International Association
for Cryptologic Research

We describe an approach to zero-sum partitions using Todo's division property at EUROCRYPT 2015. It follows the inside-out methodology, and includes MILP-assisted search for the forward and backward trails, and subspace approach to connect those two trails that is less restrictive than commonly done.

As an application we choose PHOTON, a family of sponge-like hash function proposals that was recently standardized by ISO. With respect to the security claims made by the designers, we for the first time show zero-sum partitions for almost all of those full 12-round permutation variants that use a 4-bit S-Box. As with essentially any other zero-sum property in the literature, also here the gap between a generic attack and the shortcut is small.

We present here new multivariate schemes that can be seen as HFE generalization having a property called `Two-Face'. Particularly, we present five such families of algorithms named `Dob', `Simple Pat', `General Pat', `Mac', and `Super Two-Face'. These families have connections between them, some of them are refinements or generalizations of others. Notably, some of these schemes can be used for public key encryption, and some for public key signature. We introduce also new multivariate quadratic permutations that may have interest beyond cryptography.

In this paper we improve Wu and Wang's method for finding impossible differentials of block cipher structures. This improvement is more general than Wu and Wang's method that it can find more impossible differentials with less time. We apply it on Gen-CAST256, Misty, Gen-Skipjack, Four-Cell, Gen-MARS, SMS4, MIBS, Camellia*, LBlock, E2 and SNAKE block ciphers. All impossible differentials discovered by the algorithm are the same as Wu's method. Besides, for the 8-round MIBS block cipher, we find 4 new impossible differentials, which are not listed in Wu and Wang's results. The experiment results show that the improved algorithm can not only find more impossible differentials, but also largely reduce the search time.

Cloud storage is in widespread use by individuals and enterprises but introduces a wide array of attack vectors. A basic step for users is to encrypt their data, but it is not obvious what precise security properties are required for encryption. Furthermore, cloud storage providers often use techniques such as data deduplication for improving efficiency which restricts the application of semantically-secure encryption. Generic security goals and attack models have thus far proved elusive: primitives are considered in isolation and protocols are often proved secure under ad hoc models for restricted classes of adversaries.

We provide a generic syntax for storage systems that allows us to formally model natural security notions for cloud storage and deduplication. We define security notions for confidentiality and integrity in encrypted cloud storage and determine relations between these notions. We show how to build cloud storage systems that satisfy our defined security notions using generic cryptographic components.

A new unconditionally secure multi-party quantum commitment is proposed in this paperby encoding the committed message to the phase of a quantum state. Multi-party means that there are more than one recipient in our scheme. We show that our quantum commitment scheme is unconditional hiding and binding, and hiding is perfect. Our technique is based on the interference of phase-encoded coherent states of light. Its security proof relies on the no-cloning theorem of quantum theory and the properties of quantum information.

This paper designs and analyzes a quantum algorithm to solve a system of $m$ quadratic equations in $n$ variables over a finite field ${\bf F}_q$. In the case $m=n$ and $q=2$, under standard assumptions, the algorithm takes time $2^{(t+o(1))n}$ on a mesh-connected computer of area $2^{(a+o(1))n}$, where $t\approx 0.45743$ and $a\approx 0.01467$. The area-time product has asymptotic exponent $t+a\approx 0.47210$.

For comparison, the area-time product of Grover's algorithm has asymptotic exponent $0.50000$. Parallelizing Grover's algorithm to reach asymptotic time exponent $0.45743$ requires asymptotic area exponent $0.08514$, much larger than $0.01467$.

Scalar multiplication on Legendre form elliptic curves can be speeded up in two ways. One can perform the bulk of the computation either on the associated Kummer line or on an appropriate twisted Edwards form elliptic curve. This paper provides details of moving to and from between Legendre form elliptic curves and associated Kummer line and moving to and from between Legendre form elliptic curves and related twisted Edwards form elliptic curves. Further, concrete twisted Edwards form elliptic curves are identified which correspond to known Kummer lines at the 128-bit security level which provide very fast scalar multiplication on modern architectures supporting SIMD operations.

Side-channel attacks are a threat to cryptographic algorithms running on embedded devices. Public-key cryptosystems, including elliptic curve cryptography (ECC), are particularly vulnerable because their private keys are usually long-term. Well known countermeasures like regularity, projective coordinates and scalar randomization, among others, are used to harden implementations against common side-channel attacks like DPA. Horizontal clustering attacks can theoretically overcome these countermeasures by attacking individual side-channel traces. In practice horizontal attacks have been applied to overcome protected ECC implementations on FPGAs. However, it has not been known yet whether such attacks can be applied to protected implementations working on embedded devices, especially in a non-profiled setting. In this paper we mount non-profiled horizontal clustering attacks on two protected implementations of the Montgomery Ladder on Curve25519 available in the µNaCl library targeting electromagnetic (EM) emanations. The first implementation performs the conditional swap (cswap) operation through arithmetic of field elements (cswap-arith), while the second does so by swapping the pointers (cswap-pointer). They run on a 32-bit ARM Cortex-M4F core. Our best attack has success rates of 97.64% and 99.60% for cswap-arith and cswap-pointer, respectively. This means that at most 6 and 2 bits are incorrectly recovered, and therefore, a subsequent brute-force can fix them in reasonable time. Furthermore, our horizontal clustering framework used for the aforementioned attacks can be applied against other protected implementations.

Double-authentication-preventing signatures (DAPS) are signatures designed with the aim that signing two messages with an identical first part (called address) but different second parts (called payload) allows to publicly extract the secret signing key from two such signatures. A prime application for DAPS is disincentivizing and/or penalizing the creation of two signatures on different payloads within the same address, such as penalizing double spending of transactions in Bitcoin by the loss of the double spender's money.

So far DAPS have been constructed from very specific signature schemes not used in practice and using existing techniques it has proved elusive to construct DAPS schemes from signatures widely used in practice. This, unfortunately, has prevented practical adoption of this interesting tool so far. In this paper we ask whether one can construct DAPS from signature schemes used in practice. We affirmatively answer this question by presenting novel techniques to generically construct provably secure DAPS from a large class of discrete logarithm based signatures. This class includes schemes like Schnorr, DSA, EdDSA, and, most interestingly for practical applications, the widely used ECDSA signature scheme. The resulting DAPS are highly efficient and the shortest among all existing DAPS schemes. They are nearly half of the size of the most efficient factoring based schemes (IACR PKC'17) and improve by a factor of 100 over the most efficient discrete logarithm based ones (ACM CCS'15). Although this efficiency comes at the cost of a reduced address space, i.e., size of keys linear in the number of addresses, we will show that this is not a limitation in practice. Moreover, we generalize DAPS to any N>2, which we denote as N-times-authentication-preventing signatures (NAPS). Finally, we also provide an integration of our ECDSA-based DAPS into the OpenSSL library and perform an extensive comparison with existing approaches.

We propose a variant of the CGL hash, Charles et al. 2009, that is significantly faster than the original algorithm, and prove that it is preimage and collision resistant. For $n = \log p$ where $p$ is the characteristic of the finite field, the performance ratio between CGL and the new proposal is $(2n + 104.8) / (1.8\log n + 12.6)$. Assuming the best quantum preimage attack on the hash has complexity $O(p^{\frac{1}{4}})$, we attain a concrete speed-up for a 256-bit quantum preimage security level by a factor 70.35. For a 384-bit quantum preimage security level, the speed-up is by a factor 100.36.

A cryptographic watermarking scheme embeds message into a program while preserving its functionality. Essential security of the watermarking schemes requires that no one could remove the marking message of a marked program without substantially changing its functionality. In practical applications, it is common to mark a program with multiple different messages, e.g. in the secret leaker tracing scenarios. Thus, it is usually required that the watermarking scheme should be secure against the “collusion attacks”, where the adversary can obtain multiple watermarked programs embedded with different messages for the same functionality. However, current works in this area have not formally considered this requirement.

In this paper, we formally address the problem and give new security definition for watermarking schemes that captures the collusion attacks. Then we explore the existence of watermarking schemes secure under our new security defjnition:

– On the negative side, we observe that all current watermarking schemes either do not support multi-message embedding inherently or are vulnerable to the collusion attacks.

– On the positive side, we construct watermarking scheme secure against the collusion attacks for pseudorandom function (PRF). This is achieved by introducing a new message-embedding technique in the watermarking settings and is built on a newly presented primitive, namely, private multi-programmable PRF. Based on our watermarking scheme for PRF, we also construct watermarking schemes for various other cryptographic functionalities.

In this paper, we present new key-recovery attacks on AES with a single secret S-Box. Several attacks for this model have been proposed in literature, the most recent ones at Crypto’16 and FSE’17. Both these attacks exploit a particular property of the MixColumns matrix to recover the secret-key.

In this work, we show that the same attacks work exploiting a weaker property of the MixColumns matrix. As first result, this allows to (largely) increase the number of MixColumns matrices for which it is possible to set up all these attacks. As a second result, we present new attacks on 5-round AES with a single secret S-Box that exploit the new multiple-of-n property recently proposed at Eurocrypt’17. This property is based on the fact that choosing a particular set of plaintexts, the number of pairs of ciphertexts that lie in a particular subspace is a multiple of n.

Post-quantum cryptography has attracted much attention from worldwide cryptologists. At Asiacrypt 2017, Leander and May combines Grover and Simon algorithms to quantumly break FX-based block ciphers. In this paper, we study the Feistel constructions with Grover and Simon algorithms and give some new quantum key-recovery attacks on different rounds of Feistel constructions. Our attacks requires $2^{nr/4-3n/4}$ quantum queries. When comparing with the quantum brute force search, the time complexity is reduced by a factor of $2^{0.75n}$. When comparing with the best classical attacks, the time complexity is reduced by a factor $2^{0.5n}$ without any memory cost.

A recent paper by Costello and Hisil at Asiacrypt'17 presents efficient formulas for computing isogenies with odd-degree cyclic kernels on Montgomery curves. We provide a constructive proof of a generalization of this theorem which shows the connection between the shape of the isogeny and the simple action of the point (0,0). This generalization removes the restriction of a cyclic kernel and allows for any separable isogeny whose kernel does not contain (0,0). As a particular case, we provide efficient formulas for 2-isogenies between Montgomery curves and show that these formulas can be used in isogeny-based cryptosystems without expensive square root computations and without knowledge of a special point of order 8. We also consider elliptic curves in triangular form containing an explicit point of order 3.

The Camenisch-Lysyanskaya (CL) signature is a very popular tool in cryptography, especially among privacy-preserving constructions. Indeed, the latter benefit from their numerous features such as randomizability. Following the evolution of pairing-based cryptography, with the move from symmetric pairings to asymmetric pairings, Pointcheval and Sanders (PS) proposed at CT-RSA '16 an alternative scheme which improves performances while keeping the same properties. Unfortunately, CL and PS signatures raise concerns in the cryptographic community because they both rely on interactive assumptions that essentially state their EUF-CMA security. This lack of precise security assessment is obviously a barrier to a widespread use of these signatures and a reason for preferring other constructions, such as the ones relying on $q$-type assumptions.

In this paper, we study more thoroughly the security of these signatures and prove that it actually relies, for both constructions, on simple variants of the $\textsf{SDH}$ assumption, assuming a slight modification of the original constructions. Our work thus shows that the CL and PS signature schemes offer similar security guarantees as those provided by several other constructions using bilinear groups, and so that one can benefit from their interesting features without jeopardizing security.

Secure Remote Password (SRP) protocol is an augmented Password-based Authenticated Key Exchange (PAKE) protocol based on discrete logarithm problem (DLP) with various attractive security features. Compared with basic PAKE protocols, SRP does not require server to store user's password and user does not send password to server to authenticate. These features are desirable for secure client-server applications. SRP has gained extensive real-world deployment, including Apple iCloud, 1Password etc. However, with the advent of quantum computer and Shor's algorithm, classic DLP-based public key cryptography algorithms are no longer secure, including SRP. Motivated by importance of SRP and threat from quantum attacks, we propose a RLWE-based SRP protocol (RLWE-SRP) which inherit advantages from SRP and elegant design from RLWE key exchange. We also present parameter choice and efficient portable C++ implementation of RLWE-SRP. Implementation of our 209-bit secure RLWE-SRP is more than 3x faster than 112-bit secure original SRP protocol, 5.5x faster than 80-bit secure J-PAKE and 14x faster than two 184-bit secure RLWE-based PAKE protocols with more desired properties.

In this paper, we introduce CAPA: a combined countermeasure against physical attacks. Our countermeasure provides security against higher-order SCA, multiple-shot DFA and combined attacks, scales to arbitrary protection order and is suitable for implementation in embedded hardware and software. The methodology is based on an attack model which we call tile-probe-and-fault, which is an extension (in both attack surface and capabilities) of prior work such as the wire-probe model. The tile-probe-and-fault leads one to naturally look (by analogy) at actively secure multi-party computation protocols such as SPDZ. We detail several proof-of-concept designs using the CAPA methodology: a hardware implementation of the KATAN and AES block ciphers, as well as a software bitsliced AES S-box implementation. We program a second-order secure version of the KATAN design into a Spartan-6 FPGA and perform a side-channel evaluation. No leakage is detected with up to 18 million traces. We also deploy a second-order secure software AES S-box implementation into an ARM Cortex-M4. Neither first- nor second-order leakage is detected with up to 200 000 traces. Both our implementations can detect faults within a strong adversarial model with arbitrarily high probability.

Nachef et al used differential cryptanalysis to study four types of Generalized Feistel Scheme (GFS). They gave the lower bound of maximum number of rounds that is indistinguishable from a random permutation. In this paper, we study the security of several types of GFS by exploiting the asymmetric property. We show that better lower bounds can be achieved for the Type-1 GFS, Type-3 GFS and Alternating Feistel Scheme. Furthermore, we give the first general results regarding to the lower bound of the Unbalanced Feistel Scheme.

Glitches entail a great issue when securing a cryptographic implementation in hardware. Several masking schemes have been proposed in the literature that provide security even in the presence of glitches. The key property that allows this protection was introduced in threshold implementations as non-completeness. We address crucial points to ensure the right compliance of this property especially for low-latency implementations. Specifically, we first discuss the existence of a flaw in DSD 2017 implementation of Keccak by Gross et al. in violation of the non-completeness property and propose a solution. We perform a side-channel evaluation on the first-order and second-order implementations of the proposed design where no leakage is detected with up to 55 million traces. Then, we present a method to ensure a non-complete scheme of an unrolled implementation applicable to any order of security or algebraic degree of the shared function. By using this method we design a two-rounds unrolled first-order Keccak-f [200] implementation that completes an encryption in 20.61ns, the fastest implementation in the literature to this date.

Two post-quantum password-based authenticated key exchange (PAKE) protocols were proposed at CT-RSA 2017. Following this work, we give much more efficient and portable C++ implementation of these two protocols. We also choose more compact parameters providing 200-bit security. Compared with original implementation, we achieve 21.5x and 18.5x speedup for RLWE-PAK and RLWE-PPK respectively. Compare with quantum-vulnerable J-PAKE protocol, we achieve nearly 8x speedup. We also integrate RLWE-PPK into TLS to construct a post-quantum TLS ciphersuite. This allows simpler key management, mutual authentication and resistant to phishing attack. Benchmark shows that our ciphersuite is indeed practical.