A Statistical Side-Channel Risk Model for Timing Variability in Lattice-Based Post-Quantum Cryptography
Aayush Mainali, Sirjan Ghimire · 2025
Read A Statistical Side-Channel Risk Model for Timing Variability in Lattice-Based Post-Quantum Cryptography by Aayush Mainali, Sirjan Ghimire (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2512-22301.txt.
Introduction
Aayush Mainali and Sirjan Ghimire model the risk that lattice-based post-quantum cryptography leaks secrets through timing. These schemes' complex arithmetic and control flow can make run time depend on secret values, yet real readings are muddied by scheduling jitter, contention, and heavy-tailed delays. The authors treat leakage as a question of whether two timing distributions can be told apart under controlled conditions.
They synthesize traces for two secret classes under idle, jittery, and loaded scenarios across several leakage models, then score leakage with statistical tests and information measures combined into one ranking score. Across Kyber, Saber, and Frodo key encapsulation families, idle conditions leaked most clearly, while noise eroded the signal. Cache-index and branch-style leakage looked riskiest, enabling early design-stage comparisons.
Catalog record
| CREATOR | Aayush Mainali, Sirjan Ghimire |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 41 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-22301
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A Statistical Side-Channel Risk Model for Timing Variability in Lattice-Based Post-Quantum Cryptography Aayush Mainali, Sirjan Ghimire Abstract Timing side-channels are an important threat to cryptography that still needs to be addressed in implementations, and the advent of post-quantum cryptography raises this issue because the lattice-based schemes may produce secret-dependent timing variability with the help of complex arithmetic and control flow. Since also real timing measurements are affected by environmental noise (e.g. scheduling effects, contention, heavy tailed delays), in this work a scenario-based statistical risk model is proposed for timing leakage as a problem of distributional distinguishability under controlled execution conditions. We synthesize traces for two secret classes in idle, jitter and loaded scenarios and for multiple leakage models and quantify leakage with Welch’s t-test, KS distance, Cliff’s delta, mutual information, and distribution overlap to combine in a TLRI like manner to obtain a consistent score for ranking scenarios. Across representative lattice-based KEM families (Kyber, Saber, Frodo), idle conditions generally have the best distinguishability, jitter and loaded conditions erode distinguishability through an increase in variance and increase in overlap; cache-index and branch-style leakage tends to give the highest risk signals, and faster schemes can have a higher peak risk given similar leakage assumptions, allowing reproducible comparisons at an early design stage, prior to platform-specific validation. Keywords: post-quantum cryptography; lattice-based KEM; timing side-channel; leakage modeling; statistical distinguishability; mutual information; Kolmogorov–Smirnov distance; TLRI risk score 1 Introduction Post-quantum cryptography (PQC) is being adopted to insure against long-term secret information from adversaries with quantum capabilities. While many PQC candidates are highly secure mathematically, to make PQC a practical choice, they also need to resist to deployments at an implementation level. Timing side-channels are of particular importance because they may leak information due to small differences in execution time caused by secret-dependent control flow, the (variable) latency of arithmetics or memory accesses. In the case of lattice-based constructions, performance-critical operations such as polynomial arithmetic, transforms, encoding/decoding and validation logic can inadvertently introduce measurable timing variability even with the assumption that the underlying protocol is formally secure. One of the main challenges in the timing of side-channel analysis is that the actual timing measurements are influenced not only by effects of the secret, but also by the execution environment. Scheduling jitter, interrupts, cache contention and other system-level behaviors result in changes to the timing distributions, which will affect the leakage, masking it or, in some cases, multiplying it. As a result, the practical question is often not whether leakage exists in principle, but rather how consistently an attacker can look for secrecy differences for realistically-ranging variability and limited samples. This leads to the need for methods for assessing-malexia in which environment noise is explicitly added, and which allow comparison of the leakage risk for different algorithmic styles, platforms, and assumptions about the threat. Existing side channel research often shows convincing examples of effective attacks or suggests specific counter-measures, however, much fewer methods feature a unified and reproducible methodology to quantify and rank risk for timing leakage in multiple environments and leakage mechanisms. In particular, the variation in timing is frequently treated as a nuisance variable rather than as a first-class modelling aim, and assessments can be scattered over scheme households and measuring measurement environments. This restricts the ability of practitioners to reason in a uniform manner about which combinations of scheme, environment and leakage mechanism produce the greatest practical risk. To address these gaps, a scenario-based statistical risk model for the timing of leakage is introduced in this paper for lattice-based PQC. We define a scenario to consist of the combination of a scheme-level baseline timing scale, an environment regime (idle, jitter, loaded), a leakage model that describes the modulation of timing secrets and the leakage strength parameter which controls the magnitude of the effect. For each scenario, we obtain timing traces for two balanced secret classes and measure distinguishability with a set of complementary statistical and information-theoretic metrics. Finally, we combine these metrics into a TLRI-type composite score which assists in consistent ranking of the risk of leakage for scenarios. The main contributions of this work are: • a re-playable scenario-based simulator that separates the baseline cost, environment noise, and secret dependent leakage; • a multi-metric evaluation pipeline based on mean-based, distributional and dependency measures; • a bounded composite risk score to allow cross scenario comparison; and • empirical comparative study of representatives of lattice based KEM families including Kyber, Saber and Frodo and various leakage mechanisms Together, these contributions offer an early stage risk modelling tool that would help to guide where deeper platform specific measurements and countermeasures efforts should be prioritized. 2 Literature Review 2.1 Timing Side-Channels: Foundations and Modern Threat Surfaces Timing side channels are some of the oldest and most significant classes of implementation vulnerabilities in cryptography, as they illustrate how even with secure underlying mathematics, differences in the time to perform some computer operation may reveal something sensitive. Classical work showed that timing differences arise due to the use of conditional branches, variable-latency arithmetics, and memory access patterns can be used to recover secrets within popular cryptosystems and the foundational adversarial principle, “time is an information channel” [ 1 ] . This is a basic understanding that is still very pertinent in post-quantum cryptography (PQC) where implementations frequently involve the use of large polynomial arithmetics, structured transforms and complex decoding paths which can inadvertently create secret dependent variability. In parallel, modern systems enlarge the timing attack surface to include more than user-space cryptographic routines. Practical attacks have also demonstrated that timing can be used to disclose characteristics of the privilege address space and layout randomization properties, to illustrate the impact of the operating system and microarchitectural behavior on the timing distributions that can be measured [ 2 ] . In PQC deployments, these system-level effects are important because timing measurements can be impacted by scheduling effects, interrupts, cache contention, and power management behavior that can both mask as well as amplify leakage signals based on the environment. A lot of work is also done for memory timing and cache timing leakage by treating them as first-class threats through proposing different mitigation mechanisms to change the execution with lower timing dependence of memory access behavior. Techniques such as mitigating memory timing side-channels through changes to the execution structure [ 3 ] and cache locking with fine-grained partitioning [ 4 ] reflect the fact that mitigating timing side-channel attacks is often required at the memory hierarchy level, rather than at the algorithmic level. These insights inspire evaluation methods that explicitly account for variation of platforms and environments when reasoning about timing leakage risk. 2.2 Side-Channel and Fault Attacks in the Post-Quantum Era As PQC candidates were progressing towards standardization, research increasingly showed that post quantum primitives are vulnerable to both passive and active implementation attacks. Fault-injection analyses against Round 3 KEM candidates find that well-designed induced faults can result in only partial or complete secret recovery as well as counter security presumptions when the implementations do not come with proper countermeasures and verification measures [ 5 ] . This work focusing on timing leakage is complementary to timing leakage studies, and provides an important reminder that the security a system offers in