Secret-Key Generation from Private Identifiers under Channel Uncertainty
Vamoua Yachongka, Rémi A. Chou · 2025
Read Secret-Key Generation from Private Identifiers under Channel Uncertainty by Vamoua Yachongka, Rémi A. Chou (2025) 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-2503-08632.txt.
Introduction
Vamoua Yachongka and Rémi Chou study secret keys built from a device's own fingerprint, in a 2025 paper. Each chip has small flaws no one can clone, and those flaws can seed a secret key. One terminal makes the key and stores public helper data. A second terminal uses its own reading of the chip, plus the helper data, to rebuild the key.
An eavesdropper may hold her own reading and can see the helper data too. The study asks how much key can be made while little leaks to her. The twist is that the channel stats are not known for sure. The authors find inner and outer bounds on the best trade-off among key size, storage, and leakage. For Gaussian sources, the bounds meet, so the trade-off is exact.
Catalog record
| CREATOR | Vamoua Yachongka, Rémi A. Chou |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 100 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-08632
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Secret-Key Generation from Private Identifiers under Channel Uncertainty Vamoua Yachongka, Rémi A. Chou Abstract This study investigates secret-key generation for device authentication using physical identifiers, such as responses from physical unclonable functions (PUFs). The system includes two legitimate terminals (encoder and decoder) and an eavesdropper (Eve), each with access to different measurements of the identifier. From the device identifier, the encoder generates a secret key, which is securely stored in a private database, along with helper data that is saved in a public database accessible by the decoder for key reconstruction. Eve, who also has access to the public database, may use both her own measurements and the helper data to attempt to estimate the secret key and identifier. Our setup focuses on authentication scenarios where channel statistics are uncertain, with the involved parties employing multiple antennas to enhance signal reception. Our contributions include deriving inner and outer bounds on the optimal trade-off among secret-key, storage, and privacy-leakage rates for general discrete sources, and showing that these bounds are tight for Gaussian sources. I Introduction The Internet of Things (IoT) is a rapidly growing technology that enables numerous sensors and small-chip devices to interact and exchange information over the internet. However, ensuring security and privacy in IoT communications presents significant challenges compared to conventional networks due to the diverse range of applications and resource constraints of these devices [ 2 ] . To help address these difficulties, recent efforts have focused on developing security protocols at the physical layer for authenticating devices. Secret-key generation using physical identifiers, such as responses from physical unclonable functions (PUFs), is a promising protocol for device authentication because it offers several advantages, including simple designs, low costs, and eliminating the need to save the secret key on the device [ 3 ] . A PUF is defined as a physical function that for a given input (challenge), provides an output (response) that serves as a unique identifier for each device [ 4 , 5 ] . Some examples of PUFs are static random access memory (SRAM) PUFs and ring-oscillator (RO) PUFs. Fig. 1 : An authentication scheme based on secret-key generation with PUFs. The eavesdropper (Eve) is a passive adversary who is interested in learning the secret key and the source identifier, but does not interfere with the communication mechanism of the system. Secret-key authentication using PUFs is illustrated in Fig. 1 [ 6 ] 1 1 1 [ 6 ] does not consider the presence of Eve, but is included in the figure to facilitate understanding of our system model in Section II . and consists of an enrollment phase and an authentication phase. During the enrollment phase, the terminal (encoder) challenges the device, i.e., the PUF embedded in the device, with a challenge C and gets a response \widetilde{\rm R} , from which the encoder generates a secret key and helper data. The secret key is securely stored in a private database, while the helper data are saved in a public database, which can be accessed by both the decoder and Eve. In the authentication phase, the terminal (decoder) challenges the device by sending the same challenge C, which produces a different response R due to noise effects. The decoder reconstructs the secret key based on both the response and helper data from the public database and then compares it with the one saved in the private database. If they match, the device is successfully authenticated; otherwise, the authentication fails. In subsequent discussions, the response of a PUF unaffected by noise is referred to as the source identifier . The responses of a PUF observed at the terminals and Eve through communication channels are called the observed identifiers . It is worth noting that the source identifiers are assumed to be fixed and thus the secret-key generation model considered in this paper corresponds to the source-type model, unlike the channel-type models, where the source distribution can be controlled [ 7 ] . I-A Motivations We study the capacity region of secret-key generation from source identifiers for a setup involving compound authentication channels with multiple outputs. The motivation for considering compound-channel settings is to capture a situation in which device authentication takes place in environments where the channel statistics may not be perfectly known. This contrasts with most previous studies, which assume that the encoder and decoder have perfect knowledge of the source and channel statistics of the systems. For example, as shown in Fig. 1 , consider a situation where the decoder needs to authenticate a flying drone. As the channel state information (CSI) of the channel from the drone to the decoder may fluctuate, it makes it difficult for the decoder to obtain the exact CSI. Additionally, Eve is unlikely to share her CSI with the legitimate terminals. Thus, compound channels are used to model the channels to the decoder and Eve. In this setting, the encoder and decoder do not possess precise CSI of the relevant channels but are aware that these channels belong to certain predefined sets. Additionally, we consider multiple outputs for the channels to the encoder, decoder, and Eve to capture the circumstance where these parties may deploy multiple antennas to enhance signal reception. Note that having more antennas can increase the correlation between Eve’s observation and the source identifier, giving her an advantage in learning the secret key and the source identifier. Hence, in this setting, we want to quantify the potential leakage to Eve from both security and privacy perspectives. Finally, in practice, certain types of PUFs produce continuous-value identifiers. For example, the source of RO PUFs can be modeled as a Gaussian distribution [ 8 ] . Additionally, a number of communication channels are sometimes approximated as additive white Gaussian noise (AWGN) channels. This motivates us to study setups with Gaussian sources and AWGN channels. I-B Related Work Secret-key generation using PUFs 2 2 2 PUF and biometric identifiers share similar characteristics, and thus, the theoretical results developed for one can be applied to the other as well [ 3 ] . has been studied from information-theoretic perspectives in [ 9 , 10 ] . Later, several extensions of this model were found in [ 11 , Ch. 4] for Gaussian sources, [ 12 ] for separated and combined enrollments, and [ 13 ] for multiple rounds of enrollments and authentications. Limited storage rate was introduced to the model in [ 14 ] . Furthermore, the fundamental limits among secret-key, storage, and privacy-leakage rates when Eve also has a correlated sequence of the source were characterized in [ 15 ] for discrete sources and [ 16 ] for Gaussian sources. This model is similar to the key-agreement problem with forward communication only studied in [ 17 ] , [ 18 ] , [ 19 ] , but an additional privacy constraint is imposed in the problem formulation to limit information leakage on the source identifier. More recent works have considered a setup that incorporates a noisy channel in the enrollment phase [ 20 , 21 , 22 ] . The channel is modeled to account for the noise introduced to the source identifier during the enrollment process, providing a more general framework, as signals generated