Options
2023
Book Article
Titel
Generating Trust in Hardware through Physical Inspection
Abstract
A globally distributed semiconductor supply chain, if not properly secured, may lead to counterfeiting, malicious modification, and IP piracy. Countermeasures like a secured design flow, including development tools, and physical and functional verification methods are paramount for building a trusted supply chain. In this context, we present selected image processing methods for physical inspection that aim to provide trust in the hardware produced. We consider aspects of the manufacturing process as well as the level of the physical layout. In addition, from the perspective of trust, we discuss the potential and risks of artificial intelligence compared to traditional image processing methods. We showcase the presented methods for a 28 nm process, and propose a quantitative trust evaluation scheme on the basis of feature similarities. A process for physical analysis and inspection that is free of anomalies or failures, of course, cannot be reached. However, from the trust point of view, it is crucial to be able to identify the sources of anomalies and clearly distinguish variations of the manufacturing process and artefacts of the analysis process from the signatures of potentially malicious activities. Finally, we propose a novel quantitative trust evaluation scheme which is partially based on the physical inspection methods outlined in this work.
Author(s)