IACR News
Updates on the COVID-19 situation are on the
Announcement channel.
Here you can see all recent updates to the IACR webpage. These updates are also available:
23 May 2022
Andriyan Bilyk, Javad Doliskani, Zhiyong Gong
We investigate the security assumptions behind three public-key quantum money schemes. Aaronson and Christiano proposed a scheme based on hidden subspaces of the vector space $\mathbb{F}_2^n$ in 2012. It was conjectured by Pena et al in 2015 that the hard problem underlying the scheme can be solved in quasi-polynomial time. We confirm this conjecture by giving a polynomial time quantum algorithm for the underlying problem. Our algorithm is based on computing the Zariski tangent space of a random point in the hidden subspace.
Zhandry proposed a scheme based on multivariate hash functions in 2017. We give a polynomial time quantum algorithm for cloning a money state with high probability. Our algorithm uses the verification circuit of the scheme to produce a banknote from a given serial number.
Kane proposed a scheme based on modular forms in 2018. The underlying hard problem in Kane's scheme is cloning a quantum state that represents an eigenvector of a set of Hecke operators. We give a polynomial time quantum reduction from this hard problem to a linear algebra problem. The latter problem is much easier to understand, and we hope that our reduction opens new avenues to future cryptanalyses of this scheme.
Anders Dalskov, Daniel Escudero, Ariel Nof
We introduce a new MPC protocol to securely compute any functionality over an arbitrary black-box finite ring (which may not be commutative), tolerating $t
Furthermore, the amortized communication complexity of our protocol is not only constant, but very small: only $1 + \frac{t-1}{n}<1\frac{1}{3}$ ring elements per party, per multiplication gate over two rounds of interaction. This improves over the state-of-the art protocol in the same setting by Furukawa and Lindell (CCS 2019), which has a communication complexity of $2\frac{2}{3}$ \emph{field} elements per party, per multiplication gate and while achieving fairness only. As an alternative, we also describe a variant of our protocol which has only one round of interaction per multiplication gate on average, and amortized communication cost of $\le 1\frac{1}{2}$ ring elements per party on average for any natural circuit.
Motivated by the fact that efficiency of distributed protocols are much more penalized by high communication complexity than local computation/storage, we perform a detailed analysis together with experiments in order to explore how large the number of parties can be, before the storage and computation overhead becomes prohibitive. Our results show that our techniques are viable even for a moderate number of parties (e.g., $n>10$).
Furthermore, the amortized communication complexity of our protocol is not only constant, but very small: only $1 + \frac{t-1}{n}<1\frac{1}{3}$ ring elements per party, per multiplication gate over two rounds of interaction. This improves over the state-of-the art protocol in the same setting by Furukawa and Lindell (CCS 2019), which has a communication complexity of $2\frac{2}{3}$ \emph{field} elements per party, per multiplication gate and while achieving fairness only. As an alternative, we also describe a variant of our protocol which has only one round of interaction per multiplication gate on average, and amortized communication cost of $\le 1\frac{1}{2}$ ring elements per party on average for any natural circuit.
Motivated by the fact that efficiency of distributed protocols are much more penalized by high communication complexity than local computation/storage, we perform a detailed analysis together with experiments in order to explore how large the number of parties can be, before the storage and computation overhead becomes prohibitive. Our results show that our techniques are viable even for a moderate number of parties (e.g., $n>10$).
Olive Chakraborty, Martin Zuber
We design and implement a new efficient and accurate Fully homomorphic argmin/min or argmax/max comparison operator, which finds its application in numerous real-world use cases as a classifier. In particular we propose two versions of our algorithms using different tools from TFHE's functional bootstrapping toolkit. Our algorithm scales to any number of input data points with linear time complexity and logarithmic noise-propagation. Our algorithm is the fastest on the market for non-parallel comparisons with a high degree of accuracy and precision. For MNIST and SVHN datasets, which work under the PATE framework, using our algorithm, we achieve an accuracy of around 99.95 % for both.
Arantxa Zapico, Vitalik Buterin, Dmitry Khovratovich, Mary Maller, Anca Nitulescu, Mark Simkin
We present position-hiding linkability for vector commitment schemes: one can prove in zero knowledge that one or $m$ values that comprise commitment cm all belong to the vector of size $N$ committed to in C. Our construction Caulk can be used for membership proofs and lookup arguments and outperforms all existing alternatives in prover time by orders of magnitude.
For both single- and multi-membership proofs Caulk beats SNARKed Merkle proofs by the factor of 100 even if the latter instantiated with Poseidon hash. Asymptotically our prover needs $O(m^2 + m\log N)$ time to prove a batch of $m$ openings, whereas proof size is $O(1)$ and verifier time is $O(\log(\log N))$.
As a lookup argument, Caulk is the first scheme with prover time sublinear in the table size, assuming $O(N\log N)$ preprocessing time and $O(N)$ storage. It can be used as a subprimitive in verifiable computation schemes in order to drastically decrease the lookup overhead.
Our scheme comes with a reference implementation and benchmarks.
Zhenyu Huang, Siwei Sun
The significant progress in the development of quantum computers has made the study of cryptanalysis based on quantum computing an active topic. To accurately estimate the resources required to carry out quantum attacks, the involved quantum algorithms have to be synthesized into quantum circuits with basic quantum gates. In this work, we present several generic synthesis and optimization techniques for circuits implementing the quantum oracles of iterative symmetric-key ciphers that are commonly employed in quantum attacks based on Grover and Simon’s algorithms. Firstly, a general structure for implementing the round functions of block ciphers in-place is proposed. Then, we present some novel techniques for synthesizing efficient quantum circuits of linear and non-linear cryptographic building blocks. We apply these techniques to AES and systematically investigate the strategies for depth-width trade-offs. Along the way, we derive a quantum circuit
for the AES S-box with provably minimal T-depth based on some new observations on its classical circuit. As a result, the T-depth and width (number of qubits) required for implementing the quantum circuits of AES are significantly reduced. Compared with the circuit proposed in EUROCRYPT 2020, the T-depth is reduced from 60 to 40 without increasing the width or 30 with a slight increase in width. These circuits are fully implemented in Microsoft Q# and the source code is publicly
available. Compared with the circuit proposed in ASIACRYPT 2020, the width of one of our circuits is reduced from 512 to 371, and the Toffoli-depth is reduced from 2016 to 1558 at the same time. Actually, we can reduce the width to 270 at the cost of increased depth. Moreover, a full spectrum of depth-width trade-offs is provided, setting new records for the synthesis and optimization of quantum circuits of AES.
Matthieu Rambaud, Antoine Urban
We present the first proactive secret sharing under honest majority which is purely over pairwise asynchronous channels.
Moreover:
- it has robust reconstruction. In addition, provided a single broadcast in the sharing phase, then it commits the dealer to a value;
- it operates under a bare PKI and enables dynamic membership;
- the standard version carries over the model known as \cst{receiver-anonymous (Yoso)} in which participants speak only once then erase their memories.
Each refresh of a secret takes $2$ actual message delays and a total of $O(n^4)$ bits sent by the honest \players;
- it allows an optimization in $O(n^3)$ bits sent and latency of $5$ messages.
Adriano Koleci
In the recent years there has been a growing interest in ARX ciphers thanks to their performance in low cost architectures. This work is a short and simple proof that Add, Rotate and Exclusive-OR (ARX) operations generate the permutation group S_{2^n} and it is made up by elementary arguments with minimal use of group theory.
Shingo Sato, Junji Shikata
Selective opening (SO) security is one of the most important security notions of public key encryption (PKE) in a multi-user setting. Even though messages and random coins used in some ciphertexts are leaked, SO security guarantees the confidentiality of the other ciphertexts. Actually, it is shown that there exist PKE schemes which meet the standard security such as indistinguishability against chosen ciphertext attacks (IND-CCA security) but do not meet SO security against chosen ciphertext attacks. Hence, it is important to consider SO security in the multi-user setting. On the other hand, many researchers have studied cryptosystems in the security model where adversaries can submit quantum superposition queries (i.e., quantum queries) to oracles. In particular, IND-CCA secure PKE and KEM schemes in the quantum random oracle model have been intensively studied so far. In this paper, we show that two kinds of constructions of hybrid encryption schemes meet simulation-based SO security against chosen ciphertext attacks (SIM-SO-CCA security) in the quantum random oracle model or the quantum ideal cipher model. The first scheme is constructed from any IND-CCA secure KEM and any simulatable data encapsulation mechanism (DEM). The second one is constructed from any IND-CCA secure KEM based on Fujisaki-Okamoto transformation and any strongly unforgeable message authentication code (MAC). We can apply any IND-CCA secure KEM scheme to the first one if the underlying DEM scheme meets simulatability, whereas we can apply strongly unforgeable MAC to the second one if the underlying KEM is based on Fujisaki-Okamoto transformation.
Ren Ishibashi, Kazuki Yoneyama
Authenticated Key Exchange (AKE) is a cryptographic protocol to share a common session key among multiple parties. Usually, PKI-based AKE schemes are designed to guarantee secrecy of the session key and mutual authentication. However, in practice, there are many cases where mutual authentication is undesirable such as in anonymous networks like Tor and Riffle, or difficult to achieve due to the certificate management at the user level such as the Internet. Goldberg et al. formulated a model of anonymous one-sided AKE which guarantees the anonymity of the client by allowing only the client to authenticate the server, and proposed a concrete scheme. However, existing anonymous one-sided AKE schemes are only known to be secure in the random oracle model. In this paper, we propose generic constructions of anonymous one-sided AKE in the random oracle model and in the standard model, respectively. Our constructions allow us to construct the first post-quantum anonymous one-sided AKE scheme from isogenies in the standard model.
Thomas Debris-Alazard, Léo Ducas, Nicolas Resch, Jean-Pierre Tillich
In this article we revisit smoothing bounds in parallel between lattices \emph{and} codes. Initially introduced by Micciancio and Regev, these bounds were instantiated with Gaussian distributions and were crucial for arguing the security of many lattice-based cryptosystems. Unencumbered by direct application concerns, we provide a systematic study of how these bounds are obtained for both lattices \emph{and} codes, transferring techniques between both areas. We also consider various spherically symmetric noise distributions.
We found that the best strategy for a worst-case bound combines Parseval's Identity, the Cauchy-Schwarz inequality, and the second linear programming bound, and this for both codes and lattices, and for all noise distributions at hand. For an average-case analysis, the linear programming bound can be replaced by a tight average count.
This alone gives optimal results for spherically uniform noise over random codes and random lattices. This also improves previous Gaussian smoothing bound for worst-case lattices, but surprisingly this provides even better results for uniform noise than for Gaussian (or Bernouilli noise for codes).
This counter-intuitive situation can be resolved by adequate decomposition and truncation of Gaussian and Bernouilli distribution into a superposition of uniform noise, giving further improvement for those cases, and putting them on par with the uniform cases.
Yu Zhang, Zongbin Wang, Tihong Qin
Privacy preserving keyword search (PPKS) is investigated in this paper, which aims to ensure the privacy of clients and servers when a database is accessed. Range query has been recognized as a common operation in databases. In this paper, a formal definition of PPKS supporting range query is given, a scheme (PPRKS) is presented in accordance with Paillier’s cryptosystem. To the best of our knowledge, PPRKS has been the only existing scheme that can effectively preserve the privacy of range keyword search. Moreover, it is demonstrated that the security of PPRKS is dependent on the semantic security of Paillier’s cryptosystem. A detailed performance analysis and a simulation are conducted to verify the practicality of PPRKS. As revealed by the theoretical analysis and the experimental results, the proposed scheme is practical.
Marloes Venema, Greg Alpár
Ciphertext-policy attribute-based encryption is a versatile primitive that has been considered extensively to securely manage data in practice. Especially completely unbounded schemes are attractive, because they do not restrict the sets of attributes and policies. So far, any such schemes that support negations in the access policy or that have online/offline extensions have an inefficient decryption algorithm.
In this work, we propose GLUE (Generalized, Large-universe, Unbounded and Expressive), which is a novel scheme that allows for the efficient implementation of the decryption while allowing the support of both negations and online/offline extensions. We achieve these properties simultaneously by uncovering an underlying dependency between encryption and decryption, which allows for a flexible trade-off in their efficiency. For the security proof, we devise a new technique that enables us to generalize multiple existing schemes. As a result, we obtain a completely unbounded scheme supporting negations that, to the best of our knowledge, outperforms all existing schemes in the decryption algorithm.
In this work, we propose GLUE (Generalized, Large-universe, Unbounded and Expressive), which is a novel scheme that allows for the efficient implementation of the decryption while allowing the support of both negations and online/offline extensions. We achieve these properties simultaneously by uncovering an underlying dependency between encryption and decryption, which allows for a flexible trade-off in their efficiency. For the security proof, we devise a new technique that enables us to generalize multiple existing schemes. As a result, we obtain a completely unbounded scheme supporting negations that, to the best of our knowledge, outperforms all existing schemes in the decryption algorithm.
Raghvendra Rohit, Santanu Sarkar
SPEEDY is a family of ultra low latency block ciphers proposed by Leander, Moos, Moradi and Rasoolzadeh at TCHES 2021. Although the designers gave some differential/linear distinguishers for reduced rounds, a concrete cryptanalysis considering key recovery attacks on SPEEDY was completely missing. The latter is crucial to understand the security margin of designs like SPEEDY which typically use low number of rounds to have low latency. In this work, we present the first third-party cryptanalysis of SPEEDY-$r$-192, where $r \in \{5, 6, 7\}$ is the number of rounds and 192 is block and key size in bits. We identify cube distinguishers for 2 rounds with data complexities $2^{14}$ and $2^{13}$, while the differential/linear distinguishers provided by designers has a complexity of $2^{39}$. Notably, we show that there are several such cube distinguishers, and thus, we then provide a generic description of them. We also investigate the structural properties of 13-dimensional cubes and give experimental evidence that the partial algebraic normal form of certain state bits after two rounds is always the same. Next, we utilize the 2 rounds distinguishers to mount a key recovery attack on 3 rounds SPEEDY.
Our attack require $2^{17.6}$ data, $2^{25.5}$ bits of memory and $2^{52.5}$ time. Our results show that the practical variant of SPEEDY, i.e., SPEEDY-5-192 has a security margin of only 2 rounds. We believe our work will bring new insights in understanding the security of SPEEDY.
Gongyu Shi, Geng Wang, Dawu Gu
To enhance the security or the efficiency of the standard RSA cryptosystem, some variants have been proposed based on elliptic curves, Gaussian integers or Lucas sequences. A typical type of these variants which we called Type-A variants have the specified modified Euler's totient function $\psi(N)=(p^2-1)(q^2-1)$. But in 2018, based on cubic Pell equation, Murru and Saettone presented a new RSA-like cryptosystem, and it is another type of RSA variants which we called Type-B variants, since their scheme has $\psi(N)=(p^2+p+1)(q^2+q+1)$. For RSA-like cryptosystems, four key-related attacks have been widely analyzed, e.g., the small private key attack, the multiple private keys attack, the partial key exposure attack and the small prime difference attack. These attacks are well-studied on both standard RSA and Type-A variants. Recently, the small private key attack on Type-B variants has also been analyzed. In this paper, we make further cryptanalysis of Type-B variants, that is, we propose the first theoretical results of multiple private keys attack, partial key exposure attack as well as small prime difference attack on Type-B variants, and the validity of our attacks are verified by experiments. Our results show that for all three attacks, Type-B variants are less secure than standard RSA.
Tingting Pang, Nian Li, Xiangyong Zeng
In this paper, we investigate the cardinality, denoted by $(j_1,j_2,j_3,j_4)_2$, of the intersection of $(\mathcal{C}^{(2)}_{j_1}-1)\cap(\mathcal{C}^{(2)}_{j_2}-2)\cap(\mathcal{C}^{(2)}_{j_3}-3)
\cap(\mathcal{C}^{(2)}_{j_4}-4)$ for $j_1,j_2,j_3,j_4\in\{0,1\}$, where $\mathcal{C}^{(2)}_0, \mathcal{C}^{(2)}_1$ are the cyclotomic classes of order two over the finite field $\mathbb{F}_{p^n}$, $p$ is an odd prime and $n$ is a positive integer. By making most use of the results on cyclotomic classes of orders two and four as well as the cardinality of the intersection
$(\mathcal{C}^{(2)}_{i_1}-1)\cap(\mathcal{C}^{(2)}_{i_2}-2)\cap(\mathcal{C}^{(2)}_{i_3}-3)$, we compute the values of $(j_1,j_2,j_3,j_4)_2$ in the case of $p=5$, where $i_1,i_2,i_3\in\{0,1\}$. As a consequence, the power function $x^{\frac{5^n-1}{2}+2}$ over $\mathbb{F}_{5^n}$ is shown to be differentially $3$-uniform and its differential spectrum is also completely determined.
Mingxun Zhou, Wei-Kai Lin, Yiannis Tselekounis, Elaine Shi (random author ordering)
We construct a single-server
pre-processing Private Information Retrieval
(PIR) scheme
with optimal bandwidth
and server computation (up to poly-logarithmic factors), assuming
hardness of the Learning With Errors (LWE) problem.
Our scheme achieves
amortized
$\widetilde{O}_{\lambda}(\sqrt{n})$
server and client computation and $\widetilde{O}_\lambda(1)$
bandwidth per query, completes in a single roundtrip, and requires
$\widetilde{O}_\lambda(\sqrt{n})$
client storage.
In particular, we achieve a significant
reduction in bandwidth over the
state-of-the-art scheme by Corrigan-Gibbs,
Henzinger, and Kogan (Eurocrypt'22):
their scheme requires as much as
$\widetilde{O}_{\lambda}(\sqrt{n})$
bandwidth per query, with comparable
computational and storage overhead as ours.
Chen-Da Liu-Zhang, Christian Matt, Ueli Maurer, Guilherme Rito, Søren Eller Thomsen
In recent years, permisionless blockchains have received a lot of attention both from industry and academia, where substantial effort has been spent to develop consensus protocols that are secure under the assumption that less than half (or a third) of a given resource (e.g., stake or computing power) is controlled by corrupted parties. The security proofs of these consensus protocols usually assume the availability of a network functionality guaranteeing that a block sent by an honest party is received by all honest parties within some bounded time. To obtain an overall protocol that is secure under the same corruption assumption, it is therefore necessary to combine the consensus protocol with a network protocol that achieves this property under that assumption. In practice, however, the underlying network is typically implemented by flooding protocols that are not proven to be secure in the setting where a fraction of the considered total weight can be corrupted. This has led to many so-called eclipse attacks on existing protocols and tailor-made fixes against specific attacks.
To close this apparent gap, we propose a flooding protocol that provably delivers sent messages to all honest parties after a logarithmic number of steps. We prove security in the setting where all parties are publicly assigned a positive weight and the adversary can corrupt parties accumulating up to a constant fraction of the total weight. This can directly be used in the proof-of-stake setting, but is not limited to it. To prove the security of our protocol, we combine known results about the diameter of Erdős–Rényi graphs with reductions between different types of random graphs. We further show that the efficiency of our protocol is asymptotically optimal.
The practicality of our protocol is supported by extensive simulations for different numbers of parties, weight distributions, and corruption strategies. The simulations confirm our theoretical results and show that messages are delivered quickly regardless of the weight distribution, whereas protocols that are oblivious of the parties' weights completely fail if the weights are unevenly distributed. Furthermore, the average message complexity per party of our protocol is within a small constant factor of such a protocol. Hence, security in a weighted setting essentially comes for free with our techniques.
To close this apparent gap, we propose a flooding protocol that provably delivers sent messages to all honest parties after a logarithmic number of steps. We prove security in the setting where all parties are publicly assigned a positive weight and the adversary can corrupt parties accumulating up to a constant fraction of the total weight. This can directly be used in the proof-of-stake setting, but is not limited to it. To prove the security of our protocol, we combine known results about the diameter of Erdős–Rényi graphs with reductions between different types of random graphs. We further show that the efficiency of our protocol is asymptotically optimal.
The practicality of our protocol is supported by extensive simulations for different numbers of parties, weight distributions, and corruption strategies. The simulations confirm our theoretical results and show that messages are delivered quickly regardless of the weight distribution, whereas protocols that are oblivious of the parties' weights completely fail if the weights are unevenly distributed. Furthermore, the average message complexity per party of our protocol is within a small constant factor of such a protocol. Hence, security in a weighted setting essentially comes for free with our techniques.
Son Ho, Jonathan Protzenko, Abhishek Bichhawat, Karthikeyan Bhargavan
The Noise protocol framework defines a succinct notation and execution
framework for a large class of 59+ secure channel protocols, some of
which are used in popular applications such as WhatsApp and WireGuard.
We present a verified implementation of a Noise
protocol compiler that takes any Noise protocol, and produces
an optimized C implementation with extensive correctness and security
guarantees. To this end, we formalize the complete Noise stack in
F*, from the low-level cryptographic library to a high-level API.
We write our compiler also in F*, prove that it meets our formal
specification once and for all, and then specialize it on-demand for
any given Noise protocol, relying on a novel technique called
hybrid embedding. We thusa establish functional correctness,
memory safety and a form of side-channel resistance for the generated
C code for each Noise protocol. We propagate these guarantees to the
high-level API, using defensive dynamic checks to prevent incorrect
uses of the protocol. Finally, we formally state and prove the
security of our Noise code, by building on a symbolic model of
cryptography in F*, and formally link high-level API
security goals stated in terms of security levels to
low-level cryptographic guarantees.
Ours are the first comprehensive verification results for a
protocol compiler that targets C code and the first verified
implementations of any Noise protocol. We evaluate our framework by
generating implementations for all 59 Noise protocols and by comparing
the size, performance, and security of our verified code against other
(unverified) implementations and prior security analyses of Noise.
Li Duan, Yufan Jiang, Yong Li, Jörn Müller-Quade, Andy Rupp
Secure multiparty computation (MPC) allows distrustful parties to jointly compute some functions while keeping their private secrets unrevealed. MPC adversaries are often categorized as semi-honest and malicious, depending on whether they follow the protocol specifications or not. Covert security was first introduced by Aumann and Lindell in 2007, which models a third type of active adversaries who cheat but can be caught with a probability. However, this probability is predefined externally, and the misbehavior detection must be made by other honest participants with cut-and-choose in current constructions. In this paper, we propose a new security notion called security against honorific adversaries, who may cheat during the protocol execution but are extremely unwilling to be punished. Intuitively, honorific adversaries can cheat successfully, but decisive evidence of misbehavior will be left to honest parties with a probability close to one. By introducing an independent but not trusted auditor to the MPC ideal functionality in the universal composability framework (UC), we avoid heavy cryptographic machinery in detection and complicated discussion about the probability of being caught. With this new notion, we construct new provably secure protocols without cut-and-choose for garbled circuits that are much more efficient than those in the covert and malicious model, with slightly more overhead than passively secure protocols.
Alexandru Ionita
Unlike conventional ABE systems, which support Boolean attributes (with only 2 states: "1" and "0", or "Present" and "Absent"), weighted Attribute-based encryption schemes also support numerical values attached to attributes, and each terminal node of the access structure contains a threshold for a minimum weight. We propose a weighted ABE system, with access policy of logarithmic expansion, by dividing each weighted attribute in sub-attributes. On top of that, we show that the decryption can be parallelized, leading to a notable improvement in running time, compared to the serial version.