Reflective-Physically Unclonable Function based System for Anti-Counterfeiting

Product security techniques, in particular those targeting counterfeit goods, are under increased focus due to proliferation of supply and distribution chains across the world. Counterfeit goods are not restricted to specific product categories but pose problems across the board. In addition to loss of revenue, the usefulness of the product, functionality and safety are undermined by counterfeit products. A security technique that can be harmoniously applied across product categories as an anti-counterfeiting measure is explored in this dissertation. Physically unclonable functions (PUFs) are physical security mechanisms, which utilize inherent randomness in processes used to instantiate physical objects. In this dissertation, an extensive overview of both, the state of the art in implementations and the accompanying literature dealing with definition and analysis is provided. Although this is a relatively new domain, one can find established methods and metrics that can be applied in analysis of PUFs. We review these, while presenting our priorities in the framework of anti-counterfeiting application. Focus on experimental verification of unclonability, is a marked difference in our approach when compared with other implementations and analyses in this domain. The concept of the reflective physically unclonable function (r-PUF) is presented as a product security solution. The viability of the concept, while evaluating and defining the requirements of such a system is explored. The evolution process of the system design is elaborated, which allows for understanding of the compromises that were reached. The anti-counterfeiting system is a combination of physical modules –instantiation, registration, verification and software modules –data extraction, algorithms for processing the data and application layers to provide functionality. During the dissertation, end-to-end functionality of an anti-counterfeiting system using r-PUF was verified. The system variables in the r-PUF instantiation process were identified and their influence in the system was studied. Experiments were designed to understand the impact of individual variables and results were analysed. We were able to conclusively prove the unclonability of r-PUFs in an anti-counterfeiting scenario.