Post-Quantum Entropy as a Service for Embedded Systems

Javier Blanco-Romero, Yuri Melissa Garcia-Niño, Florina Almenares Mendoza, Daniel Díaz-Sánchez, Carlos García-Rubio, Celeste Campo · 2026

Read Post-Quantum Entropy as a Service for Embedded Systems by Javier Blanco-Romero, Yuri Melissa Garcia-Niño, Florina Almenares Mendoza, Daniel Díaz-Sánchez, Carlos García-Rubio, Celeste Campo (2026) 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-2603-10274.txt.

Introduction

Javier Blanco-Romero and co-authors build a service that sends quantum random entropy to small embedded devices over post-quantum secure channels. A Quantis device feeds the server, which offers direct quantum entropy or a mix with the system pool. On ESP32-class clients, the team adds secure DTLS links to the CoAP stack and a fast local pool for the delivered entropy. Benchmarks show the fully post-quantum setup finishing handshakes faster than the classical baseline, making strong entropy practical on tiny hardware.

Catalog record

CREATORJavier Blanco-Romero, Yuri Melissa Garcia-Niño, Florina Almenares Mendoza, Daniel Díaz-Sánchez, Carlos García-Rubio, Celeste Campo
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE37877
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.10274

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Post-Quantum Entropy as a Service for Embedded Systems Javier Blanco-Romero, Yuri Melissa Garcia-Niño, Florina Almenares Mendoza, Daniel Díaz-Sánchez, Carlos García-Rubio, Celeste Campo Abstract Embedded cryptography stands or falls on entropy quality, yet small devices have few trustworthy sources and little tolerance for heavyweight protocols. We build a Quantum Entropy as a Service (QEaaS) system that moves QRNG-derived entropy from a Quantis device to ESP32-class clients over post-quantum-secured channels. On the server side, the design exposes two paths: direct quantum entropy through a custom OpenSSL provider and mixed entropy through the Linux system pool. On the client side, we extend libcoap’s Zephyr support, integrate wolfSSL-based DTLS 1.3 into the CoAP stack, and add a BLAKE2s entropy pool that preserves the standard Zephyr extraction interface while introducing an injection API for server-provided entropy. Benchmarks on ESP32 hardware, targeting 100 iterations per configuration, show that ML-KEM-512 completes a DTLS 1.3 handshake in 313 ms on average without certificate verification, 35% faster than ECDHE P-256. Pairing ML-KEM-512 with ML-DSA-44 lowers the mean to 225 ms. Certificate verification adds roughly 194 ms for ECDSA but only 17 ms for ML-DSA-44, so the fully post-quantum configuration remains 63% faster than classical ECDHE P-256 with ECDSA even under full verification. Local BLAKE2s pool operations stay below 0.1 ms combined. On this platform, post-quantum key exchange and authentication are not only feasible; they are faster than the classical baseline. \DeclareSourcemap \maps [datatype=bibtex] \map \step [fieldsource=abstract,null] \step [fieldset=abstract,null] Keywords: QEaaS, QRNGs, Cryptography, Entropy, PQC, CoAP, IoT, Microcontrollers, Zephyr 1 Introduction Random numbers sit underneath nearly every security primitive: key generation, nonces, password salting, sequence numbers. If the entropy source is weak, the rest of the stack inherits that weakness. IoT devices make this problem harder rather than easier. They usually expose a hardware RNG, but the path from physical noise source to software API is often opaque, and vendor post-processing can hide the raw generator behaviour. A practical response is to combine several sources instead of trusting a single one [ 18 ] . In that setting, Quantum Random Number Generators (QRNGs), which draw randomness from irreducible quantum processes [ 29 ] , are attractive external inputs. The catch is transport: once entropy leaves the QRNG host, the distribution channel must itself survive the arrival of quantum-capable attackers. Post-quantum cryptography (PQC) addresses that problem through quantum-resistant algorithms [ 2 ] . NIST has now standardized the lattice-based ML-KEM for key encapsulation [ 13 , 26 ] and ML-DSA for digital signatures [ 25 ] . This paper studies that full path end to end. We build a Quantum Entropy as a Service (QEaaS) architecture that delivers QRNG-backed entropy to constrained devices over CoAP protected with post-quantum DTLS. Unlike earlier EaaS systems built around HTTP or HTTPS, our design is shaped around the constraints of embedded clients: CoAP transport, explicit local pooling, and an ESP32 implementation that can actually run the stack. 2 Background and Related Work Cryptographic systems need high-quality entropy to produce unpredictable keys. Vassilev et al. [ 36 ] outline Entropy-as-a-Service (EaaS) as a way to supply entropy from distributed sources, which motivates remote entropy distribution for cloud and IoT settings. Several systems already show that remote quantum entropy is viable. Huang et al. [ 20 ] built a cloud platform on Alibaba Cloud that combines four QRNG types and deployed it for production services such as Alipay. Kozlovičs et al. [ 23 ] proposed a remote QRNG service with quantum-safe links, X.509 certificates carrying post-quantum algorithms, and bidirectional WebSocket connections. Kumar et al. [ 24 ] generated quantum random numbers on IBM’s cloud quantum platform with Hadamard-gate circuits and validated the output with NIST SP 800-90B [ 35 ] and SP 800-22 tests. Blanco et al. [ 3 ] studied QRNG integration with post-quantum cryptography in TLS workflows. Taken together, these results show that remote quantum entropy is plausible, but they mostly target conventional computing environments. In parallel, PQC benchmarking on constrained hardware has moved from standalone kernels [ 22 ] to complete protocol stacks. Bürstinghaus-Steinbach et al. [ 14 ] integrate Kyber and SPHINCS+ into mbed TLS on ESP32 and show that lattice-based key exchange can compete with ECC. Segatz and Al Hafiz [ 30 ] then push Kyber further on ESP32 with dual-core parallelism. Nielsen et al. [ 27 ] show that post-quantum signatures are also feasible on LoRa ESP32 platforms across several NIST security levels. At the transport level, Sikeridis et al. [ 32 ] and Sosnowski et al. [ 33 ] measure PQC handshake latency in TLS 1.3 under realistic network conditions, and Tasopoulos et al. [ 34 ] extend that analysis to energy consumption on STM32 with wolfSSL. What is still missing is the intersection of these lines of work: remote entropy delivery for embedded systems, protected by post-quantum transport, over a protocol that constrained devices actually prefer. The local entropy-pool question is also largely open in this setting. Entropy pool design grounds our client-side architecture. Dodis et al. [ 17 ] formalized robustness requirements for PRNGs with entropy accumulation and showed that the Linux /dev/random pool fell short of this model. Coretti et al. [ 16 ] strengthened the framework by proving that cryptographic hash functions serve as seedless extractors, and the Linux kernel adopted this result by replacing its SHA-1/LFSR-based pool with BLAKE2s [ 1 , 18 ] , gaining better input mixing, deterministic output, and replacing the earlier LFSR-based mixing stage. Chung et al. [ 15 ] prove that Linux-DRBG achieves a 128-bit security bound in the seedless robustness model for the analyzed Linux 6.4.8 construction. 3 System Architecture Our system moves quantum-generated entropy from a centralized hardware source to embedded IoT devices. The architecture spans the whole path, from QRNG hardware to transport security to the local pool on the microcontroller, as shown in Figure 1 . A Quantis QRNG PCIe-240M [ 21 ] produces the raw quantum entropy. Linux kernel drivers expose it to user space, where our implementation offers two access paths: direct access through a custom OpenSSL provider, and mixed access through the Linux entropy pool after rng-tools folds the QRNG stream together with other system sources. The service exposes both HTTPS and secure CoAP, with post-quantum cryptography protecting each path. A CoAP-HTTP proxy compliant with RFC 7252 [ 31 ] forwards constrained-client requests carrying a Proxy-URI option to the HTTP backend. ESP32 clients running Zephyr RTOS use that path and mix the received bytes into a local BLAKE2s entropy pool. The server infrastructure [ 7 ] runs in Docker containers with automated setup scripts. Figure 1: QEaaS system architecture for quantum entropy distribution from Quantis QRNG via dual access (OpenSSL and Linux entropy pool) to ESP32 clients with BLAKE2s entropy pools, using post-quantum cryptography for HTTPS and CoAP. 3.1 Server Architecture The server architecture centers on a Quantis QRNG PCIe-240M generating quantum entropy, exposed through /dev/qrandom0 by its Linux kernel drivers. The PCIe-240M uses photon counting on an image sensor illuminated by a light-emitting diode; after post-processing via Universal-2 hash functions the output delivers 58\text{\,}\mathrm{Mbit}\text{\,}{\mathrm{s}}^{-1} of random data compliant with NIST SP 800-90A/B/C and SP 800-22 [ 21 ] . The rng-tools daemon can be configured to read from this device and inject

Read the complete text