International Association
for Cryptologic Research

We propose two efficient public key encryption (PKE) schemes satisfying key dependent message security against chosen ciphertext attacks (KDM-CCA security). The first one is KDM-CCA secure with respect to affine functions. The other one is KDM-CCA secure with respect to polynomial functions. Both of our schemes are based on the KDM-CPA secure PKE schemes proposed by Malkin, Teranishi, and Yung (EUROCRYPT 2011). Although our schemes satisfy KDM-CCA security, their efficiency overheads compared to Malkin et al.'s schemes are very small. Thus, efficiency of our schemes is drastically improved compared to the existing KDM-CCA secure schemes.

We achieve our results by extending the construction technique by Kitagawa and Tanaka (ASIACRYPT 2018). Our schemes are obtained via semi-generic constructions using an IND-CCA secure PKE scheme as a building block. We prove the KDM-CCA security of our schemes based on the decisional composite residuosity (DCR) assumption and the IND-CCA security of the building block PKE scheme.

Moreover, our security proofs are tight if the IND-CCA security of the building block PKE scheme is tightly reduced to its underlying computational assumption. By instantiating our schemes using existing tightly IND-CCA secure PKE schemes, we obtain the first tightly KDM-CCA secure PKE schemes whose ciphertext consists only of a constant number of group elements.

Event date: 14 April to 16 April 2020
Submission deadline: 1 December 2019
Notification: 15 January 2020

The arbitrary-centered discrete Gaussian sampler is a fundamental subroutine in implementing lattice trapdoor sampling algorithms. However, existing approaches typically rely on either a fast implementation of another discrete Gaussian sampler or pre-computations with regards to some specific discrete Gaussian distributions with fixed centers and standard deviations. These approaches may only support sampling from standard deviations within a limited range, or cannot efficiently sample from arbitrary standard deviations determined on-the-fly at run-time.

In this paper, we propose a compact and scalable rejection sampling algorithm by sampling from a continuous normal distribution and performing rejection sampling on rounded samples. Our scheme does not require pre-computations related to any specific discrete Gaussian distributions. Our scheme can sample from both arbitrary centers and arbitrary standard deviations determined on-the-fly at run-time. In addition, we show that our scheme only requires a low number of trials close to 2 per sample on average, and our scheme maintains good performance when scaling up the standard deviation. We also provide a concrete error analysis of our scheme based on the Renyi divergence.

In a lockable obfuscation scheme a party takes as input a program $P$, a lock value $\alpha$, a message $m$ and produces an obfuscated program $\tilde{P}$. The obfuscated program can be evaluated on an input $x$ to learn the message $m$ if $P(x)= \alpha$. The security of such schemes states that if $\alpha$ is randomly chosen (independent of $P$ and $m$), then one cannot distinguish an obfuscation of $P$ from a ``dummy'' obfuscation. Existing constructions of lockable obfuscation achieve provable security under the Learning with Errors assumption. One limitation of these constructions is that they achieve only statistical correctness and allow for a possible one sided error where the obfuscated program could output the $m$ on some value $x$ where $P(x) \neq \alpha$.

In this work we motivate the problem of studying perfect correctness in lockable obfuscation for the case where the party performing the obfuscation might wish to inject a backdoor or hole in correctness. We begin by studying the existing constructions and identify two components that are susceptible to imperfect correctness. The first is in the LWE-based pseudo random generators (PRGs) that are non-injective, while the second is in the last level testing procedure of the core constructions.

We address each in turn. First, we build upon previous work to design injective PRGs that are provably secure from the LWE assumption. Next, we design an alternative last level testing procedure that has additional structure to prevent correctness errors. We then provide a surgical proof of security (to avoid redundancy) that connects our construction to the construction by Goyal, Koppula, and Waters (GKW). Specifically, we show how for a random value $\alpha$ an obfuscation under our new construction is indistinguishable from an obfuscation under the existing GKW construction.

We introduce stochastic sandpile models which imitate numerous aspects of the practical behavior of the LLL algorithm with compelling accuracy. In addition, we argue that the physics and mathematics of sandpile models provide satisfactory heuristic explanations to much of the mysteries of LLL, and pleasant implications for lattice-based cryptography as a whole. Based on these successes, we suggest a paradigm in which one regards blockwise reduction algorithms as 1-d stochastic self-organized criticality(SOC) models and study them as such.

We take advantage of a recently published open source implementation of the AES protected with a mix of countermeasures against side-channel attacks to discuss both the challenges in protecting COTS devices against such attacks and the limitations of closed source security evaluations. The target implementation has been proposed by the French ANSSI (Agence Nationale de la Sécurité des Systèmes d'Information) to stimulate research on the design and evaluation of side-channel secure implementations. It combines additive and multiplicative secret sharings into an affine masking scheme that is additionally mixed with a shuffled execution. Its preliminary leakage assessment did not detect data dependencies with up to 100,000 measurements. We first exhibit the gap between such a preliminary leakage assessment and advanced attacks by exhibiting how a countermeasures' dissection exploiting a mix of dimensionality reduction, multivariate information extraction and key enumeration can recover the full key with less than 2,000 measurements. We then discuss the relevance of open source evaluations to analyze such implementations efficiently, by exhibiting that certain steps of the attack are hard to automate without implementation knowledge (even with machine learning tools), while performing them manually is trivial. Our findings are not due to design flaws but from the general difficulty to prevent side-channel attacks in COTS devices with limited noise. We anticipate that high security on such devices requires significantly more shares.

We propose SPAE, a single pass, patent free, authenticated encryption with associated data (AEAD) for AES. The algorithm has been developped to address the needs of a growing trend in IoT systems: storing code and data on a low cost flash memory external to the main SOC. Existing AEAD algorithms such as OCB, GCM, CCM, EAX , SIV, provide the required functionality however in practice each of them suffer from various drawbacks for this particular use case. Academic contributions such as ASCON and AEGIS-128 are suitable and efficient however they require the development of new hardware accelerators and they use primitives which are not ‘approved’ by governemental institutions such as NIST, BSI, ANSSI. From a silicon manufacturer point of view, an efficient AEAD which use existing AES hardware is much more enticing: the AES is required already by most industry standards invovling symmetric encryption (GSMA, EMVco, FIDO, Bluetooth, ZigBee to name few). This paper expose the properties of an ideal AEAD for external memory encryption, present the SPAE algorithm and analyze various security aspects. Performances of SPAE on actual hardware are better than OCB, GCM and CCM.

The LSA cryptosystem is an asymmetric encryption algorithm which is based on both group and number theory that follows Kerckhoffs’s principle and relies on a specific case of Gauss’s Generalization of Wilson’s Theorem. Unlike prime factorization based algorithms, the eavesdropping cryptanalyst has no indication that he has successfully decrypted the ciphertext. For this reason, we aim to show that LSAis not only more secure than existing asymmetric algorithms but has the potential to be significantly computationally faster.

Event date: 18 March to 20 March 2020
Submission deadline: 1 November 2019
Notification: 15 November 2019

Threshold Implementations (TI) are secure algorithmic countermeasures against side-channel attacks in the form of differential power analysis. The strength of TI lies in its minimal algorithmic requirements. These requirements have been studied over more than 10 years and many efficient implementations for symmetric primitives have been proposed. Thus, over the years the practice of protecting implementations matured, however, the theory behind threshold implementations remained the same. In this work, we revise this theory by looking at the properties of correctness, non-completeness, and uniformity as a composable security model. We prove that this model provides first-order and higher-order univariate security in the glitch-robust probing model which lets us expand the theoretic framework of TI. We first provide a link between uniformity and the notion of non-interference, a known composable security notion building out the probing model. We then relax the notion of non-completeness which helps the design of secure expansion and compression functions. Lastly, we provide generalisations of the threshold notions to allow for general secret sharing schemes and provide examples of how different sharing schemes affect the security and efficiency of the countermeasure.

Highly efficient encryption and authentication of short messages is an essential requirement for enabling security in constrained scenarios such as the CAN FD in automotive systems (max. message size 64 bytes), massive IoT, critical communication domains of 5G, and Narrowband IoT, to mention a few. In addition, one of the NIST lightweight cryptography project requirements is that AEAD schemes shall be “optimized to be efficient for short messages (e.g., as short as 8 bytes)”.

In this work we introduce and formalize a novel primitive in symmetric cryptography called a forkcipher. A forkcipher is a keyed function expanding a fixed-length input to a fixed-length output. We define its security as indistinguishability under chosen ciphertext attack. We give a generic construction validation via the new iterate-fork-iterate design paradigm.

We then propose ForkSkinny as a concrete forkcipher instance with a public tweak and based on SKINNY: a tweakable lightweight block cipher constructed using the TWEAKEY framework. We conduct extensive cryptanalysis of ForkSkinny against classical and structure-specific attacks.

We demonstrate the applicability of forkciphers by designing three new provably-secure, nonce-based AEAD modes which offer performance and security tradeoffs and are optimized for efficiency of very short messages. Considering a reference block size of 16 bytes, and ignoring possible hardware optimizations, our new AEAD schemes beat the best SKINNY-based AEAD modes. More generally, we show forkciphers are suited for lightweight applications dealing with predominately short messages, while at the same time allowing handling arbitrary messages sizes.

Furthermore, our hardware implementation results show that when we exploit the inherent parallelism of ForkSkinny we achieve the best performance when directly compared with the most efficient mode instantiated with the SKINNY block cipher.

Elliptic curves are typically defined by Weierstrass equations. Given a kernel, the well-known Velu’s formula shows how to explicitly write down an isogeny between Weierstrass curves. However, it is not clear how to do the same on other forms of elliptic curves without isomorphisms mapping to and from the Weierstrass form. Previous papers have shown some isogeny formulas for (twisted) Edwards, Huff, and Montgomery forms of elliptic curves. Continuing this line of work, this paper derives an explicit formula for isogenies between elliptic curves in (twisted) Hessian form.

In order to study the resistance of a block cipher against boomerang attacks, a tool called the Boomerang Connectivity Table (BCT) for S-boxes was recently introduced. Very little is known today about the properties of this table especially for bijective S-boxes defined for $n$ variables with $n\equiv 0 \mod{4}$. In this work we study the boomerang uniformity of some popular constructions used for building large S-boxes, e.g. for 8 variables, from smaller ones. We show that the BCTs of all the studied constructions have abnormally high values in some positions. This remark permits us in some cases to link the boomerang properties of an S-box with other well-known cryptanalytic techniques on such constructions while in other cases it leads to the discovery of new ones. A surprising outcome concerns notably the Feistel and MISTY networks. While these two structures are very similar, their boomerang uniformity can be very different. In a second time, we investigate the boomerang uniformity under EA-equivalence for Gold and the inverse function (as used respectively in MPC-friendly ciphers and the AES) and we prove that the boomerang uniformity is EA-invariant in these cases. Finally, we present an algorithm for inverting a given BCT and provide experimental results on the size of the BCT-equivalence classes for some $4$ and $8$-bit S-boxes.

At CRYPTO 2017, Rosca et al. introduce a new variant of the Learning With Errors (LWE) problem, called the Middle-Product LWE (MP-LWE). The hardness of this new assumption is based on the hardness of the Polynomial LWE (P-LWE) problem parameterized by a set of polynomials, making it more secure against the possible weakness of a single defining polynomial. As a cryptographic application, they also provide an encryption scheme based on the MP-LWE problem. In this paper, we propose a deterministic variant of their encryption scheme, which does not need Gaussian sampling and is thus simpler than the original one. Still, it has the same quasi-optimal asymptotic key and ciphertext sizes. The main ingredient for this purpose is the Learning With Rounding (LWR) problem which has already been used to derandomize LWE type encryption. The hardness of our scheme is based on a new assumption called Middle-Product Computational Learning With Rounding, an adaption of the computational LWR problem over rings, introduced by Chen et al. at ASIACRYPT 2018. We prove that this new assumption is as hard as the decisional version of MP-LWE and thus benefits from worst-case to average-case hardness guarantees.

We study the security of symmetric primitives against key-correlated attacks (KCA), whereby an adversary can arbitrarily correlate keys, messages, and ciphertexts. Security against KCA is required whenever a primitive should securely encrypt key-dependent data, even when it is used under related keys. KCA is a strengthening of the previously considered notions of related-key attack (RKA) and key-dependent message (KDM) security. This strengthening is strict, as we show that 2-round Even–Mansour fails to be KCA secure even though it is both RKA and KDM secure. We provide feasibility results in the ideal-cipher model for KCAs and show that 3-round Even–Mansour is KCA secure under key offsets in the random-permutation model. We also give a natural transformation that converts any authenticated encryption scheme to a KCA-secure one in the random-oracle model. Conceptually, our results allow for a unified treatment of RKA and KDM security in idealized models of computation.

In different contexts such as filtered LFSR, Goldreich's PRG, and FLIP stream ciphers, the security of a cryptographic primitive mostly depends on the algebraic properties of one Boolean function. Since the Seventies, more and more efficient attacks have been exhibited in this context, related to more and more general algebraic properties, such as the degree, the algebraic immunity, and finally, the fast algebraic immunity. Once the properties to estimate the attack complexities are identified, it remains to determine the exact parameters of interesting families of functions with these properties. Then, these functions can be combined in secondary constructions to guarantee the good algebraic properties of a main function. In particular, the family of symmetric functions, and more precisely the subclass of majority functions, has been intensively studied in the area of cryptography, because of their practical advantages and good properties.

The degree of all these functions is known, and they have been proven to reach the optimal algebraic immunity, but still very few is known about its fast algebraic immunity. For a function in $n=2^m+j$ variables, an upper bound is known for all $m$ and $j$, proving that these functions do not reach the optimal fast algebraic immunity. However, the exact fast algebraic immunity is known only for very few families indexed by $j$, where the parameter is exhibited for all members of the family since $m$ is big enough. Recent works gave exact values for $j=0$ and $j=1$ (in the first case), and for $j=2$ and $j=3$ with $m\geq2$ (in the second case). In this work, we determine the exact fast algebraic immunity for all possible values of $j$, for all member of the family assuming $m\geq 1+\log_2(j+1)$.

Two of the most sought-after properties of Multi-party Computation (MPC) protocols are fairness and guaranteed output delivery (GOD), the latter also referred to as robustness. Achieving both, however, brings in the necessary requirement of malicious-minority. In a generalised adversarial setting where the adversary is allowed to corrupt both actively and passively, the necessary bound for a $n$-party fair or robust protocol turns out to be $t_a + t_p < n$, where $t_a,t_p$ denote the threshold for active and passive corruption with the latter subsuming the former. Subsuming the malicious-minority as a boundary special case, this setting, denoted as dynamic corruption, opens up a range of possible corruption scenarios for the adversary. While dynamic corruption includes the entire range of thresholds for $(t_a,t_p)$ starting from $(\lceil \frac{n}{2} \rceil - 1 , \lfloor \frac{n}{2} \rfloor)$ to $(0,n-1)$, the boundary corruption restricts the adversary only to the boundary cases of $(\lceil \frac{n}{2} \rceil - 1, \lfloor \frac{n}{2} \rfloor)$ and $(0,n-1)$. Notably, both corruption settings empower an adversary to control majority of the parties, yet ensuring the count on active corruption never goes beyond $\lceil \frac{n}{2} \rceil - 1$.

We target the round complexity of fair and robust MPC tolerating dynamic and boundary adversaries. As it turns out, $\lceil \frac{n}{2} \rceil + 1$ rounds are necessary and sufficient for fair as well as robust MPC tolerating dynamic corruption. The non-constant barrier raised by dynamic corruption can be sailed through for a boundary adversary. The round complexity of $3$ and $4$ is necessary and sufficient for fair and GOD protocols respectively, with the latter having an exception of allowing $3$ round protocols in the presence of a single active corruption. While all our lower bounds assume pair-wise private and broadcast channels and are resilient to the presence of both public (CRS) and private (PKI) setup, our upper bounds are broadcast-only and assume only public setup. The traditional and popular setting of malicious-minority, being restricted compared to both dynamic and boundary setting, requires $3$ and $2$ rounds in the presence of public and private setup respectively for both fair as well as GOD protocols.

The Fiat-Shamir transform is an incredibly powerful technique that uses a suitable hash function to reduce the interaction of general public-coin protocols. Unfortunately, there are known counterexamples showing that this methodology may not be sound (no matter what concrete hash function is used). Still, these counterexamples are somewhat unsatisfying, as the underlying protocols were specifically tailored to make Fiat-Shamir fail. This raises the question of whether this transform is sound when applied to natural protocols.

One of the most important protocol for which we would like to reduce interaction is Kilian’s four-message argument system for all of NP, based on collision resistant hash functions (CRHF) and probabilistically checkable proofs (PCPs). Indeed, an application of the Fiat-Shamir transform to Kilian's protocol is at the heart of both theoretical results (e.g., Micali's CS proofs) as well as leading practical approaches of highly efficient non-interactive proof-systems (e.g., SNARKs and STARKs).

In this work, we show significant obstacles to establishing soundness of (what we refer to as) the "Fiat-Shamir-Kilian-Micali" (FSKM) protocol. More specifically:

- We construct a (contrived) CRHF for which FSKM is unsound for a very large class of PCPs and for any Fiat-Shamir hash function. The collision-resistance of our CRHF relies on very strong but plausible cryptographic assumptions. The statement is "tight" in the following sense: any PCP outside the scope of our result trivially implies a SNARK, eliminating the need for FSKM in the first place.

- Second, we consider a known extension of Kilian’s protocol to an interactive variant of PCPs called probabilistically checkable interactive proofs (PCIP) (also known as interactive oracle proofs or IOPs). We construct a particular (contrived) PCIP for NP for which the FSKM protocol is unsound no matter what CRHF and Fiat-Shamir hash function is used. This result is unconditional (i.e., does not rely on any cryptographic assumptions).

Put together, our results show that the soundness of FSKM must rely on some special structure of both the CRHF and PCP that underlie Kilian's protocol. We believe these negative results may cast light on how to securely instantiate the FSKM protocol by a synergistic choice of the PCP, CRHF, and Fiat-Shamir hash function.

Modern cryptography requires the ability to securely generate pseudorandom numbers. However, despite decades of work on side channel attacks, there is little discussion of their application to pseudorandom number generators (PRGs). In this work we set out to address this gap, empirically evaluating the side channel resistance of common PRG implementations.

We find that hard-learned lessons about side channel leakage from encryption primitives have not been applied to PRGs, at all levels of abstraction. At the design level, the NIST-recommended CTR_DRBG design does not have forward security if an attacker is able to compromise the state via a side-channel attack. At the primitive level, popular implementations of CTR_DRBG such as OpenSSL's FIPS module and NetBSD's kernel use leaky T-table AES as their underlying block cipher, enabling cache side-channel attacks. Finally, we find that many implementations make parameter choices that enable an attacker to fully exploit the side-channel attack in a realistic scenario and recover secret keys from TLS connections.

We empirically demonstrate our attack in two scenarios. In the first, we carry out an asynchronous cache attack that recovers the private state from vulnerable CTR_DRBG implementations under realistic conditions to recover long-term authentication keys when the attacker is a party in the TLS connection. In the second scenario, we show that an attacker can exploit the high temporal resolution provided by Intel SGX to carry out a blind attack to recover CTR\_DRBG's state within three AES encryptions, without viewing output, and thus to decrypt passively collected TLS connections from the victim.

Mix-nets constructed from homomorphic cryptosystems can be generalized to process lists of ciphertexts as units and use different public keys for different parts of such lists. We present a number of blackbox constructions that enriches the set of operations provided by such mix-nets. The constructions are simple, fully practical, and eliminates the need for some specialized protocols.