Post-Quantum Crypto: Challenges for Embedded Applications
summary
Post-quantum standards (Kyber, Dilithium) force real engineering tradeoffs: typical low-power devices have 4–16 KB SRAM and 16–32 KB flash, so memory usage is the top barrier. Optimizations push Dilithium3 signing down to under 3 KB of RAM at roughly 2× slower, but protecting FO-transform schemes multiplies complexity: FO causes chosen-ciphertext leakage that expands the protected surface, forcing masking from ~2 shares up toward 8 shares and enabling template attacks that break protected Kyber with as few as 15,000 traces. Falcon signatures fail on many embedded chips because they require double-precision support.
AI-written summary · about
chapters (24)
- 0:00Introduction
- 1:27Embedded applications in PQC
- 2:32NIST standardization
- 4:18Global PQC efforts
- 6:02Migration challenges
- 7:14Key size constraints
- 8:19Embedded systems reality
- 9:20Dilithium memory reduction
- 11:29Classical vs PQC performance
- 12:33Signature trade-offs
- 13:57Falcon for embedded
- 14:17Side-channel attacks
- 15:45FO transform security
- 18:15CCA on PQC schemes
- 20:27Masking limitations
- 21:53Kyber masking costs
- 23:41Increased shares costs
- 24:22Theory vs practical security
- 25:06PQC migration urgency
- 25:46Embedded PQC challenges
- 26:49Market size urgency
- 27:56Critical infrastructure challenges
- 29:22FO alternatives
- 30:28Q&A
related talks
43mStatus report on Post Quantum Cryptography Standardization
45mPost‑Quantum Cryptography: The Fundamentals of Tomorrow's Digital World
32mCybersecurity in Post Quantum Era
53mPanel: Migrating to Post-Quantum Schemes
36mPost-Quantum Cryptography for Digital Signing
44mPost-quantum cryptography migration strategy for cloud services
20mBenchmarking RISC-V Post-Quantum
40mPost-Quantum