International Association
for Cryptologic Research

The sum of two $n$-bit pseudorandom permutations is known to behave like a pseudorandom function with $n$ bits of security. A recent line of research has investigated the security of two public $n$-bit permutations and its degree of indifferentiability. Mandal et al. (INDOCRYPT 2010) proved $2n/3$-bit security, Mennink and Preneel (ACNS 2015) pointed out a non-trivial flaw in their analysis and re-proved $2n/3$-bit security. Bhattacharya and Nandi (EUROCRYPT 2018) eventually improved the result to $n$-bit security. Recently, Gunsing at CRYPTO 2022 already observed that a proof technique used in this line of research only holds for sequential indifferentiability. We revisit the line of research in detail, and observe that the strongest bound of $n$-bit security has two other serious issues in the reasoning, the first one is actually the same non-trivial flaw that was present in the work of Mandal et al., while the second one discards biases in the randomness influenced by the distinguisher. More concretely, we introduce two attacks that show limited potential of different approaches. We (i) show that the latter issue that discards biases only holds up to $2^{3n/4}$ queries, and (ii) perform a differentiability attack against their simulator in $2^{5n/6}$ queries. On the upside, we revive the result of Mennink and Preneel and show $2n/3$-bit regular indifferentiability security of the sum of public permutations.

We consider SPN schemes, i.e., schemes whose non-linear layer is defined as the parallel application of t ≥ 1 independent S-Boxes over F2n and whose linear layer is defined by the multiplication with a (n · t) × (n · t) matrix over F2. Even if the algebraic representation of a scheme depends on all its components, upper bounds on the growth of the algebraic degree in the literature usually only consider the details of the non-linear layer. Hence a natural question arises: (how) do the details of the linear layer influence the growth of the algebraic degree? We show that the linear layer plays a crucial role in the growth of the algebraic degree and present a new upper bound on the algebraic degree in SP-networks. As main results, we prove that in the case of low-degree round functions with large S-Boxes: (a) an initial exponential growth of the algebraic degree can be followed by a linear growth until the maximum algebraic degree is reached; (b) the rate of the linear growth is proportional to the degree of the linear layer over Ft2n. Besides providing a theoretical insight, our analysis is particularly relevant for assessing the security of the security of cryptographic permutations designed to be competitive in applications like MPC, FHE, SNARKs, and STARKs, including permutations based on the Hades design strategy. We have verified our findings on small-scale instances and we have compared them against the currently best results in the literature, showing a substantial improvement of upper bounds on the algebraic degree in case of low-degree round functions with large S-Boxes.

First of all we take a thorough look at an error in a paper by Daemen et al. (ToSC 2018) which looks at minimal requirements for tree-based hashing based on multiple primitives, including block ciphers. This reveals that the error is more fundamental than previously shown by Gunsing et al. (ToSC 2020), which is mainly interested in its effect on the security bounds. It turns out that the cause for the error is due to an essential oversight in the interaction between the different oracles used in the indifferentiability proofs. In essence, it reduces the claim from the normal indifferentiability setting to the weaker sequential indifferentiability one. As a matter of fact, this error appeared in multiple earlier indifferentiability papers, including the optimal indifferentiability of the sum of permutations (EUROCRYPT 2018) and the recent ABR+ construction (EUROCRYPT 2021). We discuss in detail how this oversight is caused and how it can be avoided. We next demonstrate how the negative effects on the security bound of the construction by Daemen et al. can be resolved. Instead of only allowing a truncated output, we generalize the construction to allow for any finalization function and investigate the security of this for five different types of finalization. Our findings, among others, show that the security of the SHA-2 mode does not degrade if the feed-forward is dropped and that the modern BLAKE3 construction is secure in principle but that its use of the extendable output requires its counter used for random access to be public. Finally, we introduce the tree sponge, a generalization of the sequential sponge construction with parallel absorbing and squeezing.

We present two tweakable wide block cipher modes from doubly-extendable cryptographic keyed (deck) functions and a keyed hash function: double-decker and docked-double-decker. Double-decker is a direct generalization of Farfalle-WBC of Bertoni et al. (ToSC 2017(4)), and is a four-round Feistel network on two arbitrarily large branches, where the middle two rounds call deck functions and the first and last rounds call the keyed hash function. Docked-double-decker is a variant of double-decker where the bulk of the input to the deck functions is moved to the keyed hash functions. We prove that the distinguishing advantage of the resulting wide block ciphers is simply two times the sum of the pseudorandom function distinguishing advantage of the deck function and the blinded keyed hashing distinguishing advantage of the keyed hash functions. We demonstrate that blinded keyed hashing is more general than the conventional notion of XOR-universality, and that it allows us to instantiate our constructions with keyed hash functions that have a very strong claim on bkh security but not necessarily on XOR-universality, such as Xoofffie (ePrint 2018/767). The bounds of double-decker and docked-double-decker are moreover reduced tweak-dependent, informally meaning that collisions on the keyed hash function for different tweaks only have a limited impact. We describe two use cases that can exploit this property opportunistically to get stronger security than what would be achieved with prior solutions: SSD encryption, where each sector can only be written to a limited number of times, and incremental tweaks, where one includes the state of the system in the variable-length tweak and appends new data incrementally.

A well-established PRP-to-PRF conversion design is truncation: one evaluates an $n$-bit pseudorandom permutation on a certain input, and truncates the result to $a$ bits. The construction is known to achieve tight $2^{n-a/2}$ security. Truncation has gained popularity due to its appearance in the GCM-SIV key derivation function (ACM CCS 2015). This key derivation function makes four evaluations of AES, truncates the outputs to $n/2$ bits, and concatenates these to get a $2n$-bit subkey. In this work, we demonstrate that truncation is wasteful. In more detail, we present the Summation-Truncation Hybrid (STH). At a high level, the construction consists of two parallel evaluations of truncation, where the truncated $(n-a)$-bit chunks are not discarded but rather summed together and appended to the output. We prove that STH achieves a similar security level as truncation, and thus that the $n-a$ bits of extra output is rendered for free. In the application of GCM-SIV, the current key derivation can be used to output $3n$ bits of random material, or it can be reduced to three primitive evaluations. Both changes come with no security loss.

In ToSC 2018(4), Daemen et al. performed an in-depth investigation of sound hashing modes based on arbitrary functions, permutations, or block ciphers. However, for the case of invertible primitives, there is a glitch. In this errata, we formally fix this glitch by adding an extra term to the security bound, q/2b−n, where q is query complexity, b the width of the permutation or the block size of the block cipher, and n the size of the hash digest. For permutations that are wider than two times the chaining value this term is negligible. For block cipher based hashing modes where the block size is close to the digest size, the term degrades the security significantly.