International Association
for Cryptologic Research

We develop an efficient algorithm to detect whether a superspecial genus 2 Jacobian is optimally (N,N)-split for each integer N <=11. Incorporating this algorithm into the best-known attack against the superspecial isogeny problem in dimension 2 gives rise to significant cryptanalytic improvements. Our implementation shows that when the underlying prime p is 100 bits, the attack is sped up by a factor 25x; when the underlying prime is 200 bits, the attack is sped up by a factor 42x; and, when the underlying prime is 1000 bits, the attack is sped up by a factor 160x.

We optimise the verification of the SQIsign signature scheme. By using field extensions in the signing procedure, we are able to significantly increase the amount of available rational 2-power torsion in verification, which achieves a significant speed-up. This, moreover, allows several other speed-ups on the level of curve arithmetic. We show that the synergy between these high-level and low-level improvements gives significant improvements, making verification 2.07 times faster, or up to 3.41 times when using size-speed trade-offs, compared to the state of the art, without majorly degrading the performance of signing.

One of the most promising avenues for realizing scalable proof systems relies on the existence of 2-cycles of pairing-friendly elliptic curves. Such a cycle consists of two elliptic curves E/GF(p) and E'/GF(q) that both have a low embedding degree and also satisfy q = #E and p = #E'. These constraints turn out to be rather restrictive; in the decade that has passed since 2-cycles were first proposed for use in proof systems, no new constructions of 2-cycles have been found. In this paper, we generalize the notion of cycles of pairing-friendly elliptic curves to study cycles of pairing-friendly abelian varieties, with a view towards realizing more efficient pairing-based SNARKs. We show that considering abelian varieties of dimension larger than 1 unlocks a number of interesting possibilities for finding pairing-friendly cycles, and we give several new constructions that can be instantiated at any security level.

<p> Isogeny-based schemes often come with special requirements on the field of definition of the involved elliptic curves. For instance, the efficiency of SQIsign, a promising candidate in the NIST signature standardisation process, requires a large power of two and a large smooth integer $T$ to divide $p^2-1$ for its prime parameter $p$. We present two new methods that combine previous techniques for finding suitable primes: sieve-and-boost and XGCD-and-boost. We use these methods to find primes for the NIST submission of SQIsign. Furthermore, we show that our methods are flexible and can be adapted to find suitable parameters for other isogeny-based schemes such as AprèsSQI or POKE. For all three schemes, the parameters we present offer the best performance among all parameters proposed in the literature. </p>

We revisit the problem of finding two consecutive $B$-smooth integers by giving an optimised implementation of the Conrey-Holm\-strom-McLaughlin ``smooth neighbors'' algorithm. While this algorithm is not guaranteed to return the complete set of $B$-smooth neighbors, in practice it returns a very close approximation to the complete set but does so in a tiny fraction of the time of its exhaustive counterparts. We exploit this algorithm to find record-sized solutions to the pure twin smooth problem, and subsequently to produce instances of cryptographic parameters whose corresponding isogeny degrees are significantly smoother than prior works. Our methods seem well-suited to finding parameters for the SQISign signature scheme, especially for instantiations looking to minimize the cost of signature generation. We give a number of examples, among which are the first parameter sets geared towards efficient SQISign instantiations at NIST's security levels III and V.

We give a new algorithm for finding an isogeny from a given supersingular elliptic curve $E/\F_{p^2}$ to a subfield elliptic curve $E'/\F_p$, which is the bottleneck step of the Delfs-Galbraith algorithm for the general supersingular isogeny problem. Our core ingredient is a novel method of rapidly determining whether a polynomial $f \in L[X]$ has any roots in a subfield $K \subset L$, while avoiding expensive root-finding algorithms. In the special case when $f=\Upphi_{\ell,p}(X,j) \in \F_{p^2}[X]$, i.e., when $f$ is the $\ell$-th modular polynomial evaluated at a supersingular $j$-invariant, this provides a means of efficiently determining whether there is an $\ell$-isogeny connecting the corresponding elliptic curve to a subfield curve. Together with the traditional Delfs-Galbraith walk, inspecting many $\ell$-isogenous neighbours in this way allows us to search through a larger proportion of the supersingular set per unit of time. Though the asymptotic $\tilde{O}(p^{1/2})$ complexity of our improved algorithm remains unchanged from that of the original Delfs-Galbraith algorithm, our theoretical analysis and practical implementation both show a significant reduction in the runtime of the subfield search. This sheds new light on the concrete hardness of the general supersingular isogeny problem (i.e. the foundational problem underlying isogeny-based cryptography), and has immediate implications on the bit-security of schemes like B-SIDH and SQISign for which Delfs-Galbraith is the best known classical attack.