Extraction of Device Fingerprints Using Built-in Erase-Suspend Operation of Flash Memory Devices

The-Nghia Nguyen, Sunghyun Park, Donghwa ShinIEEE Access 2020 · DOI

Abstract

The reliability and efficiency of a physically unclonable function (PUF) considerably depends on that of the random seed number generation process in the target hardware technology. Among the proposed hardware PUF techniques, flash memory-based approaches have several advantages because of the widespread use of flash memories in electronic devices. The operations of the flash memories such as read, program, and erase have been utilized to generate the random number based on their random process variations. In this work, we propose a random number generation method for the flash memory-based hardware PUF applications that utilize the intrinsic erase-suspend operation in modern flash memory devices. Unlike the conventional methods, the proposed method does not require any modification of the device or additional peripheral circuitry to control the operations. We evaluate the proposed method from the perspective of reliability and efficiency. The experimental results show that the proposed method can generate a 16-bit random number in 1 ms approximately while maintaining greater than 95 % reproductivity of the random number.

Introduction

In the modern globalized semiconductor market, the proliferation of counterfeit devices poses significant security and economic risks. Establishing a "Root of Trust" through hardware-based identification is essential for securing supply chains. Physically Unclonable Functions (PUFs) serve as digital biometrics for silicon, leveraging inherent manufacturing variations to create unique, non-reproducible identifiers for individual chips.

Figure 1: Conceptual diagram of device authentication in the supply chain using flash-based fingerprints.

While various PUF implementations exist, many require specialized circuitry or modifications to standard hardware. Previous flash-based PUFs often relied on stressing the memory through non-standard operations, which can degrade device longevity or require complex control logic. This research introduces a novel method to extract unique fingerprints from off-the-shelf flash memory by exploiting the standard, built-in erase-suspend operation.

Methodology

The proposed method utilizes the erase-suspend feature common in modern flash memory, which allows the system to pause a lengthy erase operation to prioritize other tasks.

The Physics of Randomness

The core of this technique lies in microscopic manufacturing variations, particularly the thickness of the tunneling oxide layer in memory cells. These variations cause different cells to erase at slightly different rates. By suspending the erase operation at a precise microsecond, a subset of "stubborn" bits remains unerased (logic 0) while others transition to the erased state (logic 1).

Figure 2: Erasure rate and its gradient as a function of suspension time.

Fingerprint Extraction Process

The extraction involves two main phases:

  1. Characterization: The system identifies the "long tail" of the erase process—the specific time window where the erasure rate stabilizes. Bits that are consistently the last to erase are highly reproducible and form the basis of the fingerprint.
  2. Incremental Generation: To account for environmental factors like temperature, which affects erase speeds, the authors developed an incremental erase-suspend algorithm. Instead of relying on a fixed timer, the controller performs short erase bursts and counts the resulting bits until the target fingerprint density is achieved.

The experimental setup utilized Microchip SST39VF1601 flash memory and an Arduino Due-based controller, tested across a temperature range of 258 K to 313 K (-15°C to 40°C).

Results and Discussion

The proposed method demonstrates high efficiency and reliability without hardware overhead.

  • Performance: The system can regenerate a 16-bit fingerprint in approximately 1 ms once characterized.
  • Reliability: The method achieved over 95% reproducibility. By focusing on the "long tail" of the erasure curve, the selected bits remain consistent even under varying thermal conditions.
  • Uniqueness and Randomness: The Inter-chip Hamming Distance was measured at 48.57%, very close to the ideal 50% for unique identification. Bit-level randomness was recorded at 47.87%.

Limitations

While promising, the method faces challenges regarding long-term aging effects, as repeated program/erase cycles may eventually shift cell characteristics. Additionally, scaling the fingerprint length for high-security cryptographic applications (e.g., 128-bit or 256-bit) would require utilizing more memory sectors and increasing characterization time.

Conclusion

This work presents a practical and efficient approach to hardware security by repurposing standard flash memory operations. By leveraging the intrinsic erase-suspend functionality, the method provides a reliable PUF with zero hardware overhead, making it a viable solution for authenticating devices in existing supply chains.

Summary updated

Abstract

The reliability and efficiency of a physically unclonable function (PUF) considerably depends on that of the random seed number generation process in the target hardware technology. Among the proposed hardware PUF techniques, flash memory-based approaches have several advantages because of the widespread use of flash memories in electronic devices. The operations of the flash memories such as read, program, and erase have been utilized to generate the random number based on their random process variations. In this work, we propose a random number generation method for the flash memory-based hardware PUF applications that utilize the intrinsic erase-suspend operation in modern flash memory devices. Unlike the conventional methods, the proposed method does not require any modification of the device or additional peripheral circuitry to control the operations. We evaluate the proposed method from the perspective of reliability and efficiency. The experimental results show that the proposed method can generate a 16-bit random number in 1 ms approximately while maintaining greater than 95 % reproductivity of the random number.
BibTeX
@inproceedings{nguyen2020extraction,
  author = {The-Nghia Nguyen and Sunghyun Park and Donghwa Shin},
  title = {{Extraction of Device Fingerprints Using Built-in Erase-Suspend Operation of Flash Memory Devices}},
  booktitle = {IEEE Access},
  year = {2020},
  doi = {10.1109/access.2020.2995891},
}
Figure