International Association
for Cryptologic Research

HALFLOOP is a family of tweakable block ciphers that are used for encrypting automatic link establishment (ALE) messages in high frequency radio, a technology commonly used by the military, other government agencies and industries which require high robustness in long-distance communications. Recently, it was shown in [DDLS22] that the smallest version of the cipher, HALFLOOP-24, can be attacked within a practical time and memory complexity. However, in the real-word ALE setting, it turns out that this attack require to wait more than 500 years to collect the necessary amount of plaintext-tweak-ciphertext pairs fulfilling the conditions of the attack.In this paper, we present real-world practical attacks against HALFLOOP-24 which are based on a probability-one differential distinguisher. In our attacks, we significantly reduce the data complexity to three differential pairs in the chosen-plaintext (CPA) setting which is optimal in the sense that even a brute force attack needs at least six plaintext-tweak-ciphertext pairs to uniquely identify the correct key. Considering the same ALE setting as [DDLS22], this translates to a reduction from 541 years to 2 hours worth of intercepted traffic.Besides, we provide the first, non generic, public cryptanalysis of HALFLOOP-48 and HALFLOOP-96. More precisely, we present Demirci-Selçuk meet-in-the-middle attacks against full-round HALFLOOP-48 and round-reduced HALFLOOP-96 to recover the complete master key in a CPA setting. However, unlike the attacks on HALFLOOP-24, our attacks on the larger versions are only theoretical. Moreover for HALFLOOP-96 the known generic time-memory trade-off attack, based on a flawed tweak handling, remains the strongest attack vector.In conclusion, we iterate what was already stated in [DDLS22]: HALFLOOP does not provide adequate protection and should not be used.

Motivated by progress in the field of zero-knowledge proofs, so-called Arithmetization-Oriented (AO) symmetric primitives have started to appear in the literature, such as MiMC, Poseidon or Rescue. Due to the design constraints implied by this setting, these algorithms are defined using simple operations over large (possibly prime) fields. In particular, many rely on simple low-degree monomials for their non-linear layers, essentially using x ↦ x3 as an S-box.In this paper, we show that the structure of the material injected in each round (be it subkeys in a block cipher or round constants in a public permutation) could allow a specific pattern, whereby a well-defined affine space is mapped to another by the round function, and then to another, etc. Such chains of one-dimensional subspaces always exist over 2 rounds, and they can be extended to an arbitrary number of rounds, for any linear layer, provided that the round-constants are well chosen.As a consequence, for several ciphers like Rescue, or a variant of AES with a monomial Sbox, there exist some round-key sequences for which the cipher has an abnormally high differential uniformity, exceeding the size of the Sbox alphabet.Well-known security arguments, in particular based on the wide-trail strategy, have been reused in the AO setting by many designers. Unfortunately, our results show that such a traditional study may not be sufficient to guarantee security. To illustrate this, we present two new primitives (the tweakable block cipher Snare and the permutation-based hash function Stir) that are built using state-of-the-art security arguments, but which are actually deeply flawed. Indeed, the key schedule of Snare ensures the presence of a subspace chain that significantly simplifies an algebraic attack against it, and the round constants of Stir force the presence of a subspace chain aligned with the rate and capacity of the permutation. This in turns implies the existence of many easy-to-find solutions to the so-called CICO problem.

When outsourcing a database to an untrusted remote server, one might want to verify the integrity of contents while accessing it. To solve this, Blum et al. [FOCS `91] propose the notion of \emph{memory checking}. Memory checking allows a user to run a RAM program on a remote server, with the ability to verify integrity of the storage with small private storage. In this work, we define and initiate the formal study of memory checking for \emph{Parallel RAMs} (PRAMs). The parallel RAM model is very expressive and captures many modern architectures such as multi-core architectures and cloud clusters. When multiple clients run a PRAM algorithm on a shared remote server, it is possible that there are concurrency issues that cause inconsistencies. Therefore, integrity verification is also a desirable property in this setting. We construct an online memory checker (one that reports faults as soon as they occur) for PRAMs with $O(\log N)$ simulation overhead in both work and depth. Moreover, we construct an offline memory checker (one that reports faults only after a long sequence of operations) with amortized $O(1)$ simulation overhead in both work and depth. As an application of our parallel memory checking constructions, we construct a \emph{maliciously secure oblivious parallel RAM} (OPRAM) with polylogarithmic overhead.

We introduce a new lattice basis reduction algorithm with approximation guarantees analogous to the LLL algorithm and practical performance that far exceeds the current state of the art. We achieve these results by iteratively applying precision management techniques within a recursive algorithm structure and show the stability of this approach. We analyze the asymptotic behavior of our algorithm, and show that the heuristic running time is $O(n^{\omega}(C+n)^{1+\varepsilon})$ for lattices of dimension $n$, $\omega\in (2,3]$ bounding the cost of size reduction, matrix multiplication, and QR factorization, and $C$ bounding the log of the condition number of the input basis $B$. This yields a running time of $O\left(n^\omega (p + n)^{1 + \varepsilon}\right)$ for precision $p = O(\log \|B\|_{max})$ in common applications. Our algorithm is fully practical, and we have published our implementation. We experimentally validate our heuristic, give extensive benchmarks against numerous classes of cryptographic lattices, and show that our algorithm significantly outperforms existing implementations.

We prove adaptive security of a simple three-round threshold Schnorr signature scheme, which we call Sparkle. The standard notion of security for threshold signatures considers a static adversary - one who must declare which parties are corrupt at the beginning of the protocol. The stronger adaptive adversary can at any time corrupt parties and learn their state. This notion is natural and practical, yet not proven to be met by most schemes in the literature. In this paper, we demonstrate that Sparkle achieves several levels of security based on different corruption models and assumptions. To begin with, Sparkle is statically secure under minimal assumptions: the discrete logarithm assumption (DL) and the random oracle model (ROM). If an adaptive adversary corrupts fewer than t/2 out of a threshold of t+1 signers, then Sparkle is adaptively secure under a weaker variant of the one-more discrete logarithm assumption (AOMDL) in the ROM. Finally, we prove that Sparkle achieves full adaptive security, with a corruption threshold of t, under AOMDL in the algebraic group model (AGM) with random oracles. Importantly, we show adaptive security without requiring secure erasures. Ours is the first proof achieving full adaptive security without exponential tightness loss for any threshold Schnorr signature scheme; moreover, the reduction is tight.

Kyber is a key-encapsulation mechanism (KEM) that was recently selected by NIST in its PQC standardization process; it is also the only scheme to be selected in the context of public-key encryption (PKE) and key establishment. The main security target for KEMs, and their associated PKE schemes, in the NIST PQC context has been IND-CCA security. However, some important modern applications also require their underlying KEMs/PKE schemes to provide anonymity (Bellare et al., ASIACRYPT 2001). Examples of such applications include anonymous credential systems, cryptocurrencies, broadcast encryption schemes, authenticated key exchange, and auction protocols. It is hence important to analyze the compatibility of NIST's new PQC standard in such "beyond IND-CCA" applications. Some starting steps were taken by Grubbs et al. (EUROCRYPT 2022) and Xagawa (EUROCRYPT 2022) wherein they studied the anonymity properties of most NIST PQC third round candidate KEMs. Unfortunately, they were unable to show the anonymity of Kyber because of certain technical barriers. In this paper, we overcome said barriers and resolve the open problems posed by Grubbs et al.(EUROCRYPT 2022) and Xagawa (EUROCRYPT 2022) by establishing the anonymity of Kyber, and the (hybrid) PKE schemes derived from it, in a post-quantum setting. Along the way, we also provide an approach to obtain tight IND-CCA security proofs for Kyber with concrete bounds; this resolves another issue identified by the aforementioned works related to the post-quantum IND-CCA security claims of Kyber from a provable security point-of-view. Our results also extend to Saber, a NIST PQC third round finalist, in a similar fashion.

In recent work, Backendal, Haller, and Paterson identified several exploitable vulnerabilities in the cloud storage provider MEGA. They demonstrated an RSA key recovery attack in which a malicious server could recover a client's private RSA key after 512 client login attempts. We show how to exploit additional information revealed by MEGA's protocol vulnerabilities to give an attack that requires only six client logins to recover the secret key. Our optimized attack combines several cryptanalytic techniques. In particular, we formulate and give a solution to a variant of the hidden number problem with small unknown multipliers, which may be of independent interest. We show that our lattice construction for this problem can be used to give improved results for the implicit factorization problem of May and Ritzenhofen.

We present an attack on SIDH utilising isogenies between polarized products of two supersingular elliptic curves. In the case of arbitrary starting curve, our attack (discovered independently from [8]) has subexponential complexity, thus significantly reducing the security of SIDH and SIKE. When the endomorphism ring of the starting curve is known, our attack (here derived from [8]) has polynomial-time complexity assuming the generalised Riemann hypothesis. Our attack applies to any isogeny-based cryptosystem that publishes the images of points under the secret isogeny, for example Séta [13] and B-SIDH [11]. It does not apply to CSIDH [9], CSI-FiSh [3], or SQISign [14].

In this work, we will give new attacks on the pseudorandomness of algebraic pseudorandom number generators (PRGs) of polynomial stretch. Our algorithms apply to a broad class of PRGs and are in the case of general local PRGs faster than currently known attacks. At the same time, in contrast to most algebraic attacks, subexponential time and space bounds will be proven for our attacks without making any assumptions of the PRGs or assuming any further conjectures. Therefore, we yield in this text the first subexponential distinguishing attacks on PRGs from constant-degree polynomials and close current gaps in the subexponential cryptoanalysis of lightweight PRGs. Concretely, against PRGs $F : \mathbb{Z}_q^{n} \rightarrow \mathbb{Z}_q^{m}$ that are computed by polynomials of degree $d$ over a field $\mathbb{Z}_q$ and have a stretch of $m = n^{1+e}$ we give an attack with space and time complexities $n^{O(n^{1 - \frac{e}{d-1}})}$ and noticeable advantage $1 - {O(n^{1 - \frac{e}{d-1}}/{q})}$. If $q$ lies in $O(n^{1 - \frac{e}{d-1}})$, we give a second attack with the same space and time complexities whose advantage is at least $q^{-O(n^{1 - \frac{e}{d-1}})}$. If $F$ is of constant \emph{locality} $d$ and $q$ is constant, we construct a third attack that has a space and time complexity of $\exp(O(n^{1 - \frac{e'}{(q-1)d-1}}))$ and noticeable advantage $1-O(n^{-\frac{e'}{(q-1)d-1}})$ for every constant $e' < e$.

We present an efficient key recovery attack on the Supersingular Isogeny Diffie-Hellman protocol (SIDH). The attack is based on Kani's "reducibility criterion" for isogenies from products of elliptic curves and strongly relies on the torsion point images that Alice and Bob exchange during the protocol. If we assume knowledge of the endomorphism ring of the starting curve then the classical running time is polynomial in the input size (heuristically), apart from the factorization of a small number of integers that only depend on the system parameters. The attack is particularly fast and easy to implement if one of the parties uses 2-isogenies and the starting curve comes equipped with a non-scalar endomorphism of very small degree; this is the case for SIKE, the instantiation of SIDH that recently advanced to the fourth round of NIST's standardization effort for post-quantum cryptography. Our Magma implementation breaks SIKEp434, which aims at security level 1, in about ten minutes on a single core.

We show that we can break SIDH in (classical) polynomial time, even with a random starting curve~$E_0$.

Cryptographic group actions are a relaxation of standard cryptographic groups that have less structure. This lack of structure allows them to be plausibly quantum resistant despite Shor's algorithm, while still having a number of applications. The most famous example of group actions are built from isogenies on elliptic curves. Our main result is that CDH for abelian group actions is quantumly equivalent to discrete log. Galbraith et al. (Mathematical Cryptology) previously showed perfectly solving CDH to be equivalent to discrete log quantumly; our result works for any non-negligible advantage. We also explore several other questions about group action and isogeny protocols.

Hashing arbitrary values to points on an elliptic curve is a required step in many cryptographic constructions, and a number of techniques have been proposed to do so over the years. One of the first ones was due to Shallue and van de Woestijne (ANTS-VII), and it had the interesting property of applying to essentially all elliptic curves over finite fields. It did not, however, have the desirable property of being *indifferentiable from a random oracle* when composed with a random oracle to the base field. Various approaches have since been considered to overcome this limitation, starting with the foundational work of Brier et al. (CRYPTO 2011). For example, if f: F_q→E(F_q) is the Shallue--van de Woestijne (SW) map and H, H' are *two* independent random oracles, we now know that m↦f(H(m))+f(H'(m)) is indifferentiable from a random oracle. Unfortunately, this approach has the drawback of being twice as expensive to compute than the straightforward, but not indifferentiable, m↦f(H(m)). Most other solutions so far have had the same issue: they are at least as costly as two base field exponentiations, whereas plain encoding maps like f cost only one exponentiation. Recently, Koshelev (DCC 2022) provided the first construction of indifferentiable hashing at the cost of one exponentiation, but only for a very specific class of curves (some of those with j-invariant 0), and using techniques that are unlikely to apply more broadly. In this work, we revisit this long-standing open problem, and observe that the SW map actually fits in a one-parameter family (f_u)_{u∈F_q} of encodings, such that for independent random oracles H, H', F: m↦f_{H'(m)}(H(m)) is indifferentiable. Moreover, on a very large class of curves (essentially those that are either of odd order or of order divisible by 4), the one-parameter family admits a rational parametrization, which lets us compute F at almost the same cost as small f, and finally achieve indifferentiable hashing to most curves with a single exponentiation. Our new approach also yields an improved variant of the Elligator Squared technique of Tibouchi (FC 2014) that represents points of arbitrary elliptic curves as close-to-uniform random strings.

A hinting PRG is a (potentially) stronger variant of PRG with a "deterministic" form of circular security with respect to the seed of the PRG (Koppula and Waters, CRYPTO 2019). Hinting PRGs enable many cryptographic applications, most notably CCA-secure public-key encryption and trapdoor functions. In this paper, we study cryptographic primitives with the hinting property, yielding the following results: - We present a novel and conceptually simpler approach for designing hinting PRGs from certain decisional assumptions over cyclic groups or isogeny-based group actions, which enables simpler security proofs as compared to the existing approaches for designing such primitives. - We introduce hinting weak PRFs, a natural extension of the hinting property to weak PRFs, and show how to realize circular/KDM-secure symmetric-key encryption from any hinting weak PRF. We demonstrate that our simple approach for building hinting PRGs can be extended to realize hinting weak PRFs from the same set of decisional assumptions. - We propose a stronger version of the hinting property, which we call the functional hinting property, that guarantees security even in the presence of hints about functions of the secret seed/key. We show how to instantiate functional hinting PRGs and functional hinting weak PRFs for certain (families of) functions by building upon our simple techniques for realizing plain hinting PRGs/weak PRFs. We also demonstrate the applicability of a functional hinting weak PRF with certain algebraic properties in realizing KDM-secure public-key encryption in a black-box manner. - Finally, we show the first black-box separation between hinting weak PRFs (and hinting PRGs) from public-key encryption using simple realizations of these primitives given only a random oracle.

Assume that distributions X_0,X_1 (respectively Y_0,Y_1) are d_X (respectively d_Y) indistinguishable for circuits of a given size. It is well known that the product distributions X_0Y_0,X_1Y_1 are d_X+d_Y indistinguishable for slightly smaller circuits. However, in probability theory where unbounded adversaries are considered through statistical distance, it is folklore knowledge that in fact X_0Y_0 and X_1Y_1 are d_x+d_y-d_x*d_y indistinguishable, and also that this bound is tight. We formulate and prove the computational analog of this tight bound. Our proof is entirely different from the proof in the statistical case, which is non-constructive. As a corollary, we show that if X and Y are d indistinguishable, then k independent copies of X and k independent copies of Y are almost 1-(1-d)^k indistinguishable for smaller circuits, as against d*k using the looser bound. Our bounds are useful in settings where only weak (i.e. non-negligible) indistinguishability is guaranteed. We demonstrate this in the context of cryptography, showing that our bounds yield simple analysis for amplification of weak oblivious transfer protocols.

Non-interactive batch arguments for NP provide a way to amortize the cost of NP verification across multiple instances. They enable a prover to convince a verifier of multiple NP statements with communication much smaller than the total witness length and verification time much smaller than individually checking each instance. In this work, we give the first construction of a non-interactive batch argument for NP from standard assumptions on groups with bilinear maps (specifically, from either the subgroup decision assumption in composite-order groups or from the k-Lin assumption in prime-order groups for any k >= 1). Previously, batch arguments for NP were only known from LWE, or a combination of multiple assumptions, or from non-standard/non-falsifiable assumptions. Moreover, our work introduces a new direct approach for batch verification and avoids heavy tools like correlation-intractable hash functions or probabilistically-checkable proofs common to previous approaches. As corollaries to our main construction, we obtain the first publicly-verifiable non-interactive delegation scheme for RAM programs (i.e., a succinct non-interactive argument (SNARG) for P) with a CRS of sublinear size (in the running time of the RAM program), as well as the first aggregate signature scheme (supporting bounded aggregation) from standard assumptions on bilinear maps.

In this article, we generalize the works of Pan et al. (Eurocrypt'21) and Porter et al. (ArXiv'21) and provide a simple condition under which an ideal lattice defines an easy instance of the shortest vector problem. Namely, we show that the more automorphisms stabilize the ideal, the easier it is to find a short vector in it. This observation was already made for prime ideals in Galois fields, and we generalize it to any ideal (whose prime factors are not ramified) of any number field. We then provide a cryptographic application of this result by showing that particular instances of the partial Vandermonde knapsack problem, also known as partial Fourier recovery problem, can be solved classically in polynomial time. As a proof of concept, we implemented our attack and managed to solve those particular instances for concrete parameter settings proposed in the literature. For random instances, we can halve the lattice dimension with non-negligible probability.

This work introduces new key recovery attacks against the Rainbow signature scheme, which is one of the three finalist signature schemes still in the NIST Post-Quantum Cryptography standardization project. The new attacks dramatically outperform previously known attacks for all the parameter sets submitted to NIST and make a key-recovery practical for the SL 1 parameters. Concretely, given a Rainbow public key for the SL 1 parameters of the second-round submission, our attack returns the corresponding public key after on average 53 hours (one weekend) of computation time on a standard laptop.

Garbled RAM (GRAM) is a powerful technique introduced by Lu and Ostrovsky that equips Garbled Circuit (GC) with a sublinear cost RAM without adding rounds of interaction. While multiple GRAM constructions are known, none are suitable for practice, due to costs that have high constants and poor scaling. We present the first GRAM suitable for practice. For computational security parameter $\kappa$ and for a size-$n$ RAM that stores blocks of size $w = \Omega(\log^2 n)$ bits, our GRAM incurs only amortized $O(w \cdot \log^2 n \cdot \kappa)$ communication and computation per access. We evaluate the concrete cost of our GRAM; our approach outperforms trivial linear-scan-based RAM for as few as $512$ $128$-bit elements.

This paper documents a practical security evaluation of the Tesla Model X keyless entry system. In contrast to other works, the keyless entry system analysed in this paper employs secure symmetric-key and public-key cryptographic primitives implemented by a Common Criteria certified Secure Element. We document the internal workings of this system, covering the key fob, the body control module and the pairing protocol. Additionally, we detail our reverse engineering techniques and document several security issues. The identified issues in the key fob firmware update mechanism and the key fob pairing protocol allow us to bypass all of the cryptographic security measures put in place. To demonstrate the practical impact of our research we develop a fully remote Proof-of-Concept attack that allows to gain access to the vehicle’s interior in a matter of minutes and pair a modified key fob, allowing to drive off. Our attack is not a relay attack, as our new key fob allows us to start the car anytime anywhere. Finally, we provide an analysis of the update performed by Tesla to mitigate our findings. Our work highlights how the increased complexity and connectivity of vehicular systems can result in a larger and easier to exploit attack surface.

We describe a garbling scheme for boolean circuits, in which XOR gates are free and AND gates require communication of $1.5\kappa + 5$ bits. This improves over the state-of-the-art ``half-gates'' scheme of Zahur, Rosulek, and Evans (Eurocrypt 2015), in which XOR gates are free and AND gates cost $2\kappa$ bits. The half-gates paper proved a lower bound of $2\kappa$ bits per AND gate, in a model that captured all known garbling techniques at the time. We bypass this lower bound with a novel technique that we call \textbf{slicing and dicing}, which involves slicing wire labels in half and operating separately on those halves. Ours is the first to bypass the lower bound while being fully compatible with free-XOR, making it a drop-in replacement for half-gates. Our construction is proven secure from a similar assumption to prior free-XOR garbling (circular correlation-robust hash), and uses only slightly more computation than half-gates.

Improved attacks on generic small-domain Feistel ciphers with alternating round tweaks are obtained using linear cryptanalysis. This results in practical distinguishing and message-recovery attacks on the United States format-preserving encryption standard FF3-1 and the South-Korean standards FEA-1 and FEA-2. The data-complexity of the proposed attacks on FF3-1 and FEA-1 is $O(N^{r/2 - 1.5})$, where $N^2$ is the domain size and $r$ is the number of rounds. For example, FF3-1 with $N = 10^3$ can be distinguished from an ideal tweakable block cipher with advantage $\ge 1/10$ using $2^{23}$ encryption queries. Recovering the left half of a message with similar advantage requires $2^{24}$ data. The analysis of FF3-1 serves as an interesting real-world application of (generalized) linear cryptanalysis over the group $\mathbb{Z}/N\mathbb{Z}$.

Liu and Pass (FOCS'20) recently demonstrated an equivalence between the existence of one-way functions and mild average-case hardness of the time-bounded Kolmogorov complexity problem. In this work, we establish a similar equivalence but to a different form of time-bounded Kolmogorov Complexity---namely, Levin's notion of Kolmogorov Complexity---whose hardness is closely related to the problem of whether $\EXP \neq \BPP$. In more detail, let $Kt(x)$ denote the Levin-Kolmogorov Complexity of the string $x$; that is, $Kt(x) = \min_{\desc \in \bitset^*, t \in \N}\{|\desc| + \lceil \log t \rceil: U(\desc, 1^t) = x\}$, where $U$ is a universal Turing machine, and let $\mktp$ denote the language of pairs $(x,k)$ having the property that $Kt(x) \leq k$. We demonstrate that: - $\mktp$ is \emph{two-sided error} mildly average-case hard (i.e., $\mktp \notin \HeurpBPP$) iff infinititely-often one-way functions exist. - $\mktp$ is \emph{errorless} mildly average-case hard (i.e., $\mktp \notin \AvgpBPP$) iff $\EXP \neq \BPP$. Thus, the only ``gap'' towards getting (infinitely-often) one-way functions from the assumption that $\EXP \neq \BPP$ is the seemingly ``minor'' technical gap between two-sided error and errorless average-case hardness of the $\mktp$ problem. As a corollary of this result, we additionally demonstrate that any reduction from errorless to two-sided error average-case hardness for $\mktp$ implies (unconditionally) that $\NP \neq \P$. We finally consider other alternative notions of Kolmogorov complexity---including space-bounded Kolmogorov complexity and conditional Kolmogorov complexity---and show how average-case hardness of problems related to them characterize log-space computable one-way functions, or one-way functions in $\NC^0$.

The HFE cryptosystem is one of the best known multivariate schemes. Especially in the area of digital signatures, the HFEv- variant offers short signatures and high performance. Recently, an instance of the HFEv- signature scheme called GeMSS was elected as one of the alternative candidates for signature schemes in the third round of the NIST Post Quantum Crypto (PQC) Standardization Project. In this paper, we propose a new key recovery attack on the HFEv- signature scheme. Our attack shows that both the Minus and the Vinegar modifi- cation do not enhance the security of the basic HFE scheme significantly. This shows that it is very difficult to build a secure and efficient signature scheme on the basis of HFE. In particular, we use our attack to show that the proposed parameters of the GeMSS scheme are not as secure as claimed.

In this paper we present a new representation of the AES key schedule, with some implications to the security of AES-based schemes. In particular, we show that the AES-128 key schedule can be split into four independent parallel computations operating on 32 bits chunks, up to linear transformation. Surprisingly, this property has not been described in the literature after more than 20 years of analysis of AES. We show two consequences of our new representation, improving previous cryptanalysis results of AES-based schemes. First, we observe that iterating an odd number of key schedule rounds results in a function with short cycles. This explains an observation of Khairallah on mixFeed, a second-round candidate in the NIST lightweight competition. Our analysis actually shows that his forgery attack on mixFeed succeeds with probability 0.44 (with data complexity 220GB), breaking the scheme in practice. The same observation also leads to a novel attack on ALE, another AES-based AEAD scheme. Our new representation also gives efficient ways to combine information from the first sub-keys and information from the last sub-keys, in order to reconstruct the corresponding master keys. In particular we improve previous impossible-differential attacks against AES-128.

We provide the first constructions of non-interactive zero-knowledge and Zap arguments for NP based on the sub-exponential hardness of Decisional Diffie-Hellman against polynomial time adversaries (without use of groups with pairings). Central to our results, and of independent interest, is a new notion of interactive trapdoor hashing protocols.

We present an algorithm solving the ROS (Random inhomogeneities in a Overdetermined Solvable system of linear equations) problem mod p in polynomial time for $l > log p$ dimensions. Our algorithm can be combined with Wagner's attack, and leads to a sub-exponential solution for any dimension $l$ with best complexity known so far. When concurrent executions are allowed, our algorithm leads to practical attacks against unforgeability of blind signature schemes such as Schnorr and Okamoto--Schnorr blind signatures, threshold signatures such as GJKR and the original version of FROST, multisignatures such as CoSI and the two-round version of MuSig, partially blind signatures such as Abe--Okamoto, and conditional blind signatures such as ZGP17. Schemes for e-cash and anonymous credentials (such as Anonymous Credentials Light) inspired from the above are also affected.

Since the celebrated work of Impagliazzo and Rudich (STOC 1989), a number of black-box impossibility results have been established. However, these works only ruled out classical black-box reductions among cryptographic primitives. Therefore it may be possible to overcome these impossibility results by using quantum reductions. To exclude such a possibility, we have to extend these impossibility results to the quantum setting. In this paper, we study black-box impossibility in the quantum setting. We first formalize a quantum counterpart of fully-black-box reduction following the formalization by Reingold, Trevisan and Vadhan (TCC 2004). Then we prove that there is no quantum fully-black-box reduction from collision-resistant hash functions to one-way permutations (or even trapdoor permutations). We take both of classical and quantum implementations of primitives into account. This is an extension to the quantum setting of the work of Simon (Eurocrypt 1998) who showed a similar result in the classical setting.

Gimli is a family of cryptographic primitives (both a hash function and an AEAD scheme) that has been selected for the second round of the NIST competition for standardizing new lightweight designs. The candidate Gimli is based on the permutation Gimli, which was presented at CHES 2017. In this paper, we study the security of both the permutation and the constructions that are based on it. We exploit the slow diffusion in Gimli and its internal symmetries to build, for the first time, a distinguisher on the full permutation of complexity $2^{64}$. We also provide a practical distinguisher on 23 out of the full 24 rounds of Gimli that has been implemented. Next, we give (full state) collision and semi-free-start collision attacks on Gimli-Hash, reaching respectively up to 12 and 18 rounds. On the practical side, we compute a collision on 8-round Gimli-Hash. In the quantum setting, these attacks reach 2 more rounds. Finally, we perform the first study of linear trails in the permutation, and we propose differential-linear cryptanalysis that reach up to 17 rounds of Gimli.

We introduce a new signature scheme, \emph{SQISign}, (for \emph{Short Quaternion and Isogeny Signature}) from isogeny graphs of supersingular elliptic curves. The signature scheme is derived from a new one-round, high soundness, interactive identification protocol. Targeting the post-quantum NIST-1 level of security, our implementation results in signatures of $204$ bytes, secret keys of $16$ bytes and public keys of $64$ bytes. In particular, the signature and public key sizes combined are an order of magnitude smaller than all other post-quantum signature schemes. On a modern workstation, our implementation in C takes 0.6s for key generation, 2.5s for signing, and 50ms for verification. While the soundness of the identification protocol follows from classical assumptions, the zero-knowledge property relies on the second main contribution of this paper. We introduce a new algorithm to find an isogeny path connecting two given supersingular elliptic curves of known endomorphism rings. A previous algorithm to solve this problem, due to Kohel, Lauter, Petit and Tignol, systematically reveals paths from the input curves to a `special' curve. This leakage would break the zero-knowledge property of the protocol. Our algorithm does not directly reveal such a path, and subject to a new computational assumption, we prove that the resulting identification protocol is zero-knowledge.

ISO/IEC 9797-1 is an international standard for block-cipher-based Message Authentication Code (MAC). The current version ISO/IEC 9797-1:2011 specifies six single-pass CBC-like MAC structures that are capped at the birthday bound security. For a higher security that is beyond-birthday bound, it recommends to use the concatenation combiner of two single-pass MACs. In this paper, we reveal the invalidity of the suggestion, by presenting a birthday bound forgery attack on the concatenation combiner, which is essentially based on Joux’s multi-collision. Notably, our new forgery attack for the concatenation of two MAC Algorithm 1 with padding scheme 2 only requires 3 queries. Moreover, we look for patches by revisiting the development of ISO/IEC 9797-1 with respect to the beyond-birthday bound security. More specifically, we evaluate the XOR combiner of single-pass CBC-like MACs, which was used in previous version of ISO/IEC 9797-1.

We present our discovery of a group of side-channel vulnerabilities in implementations of the ECDSA signature algorithm in a widely used Atmel AT90SC FIPS 140-2 certified smartcard chip and five cryptographic libraries (libgcrypt, wolfSSL, MatrixSSL, SunEC/OpenJDK/Oracle JDK, Crypto++). Vulnerable implementations leak the bit-length of the scalar used in scalar multiplication via timing. Using leaked bit-length, we mount a lattice attack on a 256-bit curve, after observing enough signing operations. We propose two new methods to recover the full private key requiring just 500 signatures for simulated leakage data, 1200 for real cryptographic library data, and 2100 for smartcard data. The number of signatures needed for a successful attack depends on the chosen method and its parameters as well as on the noise profile, influenced by the type of leakage and used computation platform. We use the set of vulnerabilities reported in this paper, together with the recently published TPM-FAIL vulnerability [MSE+20] as a basis for real-world benchmark datasets to systematically compare our newly proposed methods and all previously published applicable lattice-based key recovery methods. The resulting exhaustive comparison highlights the methods’ sensitivity to its proper parametrization and demonstrates that our methods are more efficient in most cases. For the TPM-FAIL dataset, we decreased the number of required signatures from approximately 40 000 to mere 900.

We provide a new definitional framework capturing the multi-user security of encryption schemes and pseudorandom functions in the face of adversaries that can adaptively compromise users' keys. We provide a sequence of results establishing the security of practical symmetric encryption schemes under adaptive compromise in the random oracle or ideal cipher model. The bulk of analysis complexity for adaptive compromise security is relegated to the analysis of lower-level primitives such as pseudorandom functions. We apply our framework to give proofs of security for the BurnBox system for privacy in the face of border searches and the in-use searchable symmetric encryption scheme due to Cash et al. In both cases, prior analyses had bugs that our framework helps avoid.

We present several improvements to the framework of differential-linear attacks with a special focus on ARX ciphers. As a demonstration of their impact, we apply them to Chaskey and ChaCha and we are able to significantly improve upon the best attacks published so far.

We provide a construction of chosen ciphertext secure public-key encryption from (injective) trapdoor functions. Our construction is black box and assumes no special properties (e.g. ``lossy'', ``correlated product secure'') of the trapdoor function.

In this paper, we use genus theory to analyze the hardness of the decisional Diffie--Hellman problem (DDH) for ideal class groups of imaginary quadratic orders, acting on sets of elliptic curves through isogenies; such actions are used in the Couveignes--Rostovtsev--Stolbunov protocol and in CSIDH. Concretely, genus theory equips every imaginary quadratic order $\mathcal{O}$ with a set of assigned characters $\chi : \text{cl}(\mathcal{O}) \to \{ \pm 1 \}$, and for each such character and every secret ideal class $[\mathfrak{a}]$ connecting two public elliptic curves $E$ and $E' = [\mathfrak{a}] \star E$, we show how to compute $\chi([\mathfrak{a}])$ given only $E$ and $E'$, i.e., without knowledge of $[\mathfrak{a}]$. In practice, this breaks DDH as soon as the class number is even, which is true for a density $1$ subset of all imaginary quadratic orders. For instance, our attack works very efficiently for all supersingular elliptic curves over $\mathbb{F}_p$ with $p \equiv 1 \bmod 4$. Our method relies on computing Tate pairings and walking down isogeny volcanoes.

We present the first protocols for private information retrieval that allow fast (sublinear-time) database lookups without increasing the server-side storage requirements. To achieve these efficiency goals, our protocols work in an offline/online model. In an offline phase, which takes place before the client has decided which database bit it wants to read, the client fetches a short string from the servers. In a subsequent online phase, the client can privately retrieve its desired bit of the database by making a second query to the servers. By pushing the bulk of the server-side computation into the offline phase (which is independent of the client’s query), our protocols allow the online phase to complete very quickly—in time sublinear in the size of the database. Our protocols can provide statistical security in the two-server setting and computational security in the single-server setting. Finally, we prove that, in this model, our protocols are optimal in terms of the trade-off they achieve between communication and running time.

Boneh, Waters and Zhandry (CRYPTO 2014) used multilinear maps to provide a solution to the long-standing problem of public-key broadcast encryption (BE) where all parameters in the system are small. In this work, we improve their result by providing a solution that uses only {\it bilinear} maps and Learning With Errors (LWE). Our scheme is fully collusion-resistant against any number of colluders, and can be generalized to an identity-based broadcast system with short parameters. Thus, we reclaim the problem of optimal broadcast encryption from the land of ``Obfustopia''. Our main technical contribution is a ciphertext policy attribute based encryption (CP-ABE) scheme which achieves special efficiency properties -- its ciphertext size, secret key size, and public key size are all independent of the size of the circuits supported by the scheme. We show that this special CP-ABE scheme implies BE with optimal parameters; but it may also be of independent interest. Our constructions rely on a novel interplay of bilinear maps and LWE, and are proven secure in the generic group model.

We present here a new family of trapdoor one-way functions that are Preimage Sampleable on Average (PSA) based on codes, the Wave-PSA family. The trapdoor function is one-way under two computational assumptions: the hardness of generic decoding for high weights and the indistinguishability of generalized $$(U,U+V)$$-codes. Our proof follows the GPV strategy [28]. By including rejection sampling, we ensure the proper distribution for the trapdoor inverse output. The domain sampling property of our family is ensured by using and proving a variant of the left-over hash lemma. We instantiate the new Wave-PSA family with ternary generalized $$(U,U+V)$$-codes to design a “hash-and-sign” signature scheme which achieves existential unforgeability under adaptive chosen message attacks (EUF-CMA) in the random oracle model.

The task of function inversion is central to cryptanalysis: breaking block ciphers, forging signatures, and cracking password hashes are all special cases of the function-inversion problem. In 1980, Hellman showed that it is possible to invert a random function $$f{:}\,[N] \rightarrow [N]$$ in time $$T = \widetilde{O}(N^{2/3})$$ given only $$S = \widetilde{O}(N^{2/3})$$ bits of precomputed advice about f. Hellman’s algorithm is the basis for the popular “Rainbow Tables” technique (Oechslin 2003), which achieves the same asymptotic cost and is widely used in practical cryptanalysis.Is Hellman’s method the best possible algorithm for inverting functions with preprocessed advice? The best known lower bound, due to Yao (1990), shows that $$ST = \widetilde{\Omega }(N)$$, which still admits the possibility of an $$S = T = \widetilde{O}(N^{1/2})$$ attack. There remains a long-standing and vexing gap between Hellman’s $$N^{2/3}$$ upper bound and Yao’s $$N^{1/2}$$ lower bound. Understanding the feasibility of an $$S = T = N^{1/2}$$ algorithm is cryptanalytically relevant since such an algorithm could perform a key-recovery attack on AES-128 in time $$2^{64}$$ using a precomputed table of size $$2^{64}$$.For the past 29 years, there has been no progress either in improving Hellman’s algorithm or in strengthening Yao’s lower bound. In this work, we connect function inversion to problems in other areas of theory to (1) explain why progress may be difficult and (2) explore possible ways forward.Our results are as follows:We show that any improvement on Yao’s lower bound on function-inversion algorithms will imply new lower bounds on depth-two circuits with arbitrary gates. Further, we show that proving strong lower bounds on non-adaptive function-inversion algorithms would imply breakthrough circuit lower bounds on linear-size log-depth circuits.We take first steps towards the study of the injective function-inversion problem, which has manifold cryptographic applications. In particular, we show that improved algorithms for breaking PRGs with preprocessing would give improved algorithms for inverting injective functions with preprocessing.Finally, we show that function inversion is closely related to well-studied problems in communication complexity and data structures. Through these connections we immediately obtain the best known algorithms for problems in these domains.

Implementing the masking countermeasure in hardware is a delicate task. Various solutions have been proposed for this purpose over the last years: we focus on Threshold Implementations (TIs), Domain-Oriented Masking (DOM), the Unified Masking Approach (UMA) and Generic Low Latency Masking (GLM). The latter generally come with innovative ideas to cope with physical defaults such as glitches. Yet, and in contrast to the situation in software-oriented masking, these schemes have not been formally proven at arbitrary security orders and their composability properties were left unclear. So far, only a 2-cycle implementation of the seminal masking scheme by Ishai, Sahai and Wagner has been shown secure and composable in the robust probing model – a variation of the probing model aimed to capture physical defaults such as glitches – for any number of shares.In this paper, we argue that this lack of proofs for TIs, DOM, UMA and GLM makes the interpretation of their security guarantees difficult as the number of shares increases. For this purpose, we first put forward that the higher-order variants of all these schemes are affected by (local or composability) security flaws in the (robust) probing model, due to insufficient refreshing. We then show that composability and robustness against glitches cannot be analyzed independently. We finally detail how these abstract flaws translate into concrete (experimental) attacks, and discuss the additional constraints robust probing security implies on the need of registers. Despite not systematically leading to improved complexities at low security orders, e.g., with respect to the required number of measurements for a successful attack, we argue that these weaknesses provide a case for the need of security proofs in the robust probing model (or a similar abstraction) at higher security orders.

We present practical attacks on OCB2. This mode of operation of a blockcipher was designed with the aim to provide particularly efficient and provably-secure authenticated encryption services, and since its proposal about 15 years ago it belongs to the top performers in this realm. OCB2 was included in an ISO standard in 2009.An internal building block of OCB2 is the tweakable blockcipher obtained by operating a regular blockcipher in $$ \text {XEX} ^*$$ mode. The latter provides security only when evaluated in accordance with certain technical restrictions that, as we note, are not always respected by OCB2. This leads to devastating attacks against OCB2’s security promises: We develop a range of very practical attacks that, amongst others, demonstrate universal forgeries and full plaintext recovery. We complete our report with proposals for (provably) repairing OCB2. To our understanding, as a direct consequence of our findings, OCB2 is currently in a process of removal from ISO standards. Our attacks do not apply to OCB1 and OCB3, and our privacy attacks on OCB2 require an active adversary.

We introduce models of computation that enable direct comparisons between classical and quantum algorithms. Incorporating previous work on quantum computation and error correction, we justify the use of the gate-count and depth-times-width cost metrics for quantum circuits. We demonstrate the relevance of these models to cryptanalysis by revisiting, and increasing, the security estimates for the Supersingular Isogeny Diffie–Hellman (SIDH) and Supersingular Isogeny Key Encapsulation (SIKE) schemes. Our models, analyses, and physical justifications have applications to a number of memory intensive quantum algorithms.

Attribute-based Encryption (ABE), first introduced by [SW05, GPSW06], is a public key encryption system that can support multiple users with varying decryption permissions. One of the main properties of such schemes is the supported function class of policies. While there are fully secure constructions from bilinear maps for a fairly large class of policies, the situation with lattice-based constructions is less satisfactory and many efforts were made to close this gap. Prior to this work the only known fully secure lattice construction was for the class of point functions (also known as IBE).In this work we construct for the first time a lattice-based (ciphertext-policy) ABE scheme for the function class t-CNF, which consists of CNF formulas where each clause depends on at most t bits of the input, for any constant t. This class includes NP-verification policies, bit-fixing policies and t-threshold policies. Towards this goal we also construct a fully secure single-key constrained PRF from OWF for the same function class, which might be of independent interest.

We construct a verifiable delay function (VDF). A VDF is a function whose evaluation requires running a given number of sequential steps, yet the result can be efficiently verified. They have applications in decentralised systems, such as the generation of trustworthy public randomness in a trustless environment, or resource-efficient blockchains. To construct our VDF, we actually build a trapdoor VDF. A trapdoor VDF is essentially a VDF which can be evaluated efficiently by parties who know a secret (the trapdoor). By setting up this scheme in a way that the trapdoor is unknown (not even by the party running the setup, so that there is no need for a trusted setup environment), we obtain a simple VDF. Our construction is based on groups of unknown order such as an RSA group, or the class group of an imaginary quadratic field. The output of our construction is very short (the result and the proof of correctness are each a single element of the group), and the verification of correctness is very efficient.

Public key quantum money can be seen as a version of the quantum no-cloning theorem that holds even when the quantum states can be verified by the adversary. In this work, we investigate quantum lightning where no-cloning holds even when the adversary herself generates the quantum state to be cloned. We then study quantum money and quantum lightning, showing the following results:We demonstrate the usefulness of quantum lightning beyond quantum money by showing several potential applications, such as generating random strings with a proof of entropy, to completely decentralized cryptocurrency without a block-chain, where transactions is instant and local.We give Either/Or results for quantum money/lightning, showing that either signatures/hash functions/commitment schemes meet very strong recently proposed notions of security, or they yield quantum money or lightning. Given the difficulty in constructing public key quantum money, this suggests that natural schemes do attain strong security guarantees.We show that instantiating the quantum money scheme of Aaronson and Christiano [STOC’12] with indistinguishability obfuscation that is secure against quantum computers yields a secure quantum money scheme. This construction can be seen as an instance of our Either/Or result for signatures, giving the first separation between two security notions for signatures from the literature.Finally, we give a plausible construction for quantum lightning, which we prove secure under an assumption related to the multi-collision resistance of degree-2 hash functions. Our construction is inspired by our Either/Or result for hash functions, and yields the first plausible standard model instantiation of a non-collapsing collision resistant hash function. This improves on a result of Unruh [Eurocrypt’16] which is relative to a quantum oracle.

A new approach to invariant subspaces and nonlinear invariants is developed. This results in both theoretical insights and practical attacks on block ciphers. It is shown that, with minor modifications to some of the round constants, Midori-64 has a nonlinear invariant with $$2^{96}$$ corresponding weak keys. Furthermore, this invariant corresponds to a linear hull with maximal correlation. By combining the new invariant with integral cryptanalysis, a practical key-recovery attack on 10 rounds of unmodified Midori-64 is obtained. The attack works for $$2^{96}$$ weak keys and irrespective of the choice of round constants. The data complexity is $$1.25 \cdot 2^{21}$$ chosen plaintexts and the computational cost is dominated by $$2^{56}$$ block cipher calls. Finally, it is shown that similar techniques lead to a practical key-recovery attack on MANTIS-4. The full key is recovered using 640 chosen plaintexts and the attack requires about $$2^{56}$$ block cipher calls.

An Oblivious RAM (ORAM) introduced by Goldreich and Ostrovsky [JACM’96] is a (possibly randomized) RAM, for which the memory access pattern reveals no information about the operations performed. The main performance metric of an ORAM is the bandwidth overhead, i.e., the multiplicative factor extra memory blocks that must be accessed to hide the operation sequence. In their seminal paper introducing the ORAM, Goldreich and Ostrovsky proved an amortized $$\varOmega (\lg n)$$ bandwidth overhead lower bound for ORAMs with memory size n. Their lower bound is very strong in the sense that it applies to the “offline” setting in which the ORAM knows the entire sequence of operations ahead of time.However, as pointed out by Boyle and Naor [ITCS’16] in the paper “Is there an oblivious RAM lower bound?”, there are two caveats with the lower bound of Goldreich and Ostrovsky: (1) it only applies to “balls in bins” algorithms, i.e., algorithms where the ORAM may only shuffle blocks around and not apply any sophisticated encoding of the data, and (2), it only applies to statistically secure constructions. Boyle and Naor showed that removing the “balls in bins” assumption would result in super linear lower bounds for sorting circuits, a long standing open problem in circuit complexity. As a way to circumventing this barrier, they also proposed a notion of an “online” ORAM, which is an ORAM that remains secure even if the operations arrive in an online manner. They argued that most known ORAM constructions work in the online setting as well.Our contribution is an $$\varOmega (\lg n)$$ lower bound on the bandwidth overhead of any online ORAM, even if we require only computational security and allow arbitrary representations of data, thus greatly strengthening the lower bound of Goldreich and Ostrovsky in the online setting. Our lower bound applies to ORAMs with memory size n and any word size $$r \ge 1$$ . The bound therefore asymptotically matches the known upper bounds when $$r = \varOmega (\lg ^2 n)$$ .

Non-interactive zero-knowledge (NIZK) proofs are fundamental to modern cryptography. Numerous NIZK constructions are known in both the random oracle and the common reference string (CRS) models. In the CRS model, there exist constructions from several classes of cryptographic assumptions such as trapdoor permutations, pairings, and indistinguishability obfuscation. Notably absent from this list, however, are constructions from standard lattice assumptions. While there has been partial progress in realizing NIZKs from lattices for specific languages, constructing NIZK proofs (and arguments) for all of $$\mathsf {NP}$$ from standard lattice assumptions remains open.   In this work, we make progress on this problem by giving the first construction of a multi-theorem NIZK argument for $$\mathsf {NP}$$ from standard lattice assumptions in the preprocessing model. In the preprocessing model, a (trusted) setup algorithm generates proving and verification keys. The proving key is needed to construct proofs and the verification key is needed to check proofs. In the multi-theorem setting, the proving and verification keys should be reusable for an unbounded number of theorems without compromising soundness or zero-knowledge. Existing constructions of NIZKs in the preprocessing model (or even the designated-verifier model) that rely on weaker assumptions like one-way functions or oblivious transfer are only secure in a single-theorem setting. Thus, constructing multi-theorem NIZKs in the preprocessing model does not seem to be inherently easier than constructing them in the CRS model.   We begin by constructing a multi-theorem preprocessing NIZK directly from context-hiding homomorphic signatures. Then, we show how to efficiently implement the preprocessing step using a new cryptographic primitive called blind homomorphic signatures. This primitive may be of independent interest. Finally, we show how to leverage our new lattice-based preprocessing NIZKs to obtain new malicious-secure MPC protocols purely from standard lattice assumptions.

The possibility of basing cryptography on the minimal assumption $$\mathbf{NP }\nsubseteq \mathbf{BPP }$$ NP⊈BPP is at the very heart of complexity-theoretic cryptography. The closest we have gotten so far is lattice-based cryptography whose average-case security is based on the worst-case hardness of approximate shortest vector problems on integer lattices. The state-of-the-art is the construction of a one-way function (and collision-resistant hash function) based on the hardness of the $$\tilde{O}(n)$$ O~(n)-approximate shortest independent vector problem $${\textsf {SIVP}}_{\tilde{O}(n)}$$ SIVPO~(n).Although $${\textsf {SIVP}}$$ SIVP is NP-hard in its exact version, Guruswami et al. (CCC 2004) showed that $${\textsf {gapSIVP}}_{\sqrt{n/\log n}}$$ gapSIVPn/logn is in $$\mathbf{NP } \cap \mathbf{coAM }$$ NP∩coAM and thus unlikely to be $$\mathbf{NP }$$ NP-hard. Indeed, any language that can be reduced to $${\textsf {gapSIVP}}_{\tilde{O}(\sqrt{n})}$$ gapSIVPO~(n) (under general probabilistic polynomial-time adaptive reductions) is in $$\mathbf{AM } \cap \mathbf{coAM }$$ AM∩coAM by the results of Peikert and Vaikuntanathan (CRYPTO 2008) and Mahmoody and Xiao (CCC 2010). However, none of these results apply to reductions to search problems, still leaving open a ray of hope: can $$\mathbf{NP }$$ NPbe reduced to solving search SIVP with approximation factor $$\tilde{O}(n)$$ O~(n)?We eliminate such possibility, by showing that any language that can be reduced to solving search $${\textsf {SIVP}}$$ SIVP with any approximation factor $$\lambda (n) = \omega (n\log n)$$ λ(n)=ω(nlogn) lies in AM intersect coAM.

In this work, we consider the ring- and module- variants of the LWE problem and investigate cold boot attacks on cryptographic schemes based on these problems, wherein an attacker is faced with the problem of recovering a scheme’s secret key from a noisy version of that key. The leakage resilience of cryptography based on the learning with errors (LWE) problem has been studied before, but there are only limited results considering the parameters observed in cold boot attack scenarios. There are two main encodings for storing ring- and module-LWE keys, and, as we show, the performance of cold boot attacks can be highly sensitive to the exact encoding used. The first encoding stores polynomial coefficients directly in memory. The second encoding performs a number theoretic transform (NTT) before storing the key, a commonly used method leading to more efficient implementations. We first give estimates for a cold boot attack complexity on the first encoding method based on standard algorithms; this analysis confirms that this encoding method is vulnerable to cold boot attacks only at very low bit-flip rates. We then show that, for the second encoding method, the structure introduced by using an NTT is exploitable in the cold boot setting: we develop a bespoke attack strategy that is much cheaper than our estimates for the first encoding when considering module-LWE keys. For example, at a 1% bit-flip rate (which corresponds roughly to what can be achieved in practice for cold boot attacks when applying cooling), a cold boot attack on Kyber KEM parameters has a cost of 243 operations when the second, NTT-based encoding is used for key storage, compared to 270 operations with the first encoding. On the other hand, in the case of the ring-LWE-based KEM, New Hope, the cold boot attack complexities are similar for both encoding methods.

This paper presents a cryptanalysis of full Kravatte, an instantiation of the Farfalle construction of a pseudorandom function (PRF) with variable input and output length. This new construction, proposed by Bertoni et al., introduces an efficiently parallelizable and extremely versatile building block for the design of symmetric mechanisms, e.g. message authentication codes or stream ciphers. It relies on a set of permutations and on so-called rolling functions: it can be split into a compression layer followed by a two-step expansion layer. The key is expanded and used to mask the inputs and outputs of the construction. Kravatte instantiates Farfalle using linear rolling functions and permutations obtained by iterating the Keccak round function.We develop in this paper several attacks against this PRF, based on three different attack strategies that bypass part of the construction and target a reduced number of permutation rounds. A higher order differential distinguisher exploits the possibility to build an affine space of values in the cipher state after the compression layer. An algebraic meet-in-the-middle attack can be mounted on the second step of the expansion layer. Finally, due to the linearity of the rolling function and the low algebraic degree of the Keccak round function, a linear recurrence distinguisher can be found on intermediate states of the second step of the expansion layer. All the attacks rely on the ability to invert a small number of the final rounds of the construction. In particular, the last two rounds of the construction together with the final masking by the key can be algebraically inverted, which allows to recover the key.The complexities of the devised attacks, applied to the Kravatte specifications published on the IACR ePrint in July 2017, or the strengthened version of Kravatte recently presented at ECC 2017, are far below the security claimed.

Synchrotron-based X-ray nanobeams are investigated as a tool to perturb microcontroller circuits. An intense hard X-ray focused beam of a few tens of nanometers is used to target the flash, EEPROM and RAM memory of a circuit. The obtained results show that it is possible to corrupt a single transistor in a semi-permanent state. A simple heat treatment can remove the induced effect, thus making the corruption reversible. An attack on a code stored in flash demonstrates unambiguously that this new technique can be a threat to the security of integrated circuits.