International Association
for Cryptologic Research

Secure multi-party computation (MPC) allows multiple parties to jointly compute the output of a function while preserving the privacy of any individual party's inputs to that function. As MPC protocols transition from research prototypes to real-world applications, the usability of MPC-enabled applications is increasingly critical to their successful deployment and wide adoption.

Our Web-MPC platform, designed with a focus on usability, has been deployed for privacy-preserving data aggregation initiatives with the City of Boston and the Greater Boston Chamber of Commerce. After building and deploying an initial version of this platform, we conducted a heuristic evaluation to identify additional usability improvements and implemented corresponding application enhancements. However, it is difficult to gauge the effectiveness of these changes within the context of real-world deployments using traditional web analytics tools without compromising the security guarantees of the platform. This work consists of two contributions that address this challenge: (1) the Web-MPC platform has been extended with the capability to collect web analytics using existing MPC protocols, and (2) this capability has been leveraged to conduct a usability study comparing the two version of Web-MPC (before and after the heuristic evaluation and associated improvements).

While many efforts have focused on ways to enhance the usability of privacy-preserving technologies, this study can serve as a model for using a privacy-preserving data-driven approach in evaluating or enhancing the usability of privacy-preserving websites and applications deployed in real-world scenarios. The data collected in this study yields insights about the interplay between usability and security that can help inform future implementations of applications that employ MPC.

Event date: 6 December to 8 December 2019
Submission deadline: 1 August 2019
Notification: 1 October 2019

Homomorphic encryption (HE) is often viewed as impractical, both in communication and computation. Here we provide an additively homomorphic encryption scheme based on (ring) LWE with nearly optimal rate ($1-\epsilon$ for any $\epsilon>0$). Moreover, we describe how to compress many Gentry-Sahai-Waters (GSW) ciphertexts (e.g., ciphertexts that may have come from a homomorphic evaluation) into (fewer) high-rate ciphertexts.

Using our high-rate HE scheme, we are able for the first time to describe a single-server private information retrieval (PIR) scheme with sufficiently low computational overhead so as to be practical for large databases. Single-server PIR inherently requires the server to perform at least one bit operation per database bit, and we describe a rate-(4/9) scheme with computation which is not so much worse than this inherent lower bound. In fact it is probably less than whole-database AES encryption -- specifically about 1.5 mod-$q$ multiplication per database byte, where $q$ is about 50 to 60 bits. Asymptotically, the computational overhead of our PIR scheme is $\tilde{O}(\log \log \lambda + \log \log \log N)$, where $\lambda$ is the security parameter and $N$ is the number of database files, which are assumed to be sufficiently large.

In this work, we define and construct fully homomorphic non-interactive zero knowledge (FH-NIZK) and non-interactive witness-indistinguishable (FH-NIWI) proof systems.

We focus on the NP complete language $L$, where, for a boolean circuit $C$ and a bit $b$, the pair $(C,b) \in L$ if there exists an input $w$ such that $C(w)=b$. For this language, we call a non-interactive proof system 'fully homomorphic' if, given instances $(C_i,b_i) \in L$ along with their proofs $\Pi_i$, for $i \in \{1,\ldots,k\}$, and given any circuit $D:\{0,1\}^k \rightarrow \{0,1\}$, one can efficiently compute a proof $\Pi$ for $(C^*,b) \in L$, where $C^*(w^{(1)}, \ldots,w^{(k)})=D(C_1(w^{(1)}),\ldots,C_k(w^{(k)}))$ and $D(b_1,\ldots,b_k)=b$. The key security property is 'unlinkability': the resulting proof $\Pi$ is indistinguishable from a fresh proof of the same statement.

Our first result, under the Decision Linear Assumption (DLIN), is an FH-NIZK proof system for L in the common random string model. Our more surprising second result (under a new decisional assumption on groups with bilinear maps) is an FH-NIWI proof system that requires no setup.

The Minrank (MR) problem is a computational problem closely related to attacks on code- and multivariate-based schemes. In this paper we revisit the so-called Kipnis-Shamir (KS) approach to this problem. We extend previous complexity analysis by exposing non-trivial syzygies through the analysis of the Jacobian of the resulting system, with respect to a group of variables. We focus on a particular set of instances that yield a very overdetermined system which we refer to as ``superdetermined''. We provide a tighter complexity estimate for such instances and discuss its implications for the key recovery attack on some multivariate schemes. For example, in HFE the speedup is roughly a square root.

We present a post-quantum key agreement scheme that does not require distinguishing between the initiator and the responder. This scheme is based on elliptic curve isogenies and can be viewed as a variant of the well-know SIDH protocol. Then, we provide an isogeny-based password-authenticated key exchange protocol based on our scheme. A summary of security and computational complexities are also presented. Finally, we present an efficient countermeasure against a side-channel attack that applies to both static and ephemeral versions of SIDH and our scheme.

We first introduce a family of binary $pq^2$ -periodic sequences based on the Euler quotients modulo $pq$, where $p$ and $q$ are two distinct odd primes and $p$ divides $q - 1$. The minimal polynomials and linear complexities are determined for the proposed sequences provided that $2^{q-1} \not \equiv 1 \pmod q^2$ . The results show that the proposed sequences have high linear complexities.

A key step in Regev's (2009) reduction of the Discrete Gaussian Sampling (DGS) problem to that of solving the Learning With Errors (LWE) problem is a statistical test required for verifying possible solutions to the LWE problem. Regev showed that asymptotically, the success probability of this test is exponentially close to 1. In this work, we work out the concrete details of the success probability and its effect on the tightness gap of the reduction. For lattice dimensions from 2 to 187149, the tightness gap is determined entirely by the inverse of the success probability. The actual value of the tightness gap for such lattice dimensions is huge showing that the reduction cannot be used for choosing parameters for practical cryptosystems.

TRIFLE is a Round 1 candidate of the NIST Lightweight Cryptography Standardization process. In this paper, we present an interesting 1-round iterative differential characteristic of the underlying block cipher TRIFLE-BC used in TRIFLE, which holds with probability of $2^{-3}$. Consequently, it allows to mount distinguishing attack on TRIFLE-BC for up to 43 (out of 50) rounds with data complexity $2^{124}$ and time complexity $2^{124}$. Most importantly, with such an iterative differential characteristic, the forgery attack on TRIFLE can reach up to 21 (out of 50) rounds with data complexity $2^{63}$ and time complexity $2^{63}$. Finally, to achieve key recovery attack on reduced TRIFLE, we construct a differential characteristic covering three blocks by carefully choosing the positions of the iterative differential characteristic. As a result, we can mount key-recovery attack on TRIFLE for up to 11 rounds with data complexity $2^{63}$ and time complexity $2^{104}$. Although the result in this paper cannot threaten the security margin of TRIFLE, we hope it can help further understand the security of TRIFLE.

Oblivious transfer is one of the main pillars of modern cryptography and plays a major role as a building block for other more complex cryptographic primitives. In this work, we present an efficient and versatile framework for oblivious transfer (OT) using one-round key-exchange (ORKE), a special class of key exchange (KE) where only one message is sent from each party to the other. Our contributions can be summarized as follows:

i) We carefully analyze ORKE schemes and introduce new security definitions. Namely, we introduce a new class of ORKE schemes, called Alice-Bob one-round key-exchange (A-B ORKE), and the definitions of message and key indistinguishability.

ii) We show that OT can be obtained from A-B ORKE schemes fulfilling message and key indistinguishability. We accomplish this by designing a new efficient, versatile and universally composable framework for OT in the Random Oracle Model (ROM). The efficiency of the framework presented depends almost exclusively on the efficiency of the A-B ORKE scheme used since all other operations are linear in the security parameter. Universally composable OT schemes in the ROM based on new hardness assumptions can be obtained from instantiating our framework.

Examples are presented using the classical Diffie-Hellman KE, RLWE-based KE and Supersingular Isogeny Diffie-Hellman KE.

Recently, Castryck, Lange, Martindale, Panny, and Renes proposed CSIDH (pronounced "sea-side") as a candidate post-quantum "commutative group action." It has attracted much attention and interest, in part because it enables noninteractive Diffie-Hellman-like key exchange with quite small communication. Subsequently, CSIDH has also been used as a foundation for digital signatures.

In 2003-04, Kuperberg and then Regev gave asymptotically subexponential quantum algorithms for "hidden shift" problems, which can be used to recover the CSIDH secret key from a public key. In 2013, Kuperberg gave a follow-up quantum algorithm called the collimation sieve ("c-sieve" for short), which improves the prior ones, in particular by using exponentially less quantum memory and offering more parameter tradeoffs. While recent works have analyzed the concrete cost of the original algorithms (and variants) against CSIDH, there seems not to have been any consideration of the c-sieve.

This work fills that gap. Specifically, we generalize Kuperberg's collimation sieve to work for arbitrary finite cyclic groups, provide some practical efficiency improvements, give a classical (i.e., non-quantum) simulator, run experiments for a wide range of parameters up to and including the actual CSIDH-512 group order, and concretely quantify the complexity of the c-sieve against CSIDH.

Our main conclusion is that the proposed CSIDH-512 parameters provide relatively little quantum security beyond what is given by the cost of quantumly evaluating the CSIDH group action itself (on a uniform superposition). The cost of key recovery is, for example, only about $2^{16}$ quantum evaluations using $2^{40}$ bits of quantumly accessible classical memory (plus insignificant other resources); moreover, these quantities can be traded off against each other. (This improves upon a recent estimate of $2^{32.5}$ evaluations and $2^{31}$ qubits of quantum memory, for a variant of Kuperberg's original sieve.) Therefore, under the plausible assumption that quantum evaluation does not cost very much more than indicated by a recent "best case" analysis, CSIDH-512 does not achieve the claimed 64 bits of quantum security, and it falls well short of the claimed NIST security level 1 when accounting for the MAXDEPTH restriction.

The threat of the possible advent of quantum computers has motivated the cryptographic community to search for quantum safe solutions. There have been some works in past few years showing the vulnerability of symmetric key crypto-systems in the quantum setting. Among these the works by Kuwakado et al. and Kaplan et al. use the quantum period finding procedure called Simon’s algorithm to attack several symmetric crypto-systems. In this work, we use Simon’s algorithm to break six tweakable enciphering schemes (TESs) in the quantum setting. These are CMC, EME, XCB, TET, AEZ and FAST. All of them have usual proofs of security in the classical sense. A version of EME and a version of XCB are IEEE standardised TESs.

In this work, we describe how to deploy a cryptographic secure computation protocol for routine use in industry. Based on our experience, we identify major preliminaries and enabling factors which we found to be critical to the successful deployment of such technology as a practical, and uniquely positioned method for solving the task at hand.

The specific technical problem that we tackled is that of computing Private Intersection-Sum. In this setting two parties hold datasets containing user identifiers, and one of the parties additionally has an integer value associated with each of its user identifiers. The parties want to learn (1) the number of users they have in common and (2) the sum of the integer values associated with these users, without revealing any more information about their private inputs. Private Intersection-Sum is not an arbitrary question, but rather arose naturally and was concretely defined based on a given central business need: computing aggregate conversion rate (or effectiveness) of advertising campaigns. This problem has both great practical value and important privacy considerations, and represents a type of analysis that occurs surprisingly commonly.

Among the factors that enabled our deployment, in this work we consider in more depth the technical issue of protocol choice and its performance implications. Specifically, we present a study involving three novel protocols for computing Private Intersection-Sum, which leverage three different basic protocol techniques including Random Oblivious Transfer, encrypted Bloom filters, and Diffie–Hellman style (Pohlig–Hellman specifically) double masking. We compare the three protocols under different instantiations of an additive homomorphic encryption, which is used as a building block in each protocol. We implement our constructions and compare their actual communication and computation overheads. Finally, we analyze the advantages of the DDH-based protocol which make it the solution of choice for our deployment setting.

Leakage assessment of cryptographic implementations with side-channel analysis relies on two important assumptions: leakage model and the number of side-channel traces. In the context of profiled side-channel attacks, having these assumptions correctly defined is a sufficient first step to evaluate the security of a crypto implementation with template attacks. This method assumes that the features (leakages or points of interest) follow a univariate or multi-variate Gaussian distribution for the estimation of the probability density function. When trained machine learning or neural network models are employed as classifiers for profiled attacks, a third assumption must be taken into account that it the correctness of the trained model or learning parameters. It was already proved that convolutional neural networks have advantages for side-channel analysis like bypassing trace misalignments and defeating first-order masking countermeasures in software implementations. However, if this trained model is incorrect and the test classification accuracy is close to random guessing, the correctness of the two first assumptions (number of traces and leakage model) will be insufficient and the security of the target under evaluation can be overestimated. This could lead to wrong conclusions in leakage certifications. One solution to verify if the trained model is acceptable relies on the identifying of input features that the neural network considers as points of interest. In this paper, we implement the assessment of neural network models by using the proposed backward propagation path method. Our method is employed during the profiling phase as a tool to verify what the neural network is learning from side-channel traces and to support the optimization of hyper-parameters. The method is tested against masked AES implementation. One of the main results highlights the importance of L2 regularization for the automated points of interest selection from a neural network.

In this work, we present the rst highly-optimized implementation of Supersingular Isogeny Key Encapsulation (SIKE) submitted to NIST's second round of post quantum standardization process, on 64-bit ARMv8 processors. To the best of our knowledge, this work is the rst optimized implementation of SIKE round 2 on 64-bit ARM over SIKEp434 and SIKEp610. The proposed library is explicitly optimized for these two security levels and provides constant-time implementation of the SIKE mechanism on ARMv8-powered embedded devices. We adapt dierent optimization techniques to reduce the total number of underlying arithmetic operations on the led level. In particular, the benchmark results on embedded processors equipped with ARM Cortex- A53@1.536GHz show that the entire SIKE round 2 key encapsulation mechanism takes only 84 ms at NIST's security level 1. Considering SIKE's extremely small key size in comparison to other candidates, our result implies that SIKE is one of the promising candidates for key encapsulation mechanism on embedded devices in the quantum era.

We show how to combine a fully-homomorphic encryption scheme with linear decryption and a linearly-homomorphic encryption schemes to obtain constructions with new properties. Specifically, we present the following new results.

(1) Rate-1 Fully-Homomorphic Encryption: We construct the first scheme with message-to-ciphertext length ratio (i.e., rate) $1-\sigma$ for $\sigma = o(1)$. Our scheme is based on the hardness of the Learning with Errors (LWE) problem and $\sigma$ is proportional to the noise-to-modulus ratio of the assumption. Our building block is a construction of a new high-rate linearly-homomorphic encryption. One application of this result is the first general-purpose secure function evaluation protocol in the preprocessing model where the communication complexity is within additive factor of the optimal insecure protocol. (2) Fully-Homomorphic Time-Lock Puzzles: We construct the first time-lock puzzle where one can evaluate any function over a set of puzzles without solving them, from standard assumptions. Prior work required the existence of sub-exponentially hard indistinguishability obfuscation.

Logic locking has been proposed as an obfuscation technique to protect outsourced IC designs from Intellectual Property (IP) piracy by untrusted entities in the design and fabrication process. It obfuscates the netlist by adding extra key-gates, to mislead an adversary, whose aim is to reverse engineer the netlist. The correct functionality will be obtained only if a correct key is applied to the key-gates. The key is written into a nonvolatile memory (NVM) after the fabrication by the IP owner. In the past several years, the focus of the research community has been mostly on Oracle-guided attacks, such as SAT attacks, on logic locking and proposing proper countermeasures against such attacks. However, none of the reported researches in the literature has ever challenged a more fundamental assumption of logic locking, which is the security of the key itself. In other words, if an adversary can read out the correct key after insertion, the security of the entire scheme is broken, no matter how robust is the logic-locking scheme. In this work, we first review possible adversaries for the locked circuits and their capabilities. Afterward, we demonstrate that even with the assumption of having a tamper and read-proof memory for the key storage, which is not vulnerable to any physical attacks, the key transfer between the memory and the key-gates through registers and buffers make the key extraction by an adversary possible. To support our claim, we implemented a proof-of-concept locked circuit as well as one of the common logic locking benchmarks on a 28 nm Flash-based FPGA and extract their keys using optical probing. Finally, we discuss the feasibility of the proposed attack in different scenarios and propose potential countermeasures.

Secure multi-party computation permits evaluation of any desired functionality on private data without disclosing the data to the participants and is gaining its popularity due to increasing collection of user, customer, or patient data and the need to analyze data sets distributed across different organizations without disclosing them. Because adoption of secure computation techniques depends on their performance in practice, it is important to continue improving their performance. In this work, we focus on common non-trivial operations used by many types of programs, and any advances in their performance would impact the runtime of programs that rely on them. In particular, we treat the operation of reading or writing an element of an array at a private location and integer multiplication. The focus of this work is secret sharing setting with honest majority in the semi-honest security model. We demonstrate improvement of the proposed techniques over prior constructions via analytical and empirical evaluation.

We present a framework for the study of a learning problem over abstract groups, and introduce a new technique which allows for public-key encryption using generic groups. We proved, however, that in order to obtain a quantum resistant encryption scheme, commutative groups cannot be used to instantiate this protocol.

The Hidden Subgroup Problem (HSP) aims at capturing all problems that are susceptible to be solvable in quantum polynomial time following the blueprints of Shor's celebrated algorithm. Successful solutions to this problems over various commutative groups allow to efficiently perform number-theoretic tasks such as factoring or finding discrete logarithms.

The latest successful generalization (Eisentrager et al. STOC 2014) considers the problem of finding a full-rank lattice as the hidden subgroup of the continuous vector space $\mathbb R^m$, even for large dimensions $m$. It unlocked new cryptanalytic algorithms (Biasse-Song SODA 2016, Cramer et al. EUROCRYPT 2016 and 2017), in particular to find mildly short vectors in ideal lattices.

The cryptanalytic relevance of such a problem raises the question of a more refined and quantitative complexity analysis. In the light of the increasing physical difficulty of maintaining a large entanglement of qubits, the degree of concern may be different whether the above algorithm requires only linearly many qubits or a much larger polynomial amount of qubits.

This is the question we start addressing with this work. We propose a detailed analysis of (a variation of) the aforementioned HSP algorithm, and conclude on its complexity as a function of all the relevant parameters. Incidentally, our work clarifies certain claims from the extended abstract of Eisentrager et al.