On-Demand Key Generation with Physical Unclonable Functions

You are reading version 1 of this direction, frozen 2026-10-04. The latest version is version 2.

Version 1 of 229 papers30 conceptsfrozen 2026-10-04

In traditional cryptography, a secret key is a static string of bits stored in non-volatile memory (NVM). This creates a fundamental vulnerability: if an attacker can physically access the hardware, they can use micro-probing, laser-based extraction, or side-channel analysis to "read" the secret directly from the chip. As Raphaël C-W Phan and Sung‐Ming Yen demonstrate in Amplifying Side-Channel Attacks with Techniques from Block Cipher Cryptanalysis, even attempts to mitigate one type of leak—such as pre-computing round keys to prevent timing attacks—can inadvertently increase the physical attack surface by leaving sensitive data in memory for longer periods.

The research direction of Physical Unclonable Functions (PUFs) shifts the security paradigm from storing a secret to deriving one from the unique, uncontrollable physical variations inherent in hardware manufacturing.

The Hardware Fingerprint

Every silicon chip, even those from the same wafer, possesses microscopic differences in wire thickness, gate delay, and doping concentrations. A PUF acts as a "hardware fingerprint" that challenges the device with an input and produces a response based on these unique physical traits.

Because these variations are stochastic and occur at the nanoscale, they are virtually impossible to clone, even by the original manufacturer. This allows devices to generate cryptographic keys on-the-fly. As Jiliang Zhang notes in Physical Unclonable Function-based Key Sharing for IoT Security, this eliminates the need for vulnerable on-chip key storage entirely.

Leveraging Memory Instability

One of the most practical ways to implement this is through existing memory structures. Bertrand Cambou demonstrates in A Proof of Concept SRAM-based Physically Unclonable Function (PUF) Key Generation Mechanism for IoT Devices that the power-up state of SRAM cells—which settle into a 0 or 1 based on tiny transistor mismatches—can serve as a reliable entropy source.

This is not limited to specialized hardware. Wenjie Xiong 0001 and colleagues show in Practical DRAM PUFs in Commodity Devices that even standard DRAM in smartphones can be queried for security keys by observing cell decay patterns, allowing a running system to generate a "fingerprint" without crashing.

Beyond Binary: Memristors and Ternary States

While early PUFs focused on binary responses, newer research leverages the "messy" analog properties of emerging technologies like memristors and ReRAM. Damith Ranasinghe and Yansong Gao describe a "Strong" PUF (mrSPUF) in Memristive crypto primitive for building highly secure physical unclonable functions that uses the nonlinear conductance of memristor crossbars to create an astronomical number of challenge-response pairs.

A recurring challenge with PUFs is environmental noise; a key must be identical every time it is generated. To solve this, Ashwija Reddy Korenda and Bertrand Cambou propose a "ternary-state" approach. Instead of forcing every memory cell to be a 0 or 1, the system identifies and ignores "unstable" cells that flip-flop between states, focusing only on the most robust physical features to ensure a failure rate as low as $10^{-10}$.

Applications in Hostile Environments

The ability to generate keys without storing them is transformative for devices deployed in the field where physical capture is a risk:

  • Internet of Drones: Xuanyu Chen and Chip-Hong Chang propose using PUFs combined with radio frequency fingerprinting to allow drones to authenticate each other without storing long-term credentials that could be extracted if a drone is downed.
  • Medical Implants: For devices like pacemakers, Chenglong Fu and colleagues suggest using Physical Obfuscated Keys (POKs) to prevent memory scanning attacks that could bypass security layers permanently.
  • Proof of Reality: In Secure Goods Supply Chain and Key Exchange with Virtual Proof of Reality, Yansong Gao and Damith Ranasinghe extend this concept to "Virtual Proof of Reality," where the destruction of a memristor-based PUF serves as a digital proof that a physical event (like the opening of a package) has occurred.

By turning manufacturing "defects" into a security feature, PUF-based systems move the root of trust from a vulnerable digital file to the immutable laws of physics.

Papers in this version

History & citationVersion 2 · 2026-10-04

Current · Permanent link

  1. Version 2 (latest) · 2026-10-04 · 20 papers
  2. Version 1 · 2026-10-04 · 29 papers
Cite this direction

Latest version (a living page):

@misc{lacuna_physical_key_extraction_aaed7c45,
  title = {physical key extraction},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/physical-key-extraction/rdr_aaed7c45a30046cb8987d52c57566dc1},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-20},
}

This exact version (version 1, frozen text):

@misc{lacuna_on_demand_key_generation_with_physical_u_aaed7c45_v1,
  title = {On-Demand Key Generation with Physical Unclonable Functions},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/on-demand-key-generation-with-physical-unclonable-functions/rdr_aaed7c45a30046cb8987d52c57566dc1/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-20},
}