edfas.org ELECTRONIC DEV ICE FA I LURE ANALYSIS | VOLUME 24 NO . 4 12 MEMOMETER: MEMORY PUF-BASED HARDWARE METERING METHODOLOGY FOR FPGAs Anvesh Perumalla and John M. Emmert Department of Electrical and Computer Engineering, University of Cincinnati, Ohio john.emmert@uc.edu EDFAAO (2022) 4:12-21 1537-0755/$19.00 ©ASM International® INTRODUCTION Security, assurance, and trust (SA&T) within the integrated circuit (IC) supply chain are of crucial importance to the government and the commercial sector. As the semiconductor business has shifted toward a horizontal (“fabless”) model, there is an increasing need to protect against counterfeiting, cloning, overbuilding, and intellectual property (IP) theft andpiracy. TheU.S. Department of Commerce defined a counterfeit electronic part as: 1) an unauthorized copy, 2) does not conform to original component manufacturer (OCM) design, model, and/ or performance standards, 3) is not produced by OCM or is produced by unauthorized contractors, 4) an offspecification, defective, or used OCM product sold as “new” or working, or 5) has incorrect or false markings and/or documentation.[1] A recent 2020 report by the Semiconductor Industry Association (SIA) states that the “United States today nowonly accounts for 12.5%of total installed semiconductor manufacturing, with more than 80% of production now happening in Asia.”[2] This SIA report also shows that state-of-the-art 7-nmand below IC production is happening almost exclusively outside of the United States. This creates an opportunity for untrusted agents and entities to counterfeit and overproduce ICs and place them in the supply chain. This article describes the memometer, a hardware metering technique, which addresses the supply chain integrity of field-programmable gate arrays (FPGAs). Currently, FPGAs pervade most of the semiconductor ecosystem due to their faster prototyping and time-tomarket capabilities when compared to traditional application-specific integrated circuits (ASICs). In the last three decades, the FPGA logic capacity has grown 10,000x and processing speed has grown 100x, and at the same time, the FPGA cost and energy consumption per unit function have reduced over 1000x.[3] As FPGAs have become the predominant choice of circuit realization, SA&T of these Fig. 1 An ecosystem affected by untrusted FPGAs.
RkJQdWJsaXNoZXIy MTMyMzg5NA==