Combined PUF Logic for Harder-to-Model Device Authentication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Physical Unclonable Functions (PUFs) are vulnerable to machine learning (ML) attacks, as adversaries can intercept challenge-response pairs and create models of the PUF, compromising security.
Innovation Solution
A combined PUF is designed by combining the responses of multiple PUFs using logic gates and a combination function, such as the Tribes function, to generate a single, more robust response, thereby increasing hardness against ML attacks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple PUFs are combined to increase security hardness, then resistance to ML attacks is improved, but device complexity increases
Solution Approach 1:
The patent combines multiple PUF instances into a single combined PUF system where responses from individual PUFs are merged through logic gates and combination functions. This merging approach increases security hardness by requiring attackers to compromise multiple PUFs simultaneously, while the systematic integration manages the complexity through structured combination rather than ad-hoc assembly.
Solution Approach 2:
The combined PUF is segmented into multiple independent PUF instances, each contributing to the overall security. By dividing the security function into separate PUF units that can be independently implemented and combined, the system achieves higher security hardness while managing complexity through modular design. Each segment (individual PUF) maintains simplicity while the collective provides enhanced security.
2Reliability
If a combination function is used to merge PUF responses, then robustness against ML attacks is improved, but the difficulty of detecting and measuring increases
Solution Approach 1:
The combination function acts as an intermediary layer between individual PUF responses and the final output. This intermediary transforms multiple PUF responses into a unified combined response, increasing robustness against ML attacks by obscuring the individual PUF characteristics. The intermediary function makes detection and measurement more difficult while maintaining a clear input-output relationship for legitimate verification.
Solution Approach 2:
The combination function changes the parameters of the PUF system by aggregating multiple responses into a single output with different statistical and functional properties. This parameter transformation increases robustness against ML attacks by altering the response distribution and making it harder for attackers to model individual PUF behaviors, while the deterministic nature of the combination function maintains measurability for authorized verification.
3Reliability
If attribute noise is added to the PUF system, then security against ML attacks is improved, but manufacturing precision requirements increase
Solution Approach 1:
The patent creates a composite PUF system by combining multiple PUF instances with different attribute noise characteristics. This composite approach increases security against ML attacks by leveraging the diverse noise properties of individual PUFs, making it harder for attackers to model the combined system. The composite structure manages manufacturing precision requirements by allowing each individual PUF to be manufactured within standard tolerances while the combination provides enhanced security.
Data Source
AI summary
Combined physical unclonable function (PUFs); methods, apparatuses, systems, and computer program products for enrolling combined PUFs; and methods, apparatuses, systems, and computer program products for authenticating a device physically associated with a combined PUF are described. In an example embodiment, a combined PUF includes a plurality of PUFs and one or more logic gates. Each PUF includes a plurality of stages and an arbiter configured to generate a single PUF response based on response portions generated by the plurality of stages. The one or more logic gates are configured to combine the single PUF response for each of the plurality of PUFs in accordance with a combination function to provide a combined response.