Until now, it was assumed elliptic curve key exchange was cheaper than post-quantum lattice-based methods for secure connections on IoT devices like Raspberry Pi and ESP32 microcontrollers. However, researchers have demonstrated that utilising ML-KEM-512 with ML-DSA-44 in full post-quantum configurations can be both faster and more efficient. Specifically, the best combination completes TLS handshakes up to eight point nine times quicker on an ESP32 compared to traditional approaches.

This challenges previous assumptions about performance trade-offs between these cryptographic techniques. Recent findings challenge long-held beliefs about securing internet of (IoT) devices with cryptography. Testing connections via secure TLS 1.3 exchanges, the standard for online communication, revealed that fully post-quantum cryptographic systems can outperform traditional methods in both speed and efficiency on small devices like Raspberry Pi microcontrollers and ESP32 boards.

Specifically, utilising ML-KEM-512 alongside ML-DSA-44 completed handshakes up to eight point nine times faster on an ESP32 than previous approaches. Lattice-based key exchange, a mathematical puzzle used to create secret keys instead of relying on prime numbers, appears surprisingly competitive with older techniques. This challenges assumptions about performance trade-offs and raises the question: can fully quantum-safe cryptography become the new standard even on devices with limited resources.

Post-Quantum Cryptography Achieves Efficient Performance on Resource Constrained Devices

ML-KEM-512 paired with ML-DSA-44 completes Transport Layer Security (TLS) 1.3 handshakes up to 8.9 times faster on ESP32 microcontrollers compared to X25519 with RSA-PSS; this speed was previously unattainable without significantly more powerful hardware for secure connections. This breakthrough establishes a key threshold, enabling fully post-quantum cryptographic systems to operate efficiently within the severe resource constraints of many Internet of Things devices, a feat once considered impossible due to assumed performance overheads.

Benchmarking by researchers at Edinburgh Napier University revealed that utilising ML-KEM-768 alongside ML-DSA-44 in wolfSSL resulted in approximately 41% faster TLS 1.3 handshakes than traditional pairings on the Raspberry Pi Pico W and demonstrated up to 2.12 times greater energy efficiency depending on authentication choices.

Optimising Post-Quantum Cryptography in wolfSSL for Embedded Systems

ML-KEM-512/ML-DSA-44 proved key to achieving unexpectedly efficient post-quantum connections; this lattice-based key exchange functions like solving a complex grid puzzle rather than relying on prime number calculations. The Edinburgh Napier University team focused on optimising these algorithms within wolfSSL, a lightweight cryptographic library designed for resource-constrained devices such as Raspberry Pi and ESP32 boards. Through careful testing of thirty-six different combinations, classical, hybrid, and fully post-quantum TLS 1.3 configurations, they discovered that ML-KEM’s structure allowed it to perform surprisingly well even with limited processing power and memory.

Lattice Cryptography Outperforms Elliptic Curves On Embedded Systems

Governments worldwide are pushing for a shift to post-quantum cryptography in the next decade to safeguard systems against future quantum computer attacks; this creates urgent pressure on developers of internet connected devices, from smart thermostats to industrial sensors, to adopt new security standards like TLS 1.3. While lattice-based schemes such as ML-KEM and ML-DSA offer theoretical resistance, their practical implementation was historically seen as computationally expensive until now. These findings challenge the widely held belief that securing smaller devices demands compromise between speed and quantum resistance. Fully post-quantum TLS connections using ML-KEM key exchange and ML-DSA signatures are demonstrably faster *and* more energy efficient than traditional methods relying on elliptic curve cryptography like X25519. This matters because it suggests developers needn’t immediately adopt hybrid approaches combining old and new techniques but can instead prioritize streamlined, purely post-quantum implementations for enhanced security; Edinburgh Napier University demonstrated viable fully post-quantum cryptographic connections on resource-constrained Raspberry Pi microcontrollers and ESP32 boards. By utilising ML-KEM alongside ML-DSA signatures, the team achieved TLS handshakes up to eight point nine times faster on an ESP32 compared with traditional methods while maintaining comparable or improved energy efficiency.

This research demonstrates that fully post-quantum TLS 1.3 connections are achievable on IoT devices like Raspberry Pi Picos and ESP32 boards using ML-KEM key exchange paired with ML-DSA authentication. Contrary to previous assumptions, this combination completed TLS handshakes as much as 8.9 times faster than conventional pairings of X25519 and RSA-PSS on some tested hardware, and did so with equal or better energy efficiency. These results indicate developers may be able to implement purely post-quantum cryptographic solutions without sacrificing performance in resource constrained environments; the team benchmarked a matrix of thirty-six configurations within wolfSSL to reach these conclusions.

👉 More information
🗞 Testing TLS 1.3 PQC on constrained devices and microcontrollers
✍️ Olivier Gillot, William J. Buchanan and Callum Turino (Affiliation: Edinburgh Napier University)
🧠 ArXiv: https://arxiv.org/abs/2610.27766

More like this