. In addition to our new attacks, we provide proofs showing that the conditional optimality of a previously proposed leakage-abuse attack and that inference against join operations is NP-hard in general.

Side-channel analysis is an important part of the security evaluations of hardware components and more specifically of those that include cryptographic algorithms. Profiling attacks are among the most powerful attacks as they assume the attacker has access to a clone device of the one under attack. Using the clone device allows the attacker to make a profile of physical leakages linked to the execution of algorithms. This work focuses on the characteristics of this profile and the information that can be extracted from its application to the attack traces. More specifically, looking at unsuccessful attacks, it shows that by carefully ordering the attack traces used and limiting their number, better results can be achieved with the same profile. Using this method allows us to consider the classical attack method, i.e. where the traces are randomly ordered, as the worst case scenario. The best case scenario is when the attacker is able to successfully order and select the best attack traces. A method for identifying efficient ordering when using deep learning models as profiles is also provided. A new loss function "Scoring loss" is dedicated to training machine learning models that give a score to the attack prediction and the score can be used to order the predictions.

Fault attacks that exploit the propagation of effective/ineffective faults present a richer attack surface than Differential Fault Attacks, in the sense that the adversary depends on a single bit of information to eventually leak secret cryptographic material. In the recent past, a number of propagation-based fault attacks on Lattice-based Key Encapsulation Mechanisms have been proposed; many of which have no known countermeasures. In this work, we propose an orthogonal countermeasure principle that does not follow adhoc strategies (like shuffling operations on secret coefficients), but rather depends on cryptographically-backed guarantees to provide quantifiable defence against aforementioned fault attacks. Concretely, we propose a framework that uses rejection sampling (which has been traditionally used as alternatives to trapdoors) to convert otherwise deterministic algorithms to probabilistic ones. Our specific goals allow careful selection of distributions such that our framework functions with a constant number of retries (around $2-3$) for unfaulted executions. In other words, should a fault be injected, the probability of success is negligible; for correct execution however, the probability of success is overwhelmingly high. Using our framework, we hence enable probabilistic decryptions in Kyber, NewHope, and Masked Kyber, and completely cut-off fault propagation in known attacks on these constructions, allowing a sound defence against known fault attacks in literature.

FUTURE is a recently proposed lightweight block cipher that achieved a remarkable hardware performance due to careful design decisions. FUTURE is an Advanced Encryption Standard (AES)-like Substitution-Permutation Network (SPN) with 10 rounds, whose round function consists of four components, i.e., SubCell, MixColumn, ShiftRow and AddRoundKey. Unlike AES, it is a 64-bit-size block cipher with a 128-bit secret key, and the state can be arranged into 16 cells. Therefore, the operations of FUTURE including its S-box is defined over $\mathbb{F}_2^4$. The previous studies have shown that the integral properties of 4-bit S-boxes are usually weaker than larger-size S-boxes, thus the number of rounds of FUTURE, i.e., 10 rounds only, might be too aggressive to provide enough resistance against integral cryptanalysis. In this paper, we mount the integral cryptanalysis on FUTURE. With state-of-the-art detection techniques, we identify several integral distinguishers of 7 rounds of FUTURE. By extending this 7-round distinguisher by 3 forward rounds, we manage to recover all the 128 bits secret keys from the full FUTURE cipher without the full codebook for the first time. To further achieve better time complexity, we also present a key recovery attack on full FUTURE with full codebook. Both attacks have better time complexity than existing results.

As the National Institute of Standards and Technology (NIST) concludes its post-quantum cryptography (PQC) competition, the winning algorithm, Dilithium, enters the deployment phase in 2024. This phase underscores the importance of conducting thorough practical security evaluations. Our study offers an in-depth side-channel analysis of Dilithium, showcasing the ability to recover the complete private key, ${s}_1$, within ten minutes using just two signatures and achieving a 60% success rate with a single signature. We focus on analyzing the polynomial addition in Dilithium, $z=y+{cs}_1$, by breaking down the attack into two main phases: the recovery of $y$ and ${cs}_1$ through side-channel attacks, followed by the resolution of a system of error-prone equations related to ${cs}_1$. Employing Linear Regression-based profiled attacks enables the successful recovery of the full $y$ value with a 40% success rate without the necessity for initial filtering. The extraction of ${cs}_1$ is further improved using a CNN model, which boasts an average success rate of 75%. A significant innovation of our research is the development of a constrained optimization-based residual analysis technique. This method efficiently recovers ${s}_1$ from a large set of error-containing equations concerning ${cs}_1$, proving effective even when only 10% of the equations are accurate. We conduct a practical attack on the Dilithium2 implementation on an STM32F4 platform, demonstrating that typically two signatures are sufficient for complete private key recovery, with a single signature sufficing in optimal conditions. Using a general-purpose PC, the full private key can be reconstructed in ten minutes.

Nextcloud is a leading cloud storage platform with more than 20 million users. Nextcloud offers an end-to-end encryption (E2EE) feature that is claimed to be able “to keep extremely sensitive data fully secure even in case of a full server breach”. They also claim that the Nextcloud server “has Zero Knowledge, that is, never has access to any of the data or keys in unencrypted form”. This is achieved by having encryption and decryption operations that are done using file keys that are only available to Nextcloud clients, with those file keys being protected by a key hierarchy that ultimately relies on long passphrases known exclusively to the users. We provide the first detailed documentation and security analysis of Nextcloud's E2EE feature. Nextcloud's strong security claims motivate conducting the analysis in the setting where the server itself is considered malicious. We present three distinct attacks against the E2EE security guarantees in this setting. Each one enables the confidentiality and integrity of all user files to be compromised. All three attacks are fully practical and we have built proof-of-concept implementations for each. The vulnerabilities make it trivial for a malicious Nextcloud server to access and manipulate users' data. We have responsibly disclosed the three vulnerabilities to Nextcloud. The second and third vulnerabilities have been remediated. The first was addressed by temporarily disabling file sharing from the E2EE feature until a redesign of the feature can be made. We reflect on broader lessons that can be learned for designers of E2EE systems.

We present staircase attack, the first attack on the incentive mechanism of the Proof-of-Stake (PoS) protocol used in Ethereum 2.0 beacon chain. Our attack targets the penalty of the incentive mechanism that penalizes inactive participation. Our attack can make honest validators suffer from penalties, even if they strictly follow the specification of the protocol. We show both theoretically and experimentally that if the adversary controls 29.6% stake in a moderate-size system, the attack can be launched continuously, so eventually all honest validators will lose their incentives. In contrast, the adversarial validators can still receive incentives, and the stake owned by the adversary can eventually exceed the $1/3$ threshold (system assumption), posing a threat to the security properties of the system. In practice, the attack feasibility is directly related to two parameters: the number of validators and the parameter MAX_ATTESTATION, the maximum number of attestations (i.e., votes) that can be included in each block. We further modify our attack such that, with current system setup (850,000 validators and MAX_ATTESTATION=128), our attack can be launched continuously with a probability of 80.25%. As a result, the incentives any honest validator receives are only 28.9% of its fair share.

Cache side-channels are a major threat to cryptographic implementations, particularly block ciphers. Traditional manual hardening methods transform block ciphers into Boolean circuits, a practice refined since the late 90s. The only existing automatic approach based on Boolean circuits achieves security but suffers from performance issues. This paper examines the use of Lookup Tables (LUTs) for automatic hardening of block ciphers against cache side-channel attacks. We present a novel method combining LUT-based synthesis with quantitative static analysis in our HyCaMi framework. Applied to seven block cipher implementations, HyCaMi shows significant improvement in efficiency, being 9.5$\times$ more efficient than previous methods, while effectively protecting against cache side-channel attacks. Additionally, for the first time, we explore balancing speed with security by adjusting LUT sizes, providing faster performance with slightly reduced leakage guarantees, suitable for scenarios where absolute security and speed must be balanced.

We introduce a novel template attack for secret key recovery in Kyber, leveraging side-channel information from polynomial multiplication during decapsulation. Conceptually, our attack exploits that Kyber's incomplete number-theoretic transform (NTT) causes each secret coefficient to be used multiple times, unlike when performing a complete NTT. Our attack is a single trace \emph{known} ciphertext attack that avoids machine-learning techniques and instead relies on correlation-matching only. Additionally, our template generation method is very simple and easy to replicate, and we describe different attack strategies, varying on the number of templates required. Moreover, our attack applies to both masked implementations as well as designs with multiplication shuffling. We demonstrate its effectiveness by targeting a masked implementation from the \emph{mkm4} repository. We initially perform simulations in the noisy Hamming-Weight model and achieve high success rates with just $13\,316$ templates while tolerating noise values up to $\sigma=0.3$. In a practical setup, we measure power consumption and notice that our attack falls short of expectations. However, we introduce an extension inspired by known online template attacks, enabling us to recover $128$ coefficient pairs from a single polynomial multiplication. Our results provide evidence that the incomplete NTT, which is used in Kyber-768 and similar schemes, introduces an additional side-channel weakness worth further exploration.

As a typical representative of the public key cryptosystem, RSA has attracted a great deal of cryptanalysis since its invention, among which a famous attack is the small private exponent attack. It is well-known that the best theoretical upper bound for the private exponent d that can be attacked is d ≤ N^0.292 , where N is a RSA modulus. However, this bound may not be achieved in practical attacks since the lattice constructed by Coppersmith method may have a large enough dimension and the lattice-based reduction algorithms cannot work so well in both efficiency and quality. In this paper, we propose a new practical attack based on the binary search for the most significant bits (MSBs) of prime divisors of N and the Herrmann-May’s attack in 2010. The idea of binary search is inspired by the discovery of phenomena called “multivalued-continuous phenomena”, which can significantly accelerate our attack. Together with several carefully selected parameters according to our exact and effective numerical estimations, we can improve the upper bound of d that can be practically achieved. We believe our method can provide some inspiration to practical attacks on RSA with mainstream-size moduli.

We show how fixed-unitary quantum encryption schemes can be attacked in a black-box setting. We use an efficient technique to invert a unitary transformation on a quantum computer to retrieve an encrypted secret quantum state $\ket{\psi}$. This attack has a success rate of 100% and can be executed in constant time. We name a vulnerable scheme and suggest how to improve it to invalidate this attack. The proposed attack highlights the importance of carefully designing quantum encryption schemes to ensure their security against quantum adversaries, even in a black-box setting.

The guess and determine attack is a common method in cryptanalysis. Its idea is to firstly find some variables which can deduced all remaining variables in a cipher and then traverse all values of these variables to find a solution. People usually utilize the exhausted search to find these variables. However, it is not applicable any more when the number of variables is a bit large. In this work we propose a guess and determine analysis based on set split to find as few variables as possible in the first step of guess and determine attack, which is a kind of exhausted search based on trading space for time and is more effective than the latter. Firstly we give an idea of set split in detail by introducing some conceptions such as base set, likely solution region and so on. And then we discuss how to utilize the set split to achieve a guess and determine analysis and give its specific implementation scheme. Finally, comparing it with the other two guess and determine analysis based on the exhausted search and the MILP method, we illustrate the effectiveness of our method by two ciphers Snow 2.0 and Enocoro-128v2. Our method spends about 0.000103 seconds finding a best solution of 9 variables for the former and 0.13 seconds finding a best solution of 18 variables for the latter in a personal Macbook respectively, which are better than those of both the exhausted search and the MILP method.

In the context of masking, which is the dominant technique for protecting cryptographic hardware designs against SCA attacks, the focus has long been on the design of masking schemes that guarantee provable security in the presence of glitches. Unfortunately, achieving this comes at the cost of increased latency, since registers are required to stop glitch propagation. Previous work has attempted to reduce latency by eliminating registers, but the exponential increase in area makes such approaches impractical. Some relatively new attempts have used DRP logic styles to avoid glitches in algorithmically masked circuits. Promising approaches in this area include LMDPL and SESYM, presented at CHES 2020 and CHES 2022 respectively. Both schemes allow masking of arbitrary functions with only one cycle latency. However, even if glitches no longer occur, there are other physical defaults that may violate the security of a glitch-free masked circuit. The imbalanced delay of dual rails is a known security problem for DRP logic styles such as WDDL, but is not covered by the known security models, e.g., robust probing model. In this work, we illustrate that imbalanced signal delays pose a threat to the security of algorithmically masked circuits implemented with DRP logic, both in theory and practice. Notably, we underscore the security of LMDPL even when delays are taken into account, contrasting with the vulnerability observed in SESYM under similar conditions. Consequently, our findings highlight the critical importance of addressing imbalanced delays in the design of masked circuits using DRP logic. In particular, our findings motivate the need for an appropriate security model, and imply that relying solely on the probing security model and avoiding glitches may be insufficient to construct secure circuits.

Grover’s algorithm is a very versatile cryptanalytical tool. Even though it doesn’t provide an exponential speed-up, it still changed the cryptographic requirements all over the world. Usually, Grover’s algorithm is executed with a fixed well-defined function indicating good states. In this paper, we want to investigate what happens if the function is changed over time to mark less and less good states. We compute the amplitudes after $2^{s/2}$ steps of an adjusted Grover’s algorithm proposed by Zheng et al. in Nested Quantum Search Model on Symmetric Ciphers and Its Applications (2023). We use the amplitudes to reason that such an approach always leads to a worse run-time when compared to the naïve version. We also indicate at which point in Zheng et al. the counterintuitive nature of quantum computation leads to false assumptions.

Given two elliptic curves and the degree of an isogeny between them, finding the isogeny is believed to be a difficult problem---upon which rests the security of nearly any isogeny-based scheme. If, however, to the data above we add information about the behavior of the isogeny on a large enough subgroup, the problem can become easy, as recent cryptanalyses on SIDH have shown. Between the restriction of the isogeny to a full $N$-torsion subgroup and no ''torsion information'' at all lies a spectrum of interesting intermediate problems, raising the question of how easy or hard each of them is. Here we explore modular isogeny problems where the torsion information is masked by the action of a group of $2\times 2$ matrices. We give reductions between these problems, classify them by their difficulty, and link them to security assumptions found in the literature.

In this paper, we provide new quantum cryptanalysis results on 5 rounds (balanced) Feistel schemes and on Benes schemes. More precisely, we give an attack on 5 rounds Feistel schemes in $\Theta(2^{2n/3})$ quantum complexity and an attack on Benes schemes in $\Theta(2^{2n/3})$ quantum complexity, where n is the number of bits of the internel random functions.

In this paper, we describe new quantum generic attacks on 6 rounds balanced Feistel networks with internal functions or internal permutations. In order to obtain our new quantum attacks, we revisit a result of Childs and Eisenberg that extends Ambainis' collision finding algorithm to the subset finding problem. In more details, we continue their work by carefully analyzing the time complexity of their algorithm. We also use four points structures attacks instead of two points structures attacks that leads to a complexity of $\mathcal{O}(2^{8n/5})$ instead of $\mathcal{O}(2^{2n})$. Moreover, we have also found a classical (i.e. non quantum) improved attack on $6$ rounds with internal permutations. The complexity here will be in $\mathcal{O}(2^{2n})$ instead of $\mathcal{O}(2^{3n})$ previously known.

A recent work by Guo, Johansson, and Nguyen (Eprint'23) proposes a promising adaptation of Sieving techniques from lattices to codes, in particular, by claiming concrete cryptanalytic improvements on various schemes. The core of their algorithm reduces to a Near Neighbor Search (NNS) problem, for which they devise an ad-hoc approach. In this work, we aim for a better theoretical understanding of this approach. First, we provide an asymptotic analysis which is not present in the original paper. Second, we propose a more systematic use of well-established NNS machinery, known as Locality Sensitive Hashing and Filtering (LSH/F). LSH/F is an approach that has been applied very successfully in the case of sieving over lattices. We thus establish the first baseline for the sieving approach with a decoding complexity of $2^{0.117n}$ for the conventional worst parameters (full distance decoding, where complexity is maximized over all code rates). Our cumulative improvements eventually enable us to lower the hardest parameter decoding complexity for SievingISD algorithms to $2^{0.101n}$. This approach outperforms the BJMM algorithm (Eurocrypt'12) but falls behind the most advanced conventional ISD approach by Both and May (PQCrypto'18). As for lattices, we found the Random-Spherical-Code-Product (RPC) to give the best asymptotic complexity. Moreover, we also consider an alternative that seems specific to the Hamming Sphere, which we believe could be of practical interest as it plausibly hides less sub-exponential overheads than RPC.

This paper presents practicable single trace attacks against the Hamming Quasi-Cyclic (HQC) Key Encapsulation Mechanism. These attacks are the first Soft Analytical Side-Channel Attacks (SASCA) against code-based cryptography. We mount SASCA based on Belief Propagation (BP) on several steps of HQC's decapsulation process. Firstly, we target the Reed-Solomon (RS) decoder involved in the HQC publicly known code. We perform simulated attacks under Hamming weight leakage model, and reach excellent accuracies (superior to $0.9$) up to a high noise level ($\sigma = 3$), thanks to a re-decoding strategy. In a real case attack scenario, on a STM32F407, this attack leads to a perfect success rate. Secondly, we conduct an analogous attack against the RS encoder used during the re-encryption step required by the Fujisaki-Okamoto-like transform. Both in simulation and practical instances, results are satisfactory and this attack represents a threat to the security of HQC. Finally, we analyze the strength of countermeasures based on masking and shuffling strategies. In line with previous SASCA literature targeting Kyber, we show that masking HQC is a limited countermeasure against BP attacks, as well as shuffling countermeasures adapted from Kyber. We evaluate the ``full shuffling'' strategy which thwarts our attack by introducing sufficient combinatorial complexity. Eventually, we highlight the difficulty of protecting the current RS encoder with a shuffling strategy. A possible countermeasure would be to consider another encoding algorithm for the scheme to support a full shuffling. Since the encoding subroutine is only a small part of the implementation, it would come at a small cost.

In this study, we investigate the newly developed low energy lightweight block cipher (LELBC), specifically designed for resource-constrained Internet of Things (IoT) devices in smart agriculture. The designers conducted a preliminary differential cryptanalysis of LELBC through mixed-integer linear programming (MILP). This paper further delves into LELBC’s differential characteristics in both single and related-key frameworks using MILP, identifying a nine-round differential characteristic with a probability of $2^{-60}$ in a single-key framework and a 12-round differential characteristic with a probability of $2^{-60}$ in a related-key framework.

Side-channel analysis has become a cornerstone of modern hardware security evaluation for cryptographic accelerators. Recently, these techniques are also being applied in fields such as AI and Machine Learning to investigate possible threats. Security evaluations are reliant on standard test setups including commercial and open-source evaluation boards such as, SASEBO/SAKURA and ChipWhisperer. However, with shrinking design footprints and overlapping tasks on the same platforms, the quality of the side channel information as well as the speed of data capture can significantly influence security assessment. In this work, we designed EFFLUX-F2, a hardware security evaluation board to improve the quality and speed of side-channel information capture. We also designed a measurement setup to benchmark the signal differences between target boards. Multiple experimental evaluations like noise analysis, CPA and TVLA performed on EFFLUX-F2 and competing evaluation boards showcase the significant superiority of our design in all aspects.

In this paper, we present an efficient attack against ${\tt PROV}$, a recent variant of the popular Unbalanced Oil and Vinegar (${\tt UOV}$) multivariate signature scheme, that has been submitted to the ongoing ${\tt NIST}$ standardization process for additional post-quantum signature schemes. A notable feature of ${\tt PROV}$ is its proof of security, namely, existential unforgeability under a chosen-message attack (${\tt EUF-CMA}$), assuming the hardness of solving the system formed by the public-key non-linear equations. We present a polynomial-time key-recovery attack against the first specification of ${\tt PROV}$ (v$1.0$). To do so, we remark that a small fraction of the ${\tt PROV}$ secret-key is leaked during the signature process. Adapting and extending previous works on basic ${\tt UOV}$, we show that the entire secret-key can be then recovered from such a small fraction in polynomial-time. This leads to an efficient attack against ${\tt PROV}$ that we validated in practice. For all the security parameters suggested in by the authors of ${\tt PROV}$, our attack recovers the secret-key in at most $8$ seconds. We conclude the paper by discussing the apparent mismatch between such a practical attack and the theoretical security claimed by ${\tt PROV}$ designers. Our attack is not structural but exploits that the current specification of ${\tt PROV}$ differs from the required security model. A simple countermeasure makes ${\tt PROV}$ immune against the attack presented here and led the designers to update the specification of ${\tt PROV}$ (v$1.1$).

Since the seminal works by Sasaki and Aoki, Meet-in-the-Middle (MITM) attacks are recognized as an effective technique for preimage and collision attacks on hash functions. At Eurocrypt 2021, Bao et al. automated MITM attacks on AES-like hashing and improved upon the best manual result. The attack framework has been furnished by subsequent works, yet far from complete. This paper elucidates three key contributions dedicated in further generalizing the idea of MITM and refining the automatic model on AES-like hashing. (1) We introduce S-box linearization to MITM pseudo-preimage attacks on AES-like hashing. The technique suits perfectly with superposition states to preserve information after S-box with an affordable cost. (2) We propose distributed initial structures, an extension on the original concept of initial states, that selects initial degrees of freedom in a more versatile manner to enlarge the search space. (3) We exploit the structural similarities between encryption and key schedule in constructions (e.g. Whirlpool and Streebog) to model propagations more accurately and avoid repeated costs. Weaponed with these innovative techniques, we further empower the MITM framework and improve the attack results on AES-like designs for preimage and collision. We obtain the first preimage attacks on 10-round AES-192, 10-round Rijndael-192/256, and 7.75-round Whirlpool, reduced time and/or memory complexities for preimage attacks on 5-, 6-round Whirlpool and 7.5-, 8.5-round Streebog, as well as improved collision attacks on 6- and 6.5-round Whirlpool.

The supersingular Endomorphism Ring problem is the following: given a supersingular elliptic curve, compute all of its endomorphisms. The presumed hardness of this problem is foundational for isogeny-based cryptography. The One Endomorphism problem only asks to find a single non-scalar endomorphism. We prove that these two problems are equivalent, under probabilistic polynomial time reductions. We prove a number of consequences. First, assuming the hardness of the endomorphism ring problem, the Charles–Goren–Lauter hash function is collision resistant, and the SQIsign identification protocol is sound for uniformly random keys. Second, the endomorphism ring problem is equivalent to the problem of computing arbitrary isogenies between supersingular elliptic curves, a result previously known only for isogenies of smooth degree. Third, there exists an unconditional probabilistic algorithm to solve the endomorphism ring problem in time $\tilde O(p^{1/2})$, a result that previously required to assume the generalized Riemann hypothesis. To prove our main result, we introduce a flexible framework for the study of isogeny graphs with additional information. We prove a general and easy-to-use rapid mixing theorem.

The Alternating Trilinear Form Equivalence (ATFE) problem was recently used by Tang et al. as a hardness assumption in the design of a Fiat-Shamir digital signature scheme ALTEQ. The scheme was submitted to the additional round for digital signatures of the NIST standardization process for post-quantum cryptography. ATFE is a hard equivalence problem known to be in the class of equivalence problems that includes, for instance, the Tensor Isomorphism (TI), Quadratic Maps Linear Equivalence (QMLE) and the Matrix Code Equivalence (MCE) problems. Due to the increased cryptographic interest, the understanding of its practical hardness has also increased in the last couple of years. Currently, there are several combinatorial and algebraic algorithms for solving it, the best of which is a graph-theoretic algorithm that also includes an algebraic subroutine. In this paper, we take a purely algebraic approach to the ATFE problem, but we use a coding theory perspective to model the problem. This modelling was introduced earlier for the MCE problem. Using it, we improve the cost of algebraic attacks against ATFE compared to previously known ones. Taking into account the algebraic structure of alternating trilinear forms, we show that the obtained system has less variables but also less equations than for MCE and gives rise to structural degree-3 syzygies. Under the assumption that outside of these syzygies the system behaves semi-regularly, we provide a concrete, non-asymptotic complexity estimate of the performance of our algebraic attack. Our results show that the complexity of our attack is below the estimated security levels of ALTEQ by more than 20 bits for NIST level I (and more for the others), thus the scheme requires re-parametrization for all three NIST security levels.

Quantum computers at some future date might be able to factor large numbers, and this poses a threat to some public key and key exchange systems in use today. This overview of recent progress in devising quantum algorithms and building quantum computing devices is meant to help technologists understand the difficult problems that quantum engineers are working on, where advances have been made, and how those things affect estimates of if and when large scale quantum computation might happen.

The DECT Standard Cipher (DSC) is a proprietary stream cipher used for encryption in the Digital Enhanced Cordless Telecommunications (DECT), which is a standard for short range cordless communication and widely deployed worldwide both in residential and enterprise environments. New weaknesses of the DSC stream cipher which are not discovered in previous works are explored and analyzed in this paper. Based on these weaknesses, new practical key recovery attacks and distinguishing attack on DSC with lower time cost are proposed. The first cryptanalytic result show that DSC can be broken in about 13.12 seconds in the known IV setting, when an offline phase that takes about 58.33 minutes is completed. After then, a distinguishing attack on DSC in the related key chosen IV setting is given, which has a time complexity of only 2 encryptions and a success probability of almost 1. Finally, based on the slide property, a key recovery attack on DSC with practical complexities is proposed. The experimental result shows that DSC can be broken on a common PC within about 44.97 seconds in the multiple related key setting. The attacks on DSC proposed in this paper clearly show that a well-designed initialization is absolutely necessary to design a secure stream cipher.

The Implicit Factorization Problem (IFP) was first introduced by May and Ritzenhofen at PKC'09, which concerns the factorization of two RSA moduli $N_1=p_1q_1$ and $N_2=p_2q_2$, where $p_1$ and $p_2$ share a certain consecutive number of least significant bits. Since its introduction, many different variants of IFP have been considered, such as the cases where $p_1$ and $p_2$ share most significant bits or middle bits at the same positions. In this paper, we consider a more generalized case of IFP, in which the shared consecutive bits can be located at $any$ positions in each prime, not necessarily required to be located at the same positions as before. We propose a lattice-based algorithm to solve this problem under specific conditions, and also provide some experimental results to verify our analysis.

Before third-party evaluation and certification, manufacturers often conduct internal security evaluations on secure hardware devices, including fault injection (FI). Within this process, FI aims to identify parameter combinations that reveal device vulnerabilities. The impracticality of conducting an exhaustive search over FI parameters has prompted the development of advanced and guided algorithms. However, these proposed methods often focus on a specific, critical region, which is beneficial for attack scenarios requiring a single optimal FI parameter combination. In this work, we introduce two novel metrics that align better with the goal of identifying multiple optima. These metrics consider the number of unique vulnerable locations and clusters (regions). Furthermore, we present two methods promoting diversity in tested parameter combinations - Grid Memetic Algorithm (GridMA) and Evolution Strategy (ES). Our findings reveal that these diversity methods, though identifying fewer vulnerabilities overall than the Memetic Algorithm (MA), still outperform Random Search (RS), identifying at least $\approx8\times$ more vulnerabilities. Using our novel metrics, we observe that the number of distinct vulnerable locations is similar across all three evolutionary algorithms, with $\approx30\%$ increase over RS. Importantly, ES and GridMA prove superior in discovering multiple vulnerable regions, with ES identifying $\approx55\%$ more clusters than the worst-performing MA.

Ascon, a family of algorithms that supports authenticated encryption and hashing, has been selected as the new standard for lightweight cryptography in the NIST Lightweight Cryptography Project. Ascon’s permutation and authenticated encryption have been actively analyzed, but there are relatively few analyses on the hashing. In this paper, we concentrate on preimage attacks on Ascon-Xof. We focus on linearizing the polynomials leaked by the hash value to find its inverse. In an attack on 2-round Ascon-Xof, we carefully construct the set of guess bits using a greedy algorithm in the context of guess-and-determine. This allows us to attack Ascon-Xof more efficiently than the method in Dobraunig et al., and we fully implement our attack to demonstrate its effectiveness. We also provide the number of guess bits required to linearize one output bit after 3- and 4-round Ascon’s permutation, respectively. In particular, for the first time, we connect the result for 3-round Ascon to a preimage attack on Ascon-Xof with a 64-bit output. Our attacks primarily focus on analyzing weakened versions of Ascon-Xof, where the weakening involves setting all the IV values to 0 and omitting the round constants. Although our attacks do not compromise the security of the full Ascon-Xof, they provide new insights into their security.

We devise algorithms for finding equivalences of trilinear forms over finite fields modulo linear group actions. Our focus is on two problems under this umbrella, Matrix Code Equivalence (MCE) and Alternating Trilinear Form Equivalence (ATFE), since their hardness is the foundation of the NIST round-$1$ signature candidates MEDS and ALTEQ respectively. We present new algorithms for MCE and ATFE, which are further developments of the algorithms for polynomial isomorphism and alternating trilinear form equivalence, in particular by Bouillaguet, Fouque, and Véber (Eurocrypt 2013), and Beullens (Crypto 2023). Key ingredients in these algorithms are new easy-to-compute distinguishing invariants under the respective group actions. For MCE, we associate new isomorphism invariants to corank-$1$ points of matrix codes, which lead to a birthday-type algorithm. We present empirical justifications that these isomorphism invariants are easy-to-compute and distinguishing, and provide an implementation of this algorithm. This algorithm has some implications to the security of MEDS. The invariant function for ATFE is similar, except it is associated with lower rank points. Modulo certain assumptions on turning the invariant function into canonical forms, our algorithm for ATFE improves on the runtime of the previously best known algorithm of Beullens (Crypto 2023). Finally, we present quantum variants of our classical algorithms with cubic runtime improvements.

The Partial Vandermonde (PV) Knapsack problem is an algebraic variant of the low-density inhomogeneous SIS problem. The problem has been used as a building block for various lattice-based constructions, including signatures (ACNS'14, ACISP'18), encryptions (DCC'15,DCC'20), and signature aggregation (Eprint'20). At Crypto'22, Boudgoust, Gachon, and Pellet-Mary proposed a key distinguishing attack on the PV Knapsack exploiting algebraic properties of the problem. Unfortunately, their attack doesn't offer key recovery, except for worst-case keys. In this paper, we propose an alternative attack on the PV Knapsack problem, which provides key recovery for a much larger set of keys. Like the Crypto'22 attack, it is based on lattice reduction and uses a dimension reduction technique to speed-up the underlying lattice reduction algorithm and enhance its performance. As a side bonus, our attack transforms the PV Knapsack problem into uSVP instances instead of SVP instances in the Crypto'22 attack. This also helps the lattice reduction algorithm, both from a theoretical and practical point of view. We use our attack to re-assess the hardness of the concrete parameters used in the literature. It appears that many contain a non-negligible fraction of weak keys, which are easily identified and extremely susceptible to our attack. For example, a fraction of $2^{-19}$ of the public keys of a parameter set from ACISP'18 can be solved in about $30$ hours on a moderate server using off-the-shelf lattice reduction. This parameter set was initially claimed to have a $129$-bit security against key recovery attack. Its security was reduced to $87$-bit security using the distinguishing attack from Crypto'22. Similarly, the ACNS'14 proposal also includes a parameter set containing a fraction of $2^{-19}$ of weak keys; those can be solved in about $17$ hours.

The Nostradamus attack was originally proposed as a security vulnerability for a hash function by Kelsey and Kohno at EUROCRYPT 2006. It requires the attacker to commit to a hash value y of an iterated hash function H. Subsequently, upon being provided with a message prefix P, the adversary’s task is to identify a suffix S such that H(P||S) equals y. Kelsey and Kohno demonstrated a herding attack requiring $O(\sqrt{n}\cdot 2^{2n/3})$ evaluations of the compression function of H, where n represents the output and state size of the hash, placing this attack between preimage attacks and collision searches in terms of complexity. At ASIACRYPT 2022, Benedikt et al. transform Kelsey and Kohno’s attack into a quantum variant, decreasing the time complexity from $O(\sqrt{n}\cdot 2^{2n/3})$ to $O(\sqrt[3]{n}\cdot 2^{3n/7})$. At ToSC 2023, Zhang et al. proposed the first dedicated Nostradamus attack on AES-like hashing in both classical and quantum settings. In this paper, we have made revisions to the multi-target technique incorporated into the meet-in-the-middle automatic search framework. This modification leads to a decrease in time complexity during the online linking phase, effectively reducing the overall attack time complexity in both classical and quantum scenarios. Specifically, we can achieve more rounds in the classical setting and reduce the time complexity for the same round in the quantum setting.

As an ISO/IEC standard, the hash function RIPEMD-160 has been used to generate the Bitcoin address with SHA-256. However, due to the complex double-branch structure of RIPEMD-160, the best collision attack only reaches 36 out of 80 steps of RIPEMD-160, and the best semi-free-start (SFS) collision attack only reaches 40 steps. To improve the 36-step collision attack proposed at EUROCRYPT 2023, we explored the possibility of using different message differences to increase the number of attacked steps, and we finally identified one choice allowing a 40-step collision attack. To find the corresponding 40-step differential characteristic, we re-implement the MILP-based method to search for signed differential characteristics with SAT/SMT. As a result, we can find a colliding message pair for 40-step RIPEMD-160 in practical time, which significantly improves the best collision attack on RIPEMD-160. For the best SFS collision attack published at ToSC 2019, we observe that the bottleneck is the probability of the right-branch differential characteristics as they are fully uncontrolled in the message modification. To address this issue, we utilize our SAT/SMT-based tool to search for high-probability differential characteristics for the right branch. Consequently, we can mount successful SFS collision attacks on 41, 42 and 43 steps of RIPEMD-160, thus significantly improving the SFS collision attacks. In addition, we also searched for a 44-step differential characteristic, but the differential probability is too low to allow a meaningful SFS collision attack.

The SHA-2 family including SHA-224, SHA-256, SHA-384, SHA-512, SHA-512/224 and SHA512/256 is a U.S. federal standard pub- lished by NIST. Especially, there is no doubt that SHA-256 is one of the most important hash functions used in real-world applications. Due to its complex design compared with SHA-1, there is almost no progress in collision attacks on SHA-2 after ASIACRYPT 2015. In this work, we retake this challenge and aim to significantly improve collision attacks on the SHA-2 family. First, we observe from many existing attacks on SHA-2 that the current advanced tool to search for SHA-2 characteristics has reached the bottleneck. Specifically, longer differential characteristics could not be found, and this causes that the collision attack could not reach more steps. To address this issue, we adopt Liu et al.’s MILP-based method and implement it with SAT/SMT for SHA-2, where we also add more techniques to detect contradictions in SHA-2 characteristics. This answers an open problem left in Liu et al.’s paper to apply the technique to SHA-2. With this SAT/SMT-based tool, we search for SHA-2 charac- teristics by controlling its sparsity in a dedicated way. As a result, we successfully find the first practical semi-free-start (SFS) colliding message pair for 39-step SHA-256, improving the best 38-step SFS collision attack published at EUROCRYPT 2013. In addition, we also report the first practical free-start (FS) collision attack on 40-step SHA-224, while the previously best theoretic 40-step attack has time complexity 2110. More- over, for the first time, we can mount practical and theoretic collision attacks on 28-step and 31-step SHA-512, respectively, which improve the best collision attack only reaching 27 steps of SHA-512 at ASIACRYPT 2015. In a word, with new techniques to find SHA-2 characteristics, we have made some notable progress in the analysis of SHA-2 after the major achievements made at EUROCRYPT 2013 and ASIACRYPT 2015.

At EUROCRYPT’23, Castryck and Decru, Maino et al., and Robert present efficient attacks against supersingular isogeny Diffie-Hellman key exchange protocol (SIDH). Drawing inspiration from these attacks, Andrea Basso, Luciano Maino, and Giacomo Pope introduce FESTA, an isogeny-based trapdoor function, along with a corresponding IND-CCA secure public key encryption (PKE) protocol at ASIACRYPT’23. FESTA incorporates either a diagonal or circulant matrix into the secret key to mask torsion points. In this paper, we employ a side-channel attack to construct an auxiliary verification oracle. By querying this oracle, we propose an adaptive attack strategy to recover the secret key in FESTA when the secret matrix is circulant. Compared with existing attacks, our strategy is more efficient and formal. Leveraging these findings, we implement our attack algorithms to recover the circulant matrix in secret key. Finally, we demonstrate that if the secret matrix is circulant, then the adversary can successfully recover FESTA’s secret key with a polynomial number of decryption machine queries. Consequently, our paper illustrates that FESTA PKE protocol with secret circulant matrix does not achieve IND-CCA security.

DME is a multivariate scheme submitted to the call for additional signatures recently launched by NIST. Its performance is one of the best among all the candidates. The public key is constructed from the alternation of very structured linear and non-linear components that constitute the private key, the latter being defined over an extension field. We exploit these structures by proposing an algebraic attack which is practical on all DME parameters.

The advancements in information technology have made the Advanced Encryption Standard (AES) and the PRESENT cipher indispensable in ensuring data security and facilitating private transactions. AES is renowned for its flexibility and widespread use in various fields, while the PRESENT cipher excels in lightweight cryptographic situations. This paper delves into a dual examination of the Key Scheduling Algorithms (KSAs) of AES and the PRESENT cipher, which play a crucial role in generating round keys for their respective encryption techniques. By implementing deep learning methods, particularly a Neural Network model, our study aims to unravel the complexities of these KSAs and shed light on their inner workings.

Learning parity with noise (LPN) has been widely studied and used in cryptography. It was recently brought to new prosperity since Boyle et al. (CCS'18), putting LPN to a central role in designing secure multi-party computation, zero-knowledge proofs, private set intersection, and many other protocols. In this paper, we thoroughly studied the security of LPN problems in this particular context. We found that some important aspects have long been ignored and many conclusions from classical LPN cryptanalysis do not apply to this new setting, due to the low noise rates, extremely high dimensions, various types (in addition to $\mathbb{F}_2$) and noise distributions. 1. For LPN over a field, we give a parameterized reduction from exact-noise LPN to regular-noise LPN. Compared to the recent result by Feneuil, Joux and Rivain (Crypto'22), we significantly reduce the security loss by paying only a small additive price in dimension and number of samples. 2. We analyze the security of LPN over a ring $\mathbb{Z}_{2^\lambda}$. Existing protocols based on LPN over integer rings use parameters as if they are over fields, but we found an attack that effectively reduces the weight of a noise by half compared to LPN over fields. Consequently, prior works that use LPN over $\mathbb{Z}_{2^\lambda}$ overestimate up to 40 bits of security. 3. We provide a complete picture of the hardness of LPN over integer rings by showing: 1) the equivalence between its search and decisional versions; 2) an efficient reduction from LPN over $\mathbb{F}_2$ to LPN over $\mathbb{Z}_{2^\lambda}$; and 3) generalization of our results to any integer ring. Finally, we provide an all-in-one estimator tool for the bit security of LPN parameters in the context of PCG, incorporating the recent advanced attacks.

LoPher brings, for the first time, cryptographic security promises to the field of logic locking in a bid to break the game of cat-and-mouse seen in logic locking. Toward this end, LoPher embeds the circuitry to lock within multiple rounds of a block cipher, by carefully configuring all the S-Boxes. To realize general Boolean functionalities and to support varying interconnect topologies, LoPher also introduces additional layers of MUXes between S-Boxes and the permutation operations. The authors of LoPher claim resilience against SAT-based attacks in particular. Here, we show the first successful attack on LoPher. First, we uncover a significant limitation for LoPher’s key-space configuration, resulting in large numbers of equivalent keys and, thus, a largely simplified search space for attackers in practice. Second, motivated by their well-proven working against ciphers, we employ a power side-channel attack against LoPher. We find that ISCAS-85 benchmarks locked with LoPher can all be broken in few thousands of traces. Finally, we also outline a simple and low-cost countermeasure to render LoPher more secure.

This paper focuses on the cryptanalysis of the ASCON family using automatic tools. We analyze two different problems with the goal to obtain new modelings, both simpler and less computationally heavy than previous works (all our models require only a small amount of code and run on regular desktop computers). The first problem is the search for Meet-in-the-middle attacks on reduced-round ASCON-Hash. Starting from the MILP modeling of Qin et al. (EUROCRYPT 2023 & ePrint 2023), we rephrase the problem in SAT, which accelerates significantly the solving time and removes the need for the ``weak diffusion structure'' heuristic. This allows us to reduce the memory complexity of Qin et al.'s attacks and to prove some optimality results. The second problem is the search for lower bounds on the probability of differential characteristics for the ASCON permutation. We introduce a lossy MILP encoding of the propagation rules based on the Hamming weight, in order to find quickly lower bounds which are comparable to the state of the art. We find a small improvement over the existing bound on 7 rounds.

Physical attacks pose a substantial threat to the secure implementation of cryptographic algorithms. While considerable research efforts are dedicated to protecting against passive physical attacks (e.g., side-channel analysis (SCA)), the landscape of protection against other types of physical attacks remains a challenge. Fault attacks (FA), though attracting growing attention in research, still lack the prevalence of provably secure designs when compared to SCA. The realm of combined attacks, which leverage the capabilities of both SCA and FA adversaries, introduces powerful adversarial models, rendering protection against them challenging. This challenge has consequently led to a relatively unexplored area of research, resulting in a notable gap in understanding and efficiently protecting against combined attacks. The CAPA countermeasure, published at CRYPTO 2018, addresses this challenge with a robust adversarial model that goes beyond conventional SCA and FA adversarial models. Drawing inspiration from the principles of Multiparty Computation (MPC), CAPA claims security against higher-order SCA, higher-order fault attacks, and their combination. In this work, we present a combined attack that breaks CAPA within the constraints of its assumed adversarial model. In response, we propose potential fixes to the design of CAPA that increase the complexity of the proposed attack, although not provably thwarting it. With this presented combined attack, we highlight the difficulty of effectively protecting against combined attacks.

Statistical Fault Attacks (SFA), introduced by Fuhr et al., exploit the statistical bias resulting from injected faults. Unlike prior fault analysis attacks, which require both faulty and correct ciphertexts under the same key, SFA leverages only faulty ciphertexts. In CHES 2018, more powerful attacks called Statistical Ineffective Fault Attacks (SIFA) have been proposed. In contrast to the previous fault attacks that utilize faulty ciphertexts, SIFA exploits the distribution of the intermediate values leading to fault-free ciphertexts. As a result, the SIFA attacks were shown to be effective even in the presence of widely used fault injection countermeasures based on detection and infection. In this work, we build upon the core idea of SIFA, and provide two main practical improvements over the previously proposed analysis methods. Firstly, we show how to perform SIFA from the input side, which in contrast to the original SIFA, requires injecting faults in the earlier rounds of an encryption or decryption operation. If we consider the start of the operation as the trigger for fault injection, the cumulative jitter in the first few rounds of a cipher is much lower than the last rounds. Hence, performing the attack in the first or second round requires a narrower parameter range for fault injection and hence less fault injection attempts to recover the secret key. Secondly, in comparison to the straightforward SIFA approach of guessing 32-bits at a time, we propose a chosen input approach that reduces the guessing effort to 16-bits at a time. This decreases the key search space for full key recovery of an AES-128 implementation from $2^{34}$ to $2^{19}$.

Since its introduction by Wagner (CRYPTO `02), the $k$-list algorithm has found significant utility in cryptanalysis. One important application thereof is in computing forgeries on several interactive signature schemes that implicitly rely on the hardness of the ROS problem formulated by Schnorr (ICICS `01). The current best attack strategy for these schemes relies the conjectured runtime of the $k$-list algorithm over $\mathbb{Z}_p$. The tightest known analysis of Wagner's algorithm over $\mathbb{Z}_p$ is due to Shallue (ANTS `08). However, it hides large polynomial factors and leaves a gap with respect to desirable concrete parameters for the attack. In this work, we develop a degraded version of the $k$-list algorithm which provably enforces the heuristic invariants in Wagner's original. In the process, we devise and analyze a new list merge procedure that we dub the interval merge. We give a thorough analysis of the runtime and success probability of our degraded algorithm, and show that it beats the projected runtime of the analysis by Shallue for parameters relevant to the generalized ROS attack of Benhamouda et al. (EUROCRYPT `21). For a $256$-bit prime $p$, and $k = 8$, our degraded $k$-list algorithm runs in time $\approx 2^{70.4}$, while Shallue's analysis states that the Wagner's original algorithm runs in time $\approx 2^{98.3}$.

Multivariate cryptography is one of the main candidates for creating post-quantum public key cryptosystems. Especially in the area of digital signatures, there exist many practical and secure multivariate schemes. The signature schemes UOV and Rainbow are two of the most promising and best studied multivariate schemes which have proven secure for more than a decade. However, so far the security of multivariate signature schemes towards physical attacks has not been appropriately assessed. Towards a better understanding of the physical security of multivariate signature schemes, this paper presents fault attacks against SingleField schemes, especially UOV and Rainbow. Our analysis shows that although promising attack vectors exist, multivariate signature schemes inherently offer a good protection against fault attacks.

Thus far, several papers estimated concrete quantum resources of Shor’s algorithm for solving a binary elliptic curve discrete logarithm problem. In particular, the complexity of computing quantum inversions over a binary field F2n is dominant when running the algorithm, where n is a degree of a binary elliptic curve. There are two major methods for quantum inversion, i.e., the quantum GCD-based inversion and the quantum FLT-based inversion. Among them, the latter method is known to require more qubits; however, the latter one is valuable since it requires much fewer Toffoli gates and less depth. When n = 571, Kim-Hong’s quantum GCD-based inversion algorithm (Quantum Information Processing 2023) and Taguchi-Takayasu’s quantum FLT-based inversion algorithm (CT-RSA 2023) require 3, 473 qubits and 8, 566 qubits, respectively. In contrast, for the same n = 571, the latter algorithm requires only 2.3% of Toffoli gates and 84% of depth compared to the former one. In this paper, we modify Taguchi-Takayasu’s quantum FLT-based inversion algorithm to reduce the required qubits. While Taguchi-Takayasu’s FLT-based inversion algorithm takes an addition chain for n−1 as input and computes a sequence whose number is the same as the length of the chain, our proposed algorithm employs an uncomputation step and stores a shorter one. As a result, our proposed algorithm requires only 3, 998 qubits for n = 571, which is only 15% more than Kim-Hong’s GCD-based inversion algorithm. Furthermore, our proposed algorithm preserves the advantage of FLT-based inversion since it requires only 3.7% of Toffoli gates and 77% of depth compared to Kim-Hong’s GCD-based inversion algorithm for n = 571.

The Regular Syndrome Decoding (RSD) problem, a variant of the Syndrome Decoding problem with a particular error distribution, was introduced almost 20 years ago by Augot et al. . In this problem, the error vector is divided into equally sized blocks, each containing a single noisy coordinate. More recently, the last five years have seen increased interest in this assumption due to its use in MPC and ZK applications. Generally referred to as "LPN with regular noise" in this context, the assumption allows to achieve better efficiency when compared to plain LPN. We present the first attack on RSD relying on Gröbner bases techniques. After a careful theoretical analysis of the underlying polynomial system, we propose concrete attacks that are able to take advantage of the regular noise distribution. In particular, we can identify several examples of concrete parameters where our techniques outperform other algorithms.

Transaction fees compensate actors for resources expended on transactions and can only be charged from transactions included in blocks. But, the expressiveness of Turing-complete contracts implies that verifying if transactions can be included requires executing them on the current blockchain state. In this work, we show that adversaries can craft malicious transactions that decouple the work imposed on blockchain actors from the compensation offered in return. We introduce three attacks: (i) ConditionalExhaust, a conditional resource-exhaustion attack against blockchain actors. (ii) MemPurge, an attack for evicting transactions from actors' mempools. (iii) GhostTX, an attack on the reputation system used in Ethereum's proposer-builder separation ecosystem. We evaluate our attacks on an Ethereum testnet and find that by combining ConditionalExhaust and MemPurge, adversaries can simultaneously burden victims' computational resources and clog their mempools to the point where victims are unable to include transactions in blocks. Thus, victims create empty blocks, thereby hurting the system's liveness. The attack's expected cost is $376, but becomes cheaper if adversaries are validators. For other attackers, costs decrease if censorship is prevalent in the network. ConditionalExhaust and MemPurge are made possible by inherent features of Turing-complete blockchains, and potential mitigations may result in reducing a ledger's scalability.

We present an attack on Ethereum's consensus mechanism which can be used by miners to obtain consistently higher mining rewards compared to the honest protocol. This attack is novel in that it does not entail withholding blocks or any behavior which has a non-zero probability of earning less than mining honestly, in contrast with the existing literature. This risk-less attack relies instead on manipulating block timestamps, and carefully choosing whether and when to do so. We present this attack as an algorithm, which we then analyze to evaluate the revenue a miner obtains from it, and its effect on a miner's absolute and relative share of the main-chain blocks. The attack allows an attacker to replace competitors' main-chain blocks after the fact with a block of its own, thus causing the replaced block's miner to lose all transactions fees for the transactions contained within the block, which will be demoted from the main-chain. This block, although ``kicked-out'' of the main-chain, will still be eligible to be referred to by other main-chain blocks, thus becoming what is commonly called in Ethereum an uncle. We proceed by defining multiple variants of this attack, and assessing whether any of these attacks has been performed in the wild. Surprisingly, we find that this is indeed true, making this the first case of a confirmed consensus-level manipulation performed on a major cryptocurrency. Additionally, we implement a variant of this attack as a patch for geth, Ethereum's most popular client, making it the first consensus-level attack on Ethereum which is implemented as a patch. Finally, we suggest concrete fixes for Ethereum's protocol and implemented them as a patch for geth which can be adopted quickly and mitigate the attack and its variants.

At Eurocrypt 2023, a differential attack on the block cipher Speedy-7-192 was presented. This note shows that the main differential characteristic that this attack is based on has probability zero.

Fully homomorphic encryption (FHE) is an advanced cryptography technique to allow computations (i.e., addition and multiplication) over encrypted data. After years of effort, the performance of FHE has been significantly improved and it has moved from theory to practice. The transciphering framework is another important technique in FHE to address the issue of ciphertext expansion and reduce the client-side computational overhead. To apply the transciphering framework to the CKKS FHE scheme, a new transciphering framework called the Real-to-Finite-Field (RtF) framework and a corresponding FHE-friendly symmetric-key primitive called HERA were proposed at ASIACRYPT 2021. Although HERA has a very similar structure to \aes, it is considerably different in the following aspects: 1) the power map $x\mapsto x^3$ is used as the S-box; 2) a randomized key schedule is used; 3) it is over a prime field $\mathbb F_p$ with $p>2^{16}$. In this work, we perform the first third-party cryptanalysis of HERA, by showing how to mount new algebraic attacks with multiple collisions in the round keys. Specifically, according to the special way to randomize the round keys in HERA, we find it possible to peel off the last nonlinear layer by using collisions in the last-round key and a simple property of the power map. In this way, we could construct an overdefined system of equations of a much lower degree in the key, and efficiently solve the system via the linearization technique. As a result, for HERA with 192 and 256 bits of security, respectively, we could break some parameters under the same assumption made by designers that the algebra constant $\omega$ for Gaussian elimination is $\omega=2$, i.e., Gaussian elimination on an $n\times n$ matrix takes $\mathcal{O}(n^{\omega})$ field operations. If using more conservative choices like $\omega\in\{2.8,3\}$, our attacks can also successfully reduce the security margins of some variants of HERA to only 1 round. However, the security of HERA with 80 and 128 bits of security is not affected by our attacks due to the high cost to find multiple collisions. In any case, our attacks reveal a weakness of HERA caused by the randomized key schedule and its small state size.

This paper introduces an algorithm to efficiently break the Decisional Diffie-Hellman (DDH) assumption in totally non-maximal imaginary quadratic orders, specifically when $\Delta_1 = 3$, and $f$ is non-prime with knowledge of a single factor. Inspired by Shanks and Dedekind's work on 3-Sylow groups, we generalize their observations to undermine DDH security.

Lattice-based cryptographic schemes such as Crystals-Kyber and Dilithium are post-quantum algorithms selected to be standardized by NIST as they are considered to be secure against quantum computing attacks. The multiplication in polynomial rings is the most time-consuming operation in many lattice-based cryptographic schemes, which is also subject to side-channel attacks. While NTT-based polynomial multiplication is almost a norm in a wide range of implementations, a relatively new method, incomplete NTT is preferred to accelerate lattice-based cryptography, especially on some computing platforms that feature special instructions. In this paper, we present a novel, efficient and non-profiled power/EM side-channel attack targeting polynomial multiplication based on the incomplete NTT algorithm. We apply the attack on the Crystals-Dilithium signature algorithm and Crystals-Kyber KEM. We demonstrate that the method accelerates attack run-time when compared to the existing approaches. While a conventional non-profiled side-channel attack tests a much larger hypothesis set because it needs to predict two coefficients of secret polynomials together, we propose a much faster zero-value filtering attack (ZV-FA), which reduces the size of the hypothesis set by targeting the coefficients individually. We also propose an effective and efficient validation and correction technique employing the inverse NTT to estimate and modify the mispredicted coefficients. Our experimental results show that we can achieve a speed-up of 1915x over brute-force.

Deep learning is very efficient at breaking masked implementations even when the attacker does not assume knowledge of the masks. However, recent works pointed out a significant challenge: overcoming the initial learning plateau. This paper discusses the advantages of multi-task learning to break through the initial plateau consistently. We investigate different ways of applying multi-task learning against masked AES implementations (via the ASCAD-r, ASCAD-v2, and CHESCTF-2023 datasets) under the assumption that the attacker cannot access masks during training. We offer evidence that multi-task learning significantly increases the consistency of convergence and performance of deep neural networks. Our work provides a wide range of experiments to understand the benefits of multi-task strategies over the current single-task state-of-the-art. Furthermore, such strategies achieve novel milestones against protected implementations as we propose models that defeat all masks of the affine masking on ASCAD-v2 for the first time.

Kyber KEM, the NIST selected PQC standard for Public Key Encryption and Key Encapsulation Mechanisms (KEMs) has been subjected to a variety of side-channel attacks, through the course of the NIST PQC standardization process. However, all these attacks targeting the decapsulation procedure of Kyber KEM either require knowledge of the ciphertexts or require to control the value of ciphertexts for key recovery. However, there are no known attacks in a blind setting, where the attacker does not have access to the ciphertexts. While blind side-channel attacks are known for symmetric key cryptographic schemes, we are not aware of such attacks for Kyber KEM. In this paper, we fill this gap by proposing the first blind side-channel attack on Kyber KEM. We target leakage of the pointwise multiplication operation in the decryption procedure to carry out practical blind side-channel attacks resulting in full key recovery. We perform practical validation of our attack using power side-channel from the reference implementation of Kyber KEM taken from the pqm4 library, implemented on the ARM Cortex-M4 microcontroller. Our experiments clearly indicate the feasibility of our proposed attack in recovering the full key in only a few hundred to few thousand traces, in the presence of a suitably accurate Hamming Weight (HW) classifier.

VOX is a UOV-like signature scheme submitted to Round 1 additional signatures of NIST PQC standardization process. In 2023 Furue and Ikematsu proposed a rectangular MinRank attack on VOX, resulting in the submitters changing their parameters to counter this attack. In this paper we propose a new type of MinRank attack called padded MinRank attack. We show that the attack is highly efficient in its running time, taking less than one minute to break eight of nine parameters and about eight hours for the remaining one. Therefore the parameters of VOX should be reexamined to ensure its safety.

Learning with Errors (LWE) is a hard math problem underlying post-quantum cryptography (PQC) systems for key exchange and digital signatures, recently standardized by NIST. Prior work [Wenger et al., 2022; Li et al., 2023a;b] proposed new machine learning (ML)-based attacks on LWE problems with small, sparse secrets, but these attacks require millions of LWE samples to train on and take days to recover secrets. We propose three key methods—better pre-processing, angular embeddings and model pre-training—to improve these attacks, speeding up preprocessing by 25× and improving model sample efficiency by 10×. We demonstrate for the first time that pre-training improves and reduces the cost of ML attacks on LWE. Our architecture improvements enable scaling to larger-dimension LWE problems: this work is the first instance of ML attacks recovering sparse binary secrets in dimension n = 1024, the smallest dimension used in practice for homomorphic encryption applications of LWE where sparse binary secrets are proposed.

Speculative out-of-order execution is a strategy of masking execution latency by allowing younger instructions to execute before older instructions. While originally considered to be innocuous, speculative out-of-order execution was brought into the spotlight with the 2018 publication of the Spectre and Meltdown attacks. These attacks demonstrated that microarchitectural side channels can leak sensitive data accessed by speculatively executed instructions that are not part of the normal program execution. Since then, a significant effort has been vested in investigating how microarchitectural side channels can leak data from speculatively executed instructions and how to control this leakage. However, much less is known about how speculative out-of-order execution affects microarchitectural side-channel attacks. In this paper, we investigate how speculative out-of-order execution affects the Evict+Time cache attack. Evict+Time is based on the observation that cache misses are slower than cache hits, hence by measuring the execution time of code, an attacker can determine if a cache miss occurred during the execution. We demonstrate that, due to limited resources for tracking out-of-order execution, under certain conditions an attacker can gain more fine-grained information and determine whether a cache miss occurred in part of the executed code. Based on the observation, we design the Evict+Spec+Time attack, a variant of Evict+Time that can learn not only whether a cache miss occurred, but also in which part of the victim code it occurred. We demonstrate that Evict+Spec+Time is an order of magnitude more efficient than Evict+Time when attacking a T-table-based implementation of AES. We further show an Evict+Spec+Time attack on an S-box-based implementation of AES, recovering the key with as little as 14389 decryptions. To the best of our knowledge, ours is the first successful Evict+Time attack on such a victim.

Biscuit is a recent multivariate signature scheme based on the MPC-in-the-Head paradigm. It has been submitted to the NIST competition for additional signature schemes. Signatures are derived from a zero-knowledge proof of knowledge of the solution of a structured polynomial system. This extra structure enables efficient proofs and compact signatures. This short note demonstrates that it also makes these polynomial systems easier to solve than random ones. As a consequence, the original parameters of Biscuit failed to meet the required security levels and had to be upgraded.

Isogeny-based cryptography is one of the candidates for post-quantum cryptography. In 2023, Kani's theorem breaks an isogeny-based scheme SIDH, which was considered a promising post-quantum scheme. Though Kani's theorem damaged isogeny-based cryptography, some researchers have been trying to dig into the applications of this theorem. A FESTA trapdoor function is an isogeny-based trapdoor function that is one trial to apply Kani's theorem to cryptography. This paper claims that there is an adaptive attack for a FESTA-based scheme if this scheme does not check the correctness of the input matrix. Our attack cannot be adapted to IND-CCA PKE schemes named FESTA proposed in the FESTA original paper so far. In this paper, we provide an adaptive attack for a FESTA trapdoor function using a specific oracle, and it reveals the secret key of the function. This oracle may be constructed if the FESTA trapdoor function is used in the wrong way (\textit{i.e.,} without the checking process of the input matrix). As an example, we explain that our attack can be adapted to a possible PKE scheme based on a FESTA trapdoor function in the wrong way.

After NIST’s selection of Dilithium as the primary future standard for quantum-secure digital signatures, increased efforts to understand its implementation security properties are required to enable widespread adoption on embedded devices. Concretely, there are still many open questions regarding the susceptibility of Dilithium to fault attacks. This is especially the case for Dilithium’s randomized (or hedged) signing mode, which, likely due to devastating implementation attacks on the deterministic mode, was selected as the default by NIST. This work takes steps towards closing this gap by presenting two new key-recovery fault attacks on randomized/hedged Dilithium. Both attacks are based on the idea of correcting faulty signatures after signing. A successful correction yields the value of a secret intermediate that carries information on the key. After gathering many faulty signatures and corresponding correction values, it is possible to solve for the signing key via either simple linear algebra or lattice-reduction techniques. Our first attack extends a previously published attack based on an instruction-skipping fault to the randomized setting. Our second attack injects faults in the matrix A, which is part of the public key. As such, it is not sensitive to side-channel leakage and has, potentially for this reason, not seen prior analysis regarding faults. We show that for Dilithium2, the attacks allow key recovery with as little as 1024 and 512 faulty signatures, respectively, with each signature generated by injecting a single targeted fault. We also demonstrate how our attacks can be adapted to circumvent several popular fault countermeasures with a moderate increase in the computational runtime and the number of required faulty signatures. These results are verified using both simulated faults and clock glitches on an ARM-based microcontroller. The presented attacks demonstrate that also randomized Dilithium can be subject to diverse fault attacks, that certain countermeasures might be easily bypassed, and that potential fault targets reach beyond side-channel sensitive operations. Still, many further operations are likely also susceptible, implying the need for increased analysis efforts in the future.

In this paper we describe attacks on the UOV-based signature scheme called MQ-Sign. MQ-Sign was submitted by Shim, Kim, and An as a a first-round candidate for standardization in the (South) Korean post-quantum cryptography competition (KpqC). The scheme makes use of sparseness of the secret central polynomials and equivalent key construction to reduce the size of the private key. The authors propose four variants exploiting different levels of sparsity, MQ-Sign-SS, MQ-Sign-RS, MQ-Sign-SR, and MQ-Sign-RR with the last one being the standard UOV signature scheme. We show that apart from the MQ-Sign-RR variant, all the others are insecure. Namely, we present a polynomial-time key-recovery attack on the variants MQ-Sign-SS and MQ-Sign-RS and a forgery attack on the variant MQ-Sign-SR below the claimed security level. Our attack exploits exactly the techniques used for reduction of keys - the sparsity of the central polynomials in combination with the specific structure of the secret linear map $\mathbf{S}$. We provide a verification script for the polynomial-time key-recovery attack, that recovers the secret key in less than seven seconds for security level V. Furthermore, we provide an implementation of the non-guessing part of the forgery attack, confirming our complexity estimates.

Post-Quantum Cryptography (PQC) was proposed due to the potential threats quantum computer attacks against conventional public key cryptosystems, and four PQC algorithms besides CRYSTALS-Dilithium (Dilithium for short) have so far been selected for NIST standardization. However, the selected algorithms are still vulnerable to side-channel attacks in practice, and their physical security need to be further evaluated. This study introduces two efficient power analysis attacks, the optimized fast two-stage approach and the single-bit approach, aimed at reducing the key guess space in NTT polynomial multiplication on an STM32F405 device (ARM Cortex-M4 core). Our findings reveal that the optimized approach outperforms the conservative approach and the fast two-stage approach proposed in ICCD 2021 by factors of 519 and 88, respectively. Similarly, the single-bit approach demonstrates speedups of 365 and 62 times compared to these two approaches, respectively.

Neural cryptanalysis is the study of cryptographic primitives throughmachine learning techniques. Following Gohr’s seminal paper at CRYPTO 2019, afocus has been placed on improving the accuracy of such distinguishers against specific primitives, using dedicated training schemes, in order to obtain better key recovery attacks based on machine learning. These distinguishers are highly specialized and not trivially applicable to other primitives. In this paper, we focus on the opposite problem: building a generic pipeline for neural cryptanalysis. Our tool is composed of two parts. The first part is an evolutionary algorithm for the search of good input differences for neural distinguishers. The second part is DBitNet, a neuraldistinguisher architecture agnostic to the structure of the cipher. We show thatthis fully automated pipeline is competitive with a highly specialized approach, inparticular for SPECK32, and SIMON32. We provide new neural distinguishers forseveral primitives (XTEA, LEA, HIGHT, SIMON128, SPECK128) and improve overthe state-of-the-art for PRESENT, KATAN, TEA and GIMLI.

Private Set Intersection (PSI) is a widely used protocol that enables two parties to securely compute a function over the intersected part of their shared datasets and has been a significant research focus over the years. However, recent studies have highlighted its vulnerability to Set Membership Inference Attacks (SMIA), where an adversary might deduce an individual's membership by invoking multiple PSI protocols. This presents a considerable risk, even in the most stringent versions of PSI, which only return the cardinality of the intersection. This paper explores the evaluation of anonymity within the PSI context. Initially, we highlight the reasons why existing works fall short in measuring privacy leakage, and subsequently propose two attack strategies that address these deficiencies. Furthermore, we provide theoretical guarantees on the performance of our proposed methods. In addition to these, we illustrate how the integration of auxiliary information, such as the sum of payloads associated with members of the intersection (PSI-SUM), can enhance attack efficiency. We conducted a comprehensive performance evaluation of various attack strategies proposed utilizing two real datasets. Our findings indicate that the methods we propose markedly enhance attack efficiency when contrasted with previous research endeavors. The effective attacking implies that depending solely on existing PSI protocols may not provide an adequate level of privacy assurance. It is recommended to combine privacy-enhancing technologies synergistically to enhance privacy protection even further.

In this article, we examine Differential Fault Attacks (DFA) targeting two stream ciphers, FLIP and FiLIP. We explore the fault model where an adversary flips a single bit of the key at an unknown position. Our analysis involves establishing complexity bounds for these attacks, contingent upon the cryptographic parameters of the Boolean functions employed as filters and the key size. Initially, we demonstrate how the concept of sensitivity enables the detection of the fault position using only a few keystream bits. This represents an enhancement over previous DFA methodologies applied to these ciphers. Subsequently, we leverage the properties of the filter's derivatives to execute attacks. This approach is universally applicable to any filter, and we delineate specific attack strategies for the two function families previously implemented in these ciphers.

SNOVA is a multivariate signature scheme submitted to the ad- ditional NIST PQC standardization project started in 2022. SNOVA is con- structed by incorporating the structure of the matrix ring over a finite field into the UOV signature scheme, and the core part of its public key is the UOV public key whose coefficients consist of matrices. As a result, SNOVA dramatically reduces the public key size compared to UOV. In this paper, we recall the construction of SNOVA, and reconsider its security analysis. In particular, we investigate key recovery attacks applied to the core part of the public key of SNOVA in detail. Due to our analysis, we show that some pa- rameters of SNOVA submitted in the additional NIST PQC standardization do not satisfy the claimed security levels.

We introduce the Order vs. Chaos (OvC) classifier, a novel language-model approach for side-channel attacks combining the strengths of multitask learning (via the use of a language model), multimodal learning, and deep metric learning. Our methodology offers a viable substitute for the multitask classifiers used for learning multiple targets, as put forward by Masure et al. We highlight some well-known issues with multitask classifiers, like scalability, balancing multiple tasks, slow learning, large model sizes, and the need for complex hyperparameter tuning. Thus, we advocate language models in side-channel attacks. We demonstrate improvements in results on different variants of ASCAD-V1 and ASCAD-V2 datasets compared to the existing state-of-the-art results. Additionally, we delve deeper with experiments on protected simulated datasets, allowing us to control noise levels and simulate specific leakage models. This exploration facilitates an understanding of the ramifications when the protective scheme's masks do not leak and allows us to further compare our approach with other approaches. Furthermore, with the help of unprotected simulated datasets, we demonstrate that the OvC classifier, uninformed of the leakage model, can parallelize the proficiency of a conventional multi-class classifier that is leakage model-aware. This finding implies that our methodology sidesteps the need for a predetermined leakage model in side-channel attacks.

Remote side-channel attacks on processors exploit hardware and micro-architectural effects observable from software measurements. So far, the analysis of micro-architectural leakages over physical side-channels (power consumption, electromagnetic field) received little treatment. In this paper, we argue that those attacks are a serious threat, especially against systems such as smartphones and Internet-of-Things (IoT) devices which are physically exposed to the end-user. Namely, we show that the observation of Dynamic Random Access Memory (DRAM) accesses with an electromagnetic (EM) probe constitutes a reliable alternative to time measurements in cache side-channel attacks. We describe the EVICT+EM attack, that allows recovering a full AES key on a T-Tables implementation with similar number of encryptions than state-of-the-art EVICT+RELOAD attacks on the studied ARM platforms. This new attack paradigm removes the need for shared memory and exploits EM radiations instead of high precision timers. Then, we introduce PRIME+EM, which goal is to reverse-engineer cache usage patterns. This attack allows to recover the layout of lookup tables within the cache. Finally, we present COLLISION+EM, a collision-based attack on a System-on-chip (SoC) that does not require malicious code execution, and show its practical efficiency in recovering key material on an ARM TrustZone application. Those results show that physical observation of the micro-architecture can lead to improved attacks.

We present a set of physical profiled attacks against CRYSTALS-Dilithium that accumulate noisy knowledge on secret keys over multiple signatures, finally leading to a full recovery attack. The methodology is composed of two steps. The first step consists of observing or inserting a bias in the posterior distribution of sensitive variables. The second step of an information processing phase which is based on belief propagation, which allows effectively exploiting that bias. The proposed concrete attacks rely on side-channel information, injection of fault attacks, or a combination of the two. Interestingly, the adversary benefits from the knowledge on the released signature, but is not dependent on it. We show that the combination of a physical attack with the binary knowledge of acceptance or rejection of a signature also leads to exploitable information on the secret key. Finally, we demonstrate that this approach is also effective against shuffled implementations of CRYSTALS-Dilithium.

In this paper, we focus on the common prime RSA variant and introduces a novel investigation into the partial key exposure attack targeting it. We explore the vulnerability of this RSA variant, which employs two common primes $p$ and $q$ defined as $p=2ga+1$ and $q=2gb+1$ for a large prime $g$. Previous cryptanalysis of common prime RSA has primarily focused on the small private key attack. In our work, we delve deeper into the realm of partial key exposure attacks by categorizing them into three distinct cases. We are able to identify weak private keys that are susceptible to partial key exposure by using the lattice-based method for solving simultaneous modular univariate linear equations. To validate the effectiveness and soundness of our proposed attacks, we conduct experimental evaluations. Through these examinations, we demonstrate the validity and practicality of the proposed partial key exposure attacks on common prime RSA.

The usage of convolutional neural networks (CNNs) to break cryptographic systems through hardware side-channels has enabled fast and adaptable attacks on devices like smart cards and TPMs. Current literature proposes fixed CNN architectures designed by domain experts to break such systems, which is time-consuming and unsuitable for attacking a new system. Recently, an approach using neural architecture search (NAS), which is able to acquire a suitable architecture automatically, has been explored. These works use the secret key information in the attack dataset for optimization and only explore two different search strategies using one-dimensional CNNs. We propose a NAS approach that relies only on using the profiling dataset for optimization, making it fully black-box. Using a large-scale experimental parameter study, we explore which choices for NAS, such as 1-D or 2-D CNNs and search strategy, produce the best results on 10 state-of-the-art datasets for Hamming weight and identity leakage models. We show that applying the random search strategy on 1-D inputs results in a high success rate and retrieves the correct secret key using a single attack trace on two of the datasets. This combination matches the attack efficiency of fixed CNN architectures, outperforming them in 4 out of 10 datasets. Our experiments also point toward the need for repeated attack evaluations of machine learning-based solutions in order to avoid biased performance estimates.

Higher order differential properties constitute a very insightful tool at the hands of a cryptanalyst allowing for probing a cryptographic primitive from an algebraic perspective. In FSE 2017, Saha et al. reported SymSum (referred to as SymSum_Vec in this paper), a new distinguisher based on higher order vectorial Boolean derivatives of SHA-3, constituting one of the best distinguishers on the latest cryptographic hash standard. SymSum_Vec exploits the difference in the algebraic degree of highest degree monomials in the algebraic normal form of SHA-3 with regards to their dependence on round constants. Later in Africacrypt 2020, Suryawanshi et al. extended SymSum_Vec using linearization techniques and in SSS 2023 also applied it to NIST-LWC finalist Xoodyak. However, a major limitation of SymSum_Vec is the maximum attainable derivative (MAD) which is less than half of the widely studied ZeroSum distinguisher. This is attributed to SymSum_Vec being dependent on m−fold vectorial derivatives while ZeroSum relies on m−fold simple derivatives. In this work we overcome this limitation of SymSum_Vec by developing and validating the theory of computing SymSum_Vec with simple derivatives. This gives us a close to 100% improvement in the MAD that can be computed. The new distinguisher reported in this work can also be combined with one/two-round linearization to penetrate more rounds. Moreover, we identify an issue with the two-round linearization claim made by Suryawanshi et al. which renders it invalid and also furnish an algebraic fix at the cost of some additional constraints. Combining all results we report SymSum_Sim , a new variant of the SymSum_Vec distinguisher based on m−fold simple derivatives that outperforms ZeroSum by a factor of $2^{257}$, $2^{129}$ for 10-round SHA-3-384 and 9-round SHA-3-512 respectively while enjoying the same MAD as ZeroSum. For every other SHA-3 variant, SymSum_Sim maintains an advantage of factor 2. Combined with one/two-round linearization, SymSum_Sim improves upon all existing ZeroSum and SymSum_Vec distinguishers on both SHA-3 and Xoodyak. As regards Keccak-p, the internal permutation of SHA-3, we report the best 15-round distinguisher with a complexity of $2^{256}$ and the first better than birthday-bound 16-round distinguisher with a complexity of $2^{512}$ (improving upon the 15/16-round results by Guo et al. in Asiacrypt 2016). We also devise the best full-round distinguisher on the Xoodoo internal permutation of Xoodyak with a practically verifiable complexity of $2^{32}$ and furnish the first third-party distinguishers on the Belarushian hash function Bash. All distinguishers furnished in this work have been verified through implementations whenever practically viable. Overall, with the MAD barrier broken, SymSum_Sim emerges as a better distinguisher than ZeroSum on all fronts and adds to the state-of-the-art of cryptanalytic tools investigating non-randomness of crypto primitives.

A recent preprint [ePrint 2023/1475] suggests the use of polynomials over a tropical algebra to construct a digital signature scheme "based on" the problem of factoring such polynomials, which is known to be NP‑hard. This short note presents two very efficient forgery attacks on the scheme, bypassing the need to factorize tropical polynomials and thus demonstrating that security in fact rests on a different, empirically easier problem.

As the message recovery-based attack poses a serious threat to lattice-based schemes, we conducted a study on the side-channel secu- rity of parallel implementations of lattice-based key encapsulation mech- anisms. Initially, we developed a power model to describe the power leakage during message encoding. Utilizing this power model, we pro- pose a multi-ciphertext message recovery attack, which can retrieve the required messages for a chosen ciphertext attack through a suitable mes- sage recovery oracle. Building upon the successful message recovery, we further develop a key recovery method based on a ciphertext-choosing strategy that maximizes key recovery accuracy, as well as a lattice reduc- tion attack capable of solving the whole private key from the target LWE instance. To assess the effectiveness of the attack, we conducted experi- ments using Kyber768 implemented on a Xilinx FPGA board. The exper- imental results demonstrate that our attack could successfully recover the private key with 9600 power traces and a computational complexity of 100 bikz, which is a significant advantage over existing attacks. Notably, our attack remains effective despite countermeasures such as masking and shuffling being implemented. This study reveals that parallel im- plementations remain vulnerable to side-channel attacks, and highlights the necessity of additional analysis and countermeasures for lattice-based schemes implemented in parallel.

We point out critical deficiencies in lattice-based cryptanalysis of common prime RSA presented in ``Remarks on the cryptanalysis of common prime RSA for IoT constrained low power devices'' [Information Sciences, 538 (2020) 54--68]. To rectify these flaws, we carefully scrutinize the relevant parameters involved in the analysis during solving a specific trivariate integer polynomial equation. Additionally, we offer a synthesized attack illustration of small private key attacks on common prime RSA.

$\Sigma$-protocols, a class of interactive two-party protocols, which are used as a framework to instantiate many other authentication schemes, are automatically a proof of knowledge (PoK) given that they satisfy the "special-soundness" property. This fact provides a convenient method to compose $\Sigma$-protocols and PoKs for complex relations. However, composing in this manner can be error-prone. While they must satisfy special-soundness, this is unfortunately not the case for many recently proposed composed practical schemes. Here we explore two schemes: ZKAttest from Faz-Hernández et al. and the ones from Agrawal et al., and show that their $\Sigma$-protocol's suffer from several security misdesigns which invalidate their security proofs, and state a practical cheap attack on ZKAttest's implementation. By exploring and resolving their misdesigns, we propose CDLS, a sound and secure variant of their protocols.

We show that the authentication scheme [Comput. Networks, 225 (2023), 109664] is flawed. (1) Some parameters are not specified. (2) Some computations are inconsistent. (3) It falsely require the control gateway to share its private key with the medical expert. (4) The scheme fails to keep user anonymity, not as claimed.

Since the existing tropical cryptographic protocols are either susceptible to the Kotov-Ushakov attack and its generalization, or to attacks based on tropical matrix periodicity and predictive behaviour, several attempts have been made to propose protocols that resist such attacks. Despite these attempts, many of the proposed protocols remain vulnerable to attacks targeting the underlying hidden problems, one of which we call the tropical two-sided discrete logarithm with shift. An illustrative case is the tropical Stickel protocol, which, when formulated with a single monomial instead of a polynomial, becomes susceptible to attacks based on solutions of the above mentioned tropical version of discrete logarithm. In this paper we will formally introduce the tropical two-sided discrete logarithm with shift, discuss how it is solved, and subsequently demonstrate an attack on a key exchange protocol based on the tropical semiring of pairs. This particular protocol is compromised due to the existence of efficient (albeit heuristic) solution of the tropical two-sided logarithm problem, and this highlights the ongoing challenges in search of a "good" key exchange protocol in tropical cryptography.

This paper reviews common attacks in classical cryptography and plausible attacks in the post-quantum era targeted at CRYSTALS-Kyber. Kyber is a recently standardized post-quantum cryptography scheme that relies on the hardness of lattice problems. Although it has undergone rigorous testing by the National Institute of Standards and Technology (NIST), there have recently been studies that have successfully executed attacks against Kyber while showing their applicability outside of controlled settings. The attacks discussed in the paper include common attacks, side-channel attacks, SCA-assisted CCA, and fault injections. In the common attacks section, attacks on symmetric primitives, multi-target attacks, and attacks exploiting decryption failures can all be deemed inviable, while recent data on attacks on module-LWE questions Kyber's security level. In the side-channel attacks section, timing attacks are proven useless due to the constant-time nature of Kyber, but SASCA attacks are still viable, though easily defended against with minimal drawbacks. Attacks targeting message encoding and attacks using deep learning, however, both prove effective, even with high-order masking. LDPC has also been proposed as a new framework for attack, proving itself potent with room for growth. In the SCA-assisted CCA section, EM attacks and CPA attacks have also both shown potential while remaining difficult to defend against. In the fault injection section, Roulette and error-tolerant key recovery have both recently been proposed, with data demonstrating their effectiveness and difficulty to defend against. This paper aims to provide future researchers insight into what areas should be focused on to strengthen current as well as future cryptosystems.

The semidirect discrete logarithm problem (SDLP) is the following analogue of the standard discrete logarithm problem in the semidirect product semigroup $G\rtimes \mathrm{End}(G)$ for a finite semigroup $G$. Given $g\in G, \sigma\in \mathrm{End}(G)$, and $h=\prod_{i=0}^{t-1}\sigma^i(g)$ for some integer $t$, the SDLP$(G,\sigma)$, for $g$ and $h$, asks to determine $t$. As Shor's algorithm crucially depends on commutativity, it is believed not to be applicable to the SDLP. Previously, the best known algorithm for the SDLP was based on Kuperberg's subexponential time quantum algorithm. Still, the problem plays a central role in the security of certain proposed cryptosystems in the family of $\textit{semidirect product key exchange}$. This includes a recently proposed signature protocol called SPDH-Sign. In this paper, we show that the SDLP is even easier in some important special cases. Specifically, for a finite group $G$, we describe quantum algorithms for the SDLP in $G\rtimes \mathrm{Aut}(G)$ for the following two classes of instances: the first one is when $G$ is solvable and the second is when $G$ is a matrix group and a power of $\sigma$ with a polynomially small exponent is an inner automorphism of $G$. We further extend the results to groups composed of factors from these classes. A consequence is that SPDH-Sign and similar cryptosystems whose security assumption is based on the presumed hardness of the SDLP in the cases described above are insecure against quantum attacks. The quantum ingredients we rely on are not new: these are Shor's factoring and discrete logarithm algorithms and well-known generalizations.

This paper presents an efficient attack that, in the standard IND-CCA2 attack model plus a one-time single-bit fault, recovers the NTRU-HRSS session key. This type of fault is expected to occur for many users through natural DRAM bit flips. In a multi-target IND-CCA2 attack model plus a one-time single-bit fault, the attack recovers every NTRU-HRSS session key that was encapsulated to the targeted public key before the fault. Software carrying out the full multi-target attack, using a simulated fault, is provided for verification. This paper also explains how a change in NTRU-HRSS in 2019 enabled this attack.

HALFLOOP-96 is a 96-bit tweakable block cipher used in high frequency radio to secure automatic link establishment messages. In this paper, we concentrate on its differential properties in the contexts of conventional, related-tweak, and related-key differential attacks. Using automatic techniques, we determine the minimum number of active S-boxes and the maximum differential probability in each of the three configurations. The resistance of HALFLOOP-96 to differential attacks in the conventional and related-tweak configurations is good, and the longest distinguishers in both configurations consist of five rounds. In contrast, the security of the cipher against differential attacks in the related-key configuration is inadequate. The most effective related-key distinguisher we can find spans eight rounds. The 8-round related-key differential distinguisher is then utilised to initiate a 9-round weak-key attack. With $2^{92.96}$ chosen-plaintexts, 38.77-bit equivalent information about the keys can be recovered. Even though the attack does not pose a significant security threat to HALFLOOP-96, its security margin in the related-key configuration is exceedingly narrow. Therefore, improper use must be avoided in the application.

Panny [3] described how to forge the “tropical signatures” proposed by Chen, Grigoriev and Shpilrain [1]. (These signatures are loosely related to the NP-complete problem of factoring tropical polynomials). We describe more methods to forge these tropical signatures. We also describe some patches that thwart all but one of these forgery methods (which we summarize as re-hashing an honest signature).

We present a side-channel attack on CRYSTALS-Dilithium, a post-quantum secure digital signature scheme, with two variants of post-processing. The side-channel attack exploits information leakage in the secret key unpacking procedure of the signing algorithm to recover the coefficients of the polynomials in the secret key vectors ${\bf s}_1$ and ${\bf s}_2$ by profiled deep learning-assisted power analysis. In the first variant, one half of the coefficients of ${\bf s}_1$ and ${\bf s}_2$ is recovered by power analysis and the rest is derived by solving a system of linear equations based on ${\bf t} = {\bf A}{\bf s}_1 + {\bf s}_2$, where ${\bf A}$ and ${\bf t}$ are parts of the public key. This case assumes knowledge of the least significant bits of the vector ${\bf t}$, ${\bf t}_0$. The second variant waives this requirement. However, to succeed, it needs a larger portion of ${\bf s}_1$ to be recovered by power analysis. The remainder of ${\bf s}_1$ is obtained by lattice reduction. Once the full ${\bf s}_1$ is recovered, all the other information necessary for generating valid signatures can be trivially derived from the public key. We evaluate both variants on an ARM Cortex-M4 implementation of Dilithium-2. The profiling stage (trace capture and neural network training) takes less than 10 hours. In the attack assuming that ${\bf t}_0$ is known, the probability of successfully recovering the full vector ${\bf s}_1$ from a single trace captured from a different from profiling device is non-negligible (9%). The success rate approaches 100% if multiple traces are available for the attack. Our results demonstrate the necessity of protecting the secret key of CRYSTALS-Dilithium from single-trace attacks and call for a reassessment of the role of compression of the public key vector ${\bf t}$ in the security of CRYSTALS-Dilithium implementations.

This work investigates the security of the Ascon authenticated encryption scheme in the context of fault attacks, with a specific focus on Differential Fault Analysis (DFA). Motivated by the growing significance of lightweight cryptographic solutions, particularly Ascon, we explore potential vulnerabilities in its design using DFA. By employing a novel approach that combines faulty forgery in the decryption query under two distinct fault models, leveraging bit-flip faults in the first phase and bit-set faults in the second, we successfully recover the complete Ascon key. This study sheds light on the impact of key whitening in the final permutation call and discusses potential threats when this safeguard is absent. Additionally, we consider the implications of injecting multiple bit-flip faults at the S-box input, suggesting alternative strategies for compromising the state space. Our findings contribute valuable insights into the gray-box security landscape of Ascon, emphasizing the need for robust defenses to ensure the integrity and resilience of lightweight cryptographic primitives against diverse fault attacks.

In recent years, deep learning-based side-channel analysis (DLSCA) has become an active research topic within the side-channel analysis community. The well-known challenge of hyperparameter tuning in DLSCA encouraged the community to use methods that reduce the effort required to identify an optimal model. One of the successful methods is ensemble learning. While ensemble methods have demonstrated their effectiveness in DLSCA, particularly with AES-based datasets, their efficacy in analyzing symmetric-key cryptographic primitives with different operational mechanics remains unexplored. Ascon was recently announced as the winner of the NIST lightweight cryptography competition. This will lead to broader use of Ascon and a crucial requirement for thorough side-channel analysis of its implementations. With these two considerations in view, we utilize an ensemble of deep neural networks to attack two implementations of Ascon. Using an ensemble of five multilayer perceptrons or convolutional neural networks, we could find the secret key for the Ascon-protected implementation with less than 3 000 traces. To the best of our knowledge, this is the best currently known result. We can also identify the correct key with less than 100 traces for the unprotected implementation of Ascon, which is on par with the state-of-the-art results.

While blockchain technologies leverage compelling characteristics in terms of decentralization, immutability, and transparency, user privacy in public blockchains remains a fundamental challenge that requires particular attention. This is mainly due to the history of all transactions being accessible and available to anyone, thus making it possible for an attacker to infer data about users that is supposed to remain private. In this paper, we provide a threat model of possible privacy attacks on users utilizing the Bitcoin blockchain. To this end, we followed the LINDDUN GO methodology to identify threats and suggest possible mitigation.

In this report, we perform an in-depth analysis of the RSA authentication feature used in the secure boot procedure of Xilinx Zynq-7000 SoC device. The First Stage Boot Loader (FSBL) is a critical piece of software executed during secure boot, which utilizes the RSA authentication feature to validate all the hardware and software partitions to be mounted on the device. We analyzed the implementation of FSBL (provided by Xilinx) for the Zynq-7000 SoC and identified a critical security flaw, whose exploitation makes it possible to load an unauthenticated application onto the Zynq device, thereby bypassing RSA authentication. We also experimentally validated the presence of the vulnerability through a Proof of Concept (PoC) attack to successfully mount an unauthenticated software application on an RSA authenticated Zynq device. The identified flaw is only present in the FSBL software and thus can be easily fixed through appropriate modification of the FSBL software. Thus, the first contribution of our work is the identification of a critical security flaw in the FSBL software to bypass RSA authentication. Upon bypassing RSA authentication, an attacker can mount any unauthenticated software application on the target device to mount a variety of attacks. Among the several possible attacks, we are interested to perform recovery of the encrypted bitstream in the target boot image of the Zynq-7000 device. To the best of our knowledge, there does not exist any prior work that has reported a practical bitstream recovery attack on the Zynq-7000 device. In the context of bitstream recovery, Ender et al. in 2020 proposed the Starbleed attack that is applicable to standalone Virtex-6 and 7-series Xilinx FPGAs. The design advisory provided by Xilinx as a response to the Starbleed attack claims that the Zynq-7000 SoC is resistant “due to the use of asymmetric and/or symmetric authentication in the boot/configuration process that ensures configuration is authenticated prior to use". Due to the security flaw found in the FSBL, we managed to identify a novel approach to mount the Starbleed attack on the Zynq-7000 device for full bitstream recovery. Thus, as a second contribution of our work, we present the first practical demonstration of the Starbleed attack on the Zynq-7000 SoC. We perform experimental validation of our proposed attacks on the PYNQ-Z1 platform based on the Zynq-7000 SoC.

The literature gives the impression that (1) existing heuristics accurately predict how effective lattice attacks are, (2) non-ternary lattice systems are not vulnerable to hybrid multi-decoding primal attacks, and (3) the asymptotic exponents of attacks against non-ternary systems have stabilized. This paper shows that 1 contradicts 2 and that 1 contradicts 3: the existing heuristics imply that hybrid primal key-recovery attacks are exponentially faster than standard non-hybrid primal key-recovery attacks against the LPR PKE with any constant error width. This is the first report since 2015 of an exponential speedup in heuristic non-quantum primal attacks against non-ternary LPR. Quantitatively, for dimension n, modulus n^{Q_0+o(1)}, and error width w, a surprisingly simple hybrid attack reduces heuristic costs from 2^{(ρ+o(1))n} to 2^{(ρ-ρ H_0+o(1))n}, where z_0=2Q_0/(Q_0+1/2)^2, ρ=z_0 log_4(3/2), and H_0=1/(1+(lg w)/0.057981z_0). This raises the questions of (1) what heuristic exponent is achieved by more sophisticated hybrid attacks and (2) what impact hybrid attacks have upon concrete cryptosystems whose security analyses have ignored hybrid attacks, such as Kyber-512.

During the standardisation process of post-quantum cryptography, NIST encourages research on side-channel analysis for candidate schemes. As the recommended lattice signature scheme, CRYSTALS-Dilithium, when implemented on hardware, has only been subjected to the side-channel attack presented by Steffen et al. in IACR ePrint 2022. This attack is not complete and requires excessive traces. Therefore, we investigate the leakage of an FPGA (Kintex7) implementation of CRYSTALS-Dilithium using the CPA method, where with a minimum of 70000 traces partial private key coefficients can be recovered. As far as we know, this is the first work that applies power leakage to sidechannel attacks on FPGA implementations of CRYSTALS-Dilithium. Furthermore, we optimise the attack by extracting Point-of-Interests using known information due to parallelism (named CPA-PoI) and by iteratively utilising parallel leakages (named CPA-ITR). We experimentally demonstrate that when recovering the same number of key coefficients, the CPA-PoI and CPA-ITR reduce the number of traces used by up to 16.67 percent and 25 percent, respectively, compared to the CPA method. When attacking with the same number of traces, the CPA-PoI method and the CPA-ITR method increase the number of recovered key coefficients by up to 55.17 percent and 93.10 percent, respectively, compared to the CPA method. Our experiments confirm that the FPGA implementation of CRYSTALS-Dilithium is also very vulnerable to side-channel analysis.

The cube attack is a powerful cryptanalysis technique against symmetric ciphers, especially stream ciphers. The adversary aims to recover secret key bits by solving equations that involve the key. To simplify the equations, a set of plaintexts called a cube is summed up together. Traditional cube attacks use only linear or quadratic superpolies, and the size of cube is limited to an experimental range, typically around 40. However, cube attack based on division property, proposed by Todo et al. at CRYPTO 2017, overcomes these limitations and enables theoretical cube attacks on many lightweight stream ciphers. For a given cube $I$, they evaluate the set $J$ of secret key bits involved in the superpoly and require $2^{|I|+|J|}$ encryptions to recover the superpoly. However, the secret variables evaluation method proposed by Todo et al. sometimes becomes unresponsive and fails to solve within a reasonable time. In this paper, we propose an improvement to Todo's method by breaking down difficult-to-solve problems into several smaller sub-problems. Our method retains the efficiency of Todo's method while effectively avoiding unresponsive situations. We apply our method to the WAGE cipher, an NLFSR-based authenticated encryption algorithm and one of the second round candidates in the NIST LWC competition. Specifically, we successfully mount cube attacks on 29-round WAGE, as well as on 24-round WAGE with a sponge constraint. To the best of our knowledge, this is the first cube attack against the WAGE cipher, which provides a more accurate characterization of the WAGE's resistance against algebraic attacks.

Deep Learning (DL) based Side-Channel Analysis (SCA) has been extremely popular recently. DL-based SCA can easily break implementations protected by masking countermeasures. DL-based SCA has also been highly successful against implementations protected by various trace desynchronization-based countermeasures like random delay, clock jitter, and shuffling. Over the years, many DL models have been explored to perform SCA. Recently, Transformer Network (TN) based model has also been introduced for SCA. Though the previously introduced TN-based model is successful against implementations jointly protected by masking and random delay countermeasures, it is not scalable to long traces (having a length greater than a few thousand) due to its quadratic time and memory complexity. This work proposes a novel shift-invariant TN-based model with linear time and memory complexity. The contributions of the work are two-fold. First, we introduce a novel TN-based model called EstraNet for SCA. EstraNet has linear time and memory complexity in trace length, significantly improving over the previously proposed TN-based model’s quadratic time and memory cost. EstraNet is also shift-invariant, making it highly effective against countermeasures like random delay and clock jitter. Secondly, we evaluated EstraNet on three SCA datasets of masked implementations with random delay and clock jitter effects. Our experimental results show that EstraNet significantly outperforms several benchmark models, demonstrating up to an order of magnitude reduction in the number of attack traces required to reach guessing entropy 1.