Dendritic Identifiers as Oracles in Microelectronics

Improvements in transparency, trust, and assurance in microelectronics manufacturing demand a new approach to establishing the identity of individual parts. Whereas permissioned blockchains are now being used to secure critical supply chains, the microelectronics industry still lacks a suitable oracle - the element which acts as the bridge between the blockchain and items in the real world - that meets cost, security, and physical compatibility requirements. In this paper, we present our pathfinding work on an approach to this issue, the Dendritic Identifier (DI), which comprises an inexpensive and physically scalable 'fingerprint', which may be applied directly to components at trusted points in the supply chain, coupled with feature detection and matching software. Item-level uniqueness emerges from the high pattern (information) entropy caused by chaotic instabilities in the material systems used, with machine readability and robustness arising from the strong keypoints and low structural entropy of rule-based pattern formation. The subtle relief in the pattern yields security and copy resistance via patterns of reflected light that change with viewing or illumination angle. In the proposed use model, pattern keypoints and reflected light constellations would be extracted from reference images, encoded to complete the digital bridge, and subsequently used to establish component identity and authenticity from similarly converted images taken at various critical interaction points along the manufacturing supply chain.