As electronic devices become increasingly interconnected and pervasive in people's lives, security, trustworthy computing, and intellectual property (IP) protection have notably emerged as important challenges for the next decade. The assumption that hardware is trustworthy and that security effort should only be focused on networks and software is no longer valid given globalization of integrated circuits and systems design and fabrication. The Semiconductor Industry Association pegged the cost of electronics counterfeiting at US $7.5 billion per year in lost revenue and tied it to the loss of 11,000 U.S. jobs. From a national defense perspective, unsecured devices can be compromised by the enemy, putting military personnel and equipment in danger. Therefore, securing integrated circuit (IC) chips, in other words, hardware security, is extremely important. This dissertation considers the design of highly secure digital signal processing circuits by employing both authentication-based and obfuscation-based approaches. In the first part of the dissertation, we focus on one emerging authentication-based solution: Physical Unclonable Function (PUF). We present novel reconfigurable PUF designs which could simultaneously achieve better reliability and security. We also present a systematic statistical analysis to quantitatively evaluate the performances of various multiplexer (MUX)-based PUFs. The statistical analysis results can be used to predict the relative advantages of various MUX-based PUF designs. These results can be used by the designer to choose a proper type of PUF or appropriate design parameters for a certain PUF based on the requirements of a specific application. Furthermore, a lightweight PUF-based local authentication scheme is also proposed, which eliminates the use of error correcting codes. In the next part of the dissertation, we consider another hardware protection method: obfuscation. Hardware obfuscation is a technique by which the description or the structure of electronic hardware is modified to intentionally conceal its functionality, which makes it significantly more difficult to reverse engineer. Unlike these prior works, We start to look at Digital Signal Processing (DSP) circuits. In the literature, security aspect of DSP circuits has only attracted little attention. However, high-level transformations of DSP circuit are intrinsically suitable for hardware obfuscation, as these techniques only alter the structure of a circuit, while maintaining the original functionality. Based on this finding, we present a novel design methodology for obfuscated DSP circuits by hiding functionality via high-level transformations. The key idea is to generate meaningful and non-meaningful design variations by using high-level transformations. In the final part of the dissertation, we consider the design and analysis of True Random Number Generator (TRNG), which is also an important topic in hardware security. We examine the modeling and statistical aspects of the proposed TRNG circuit. According to our model, we show that the performance of the beat-frequency detector based TRNG (BFD-TRNG) can be improved by appropriately adjusting design parameters. Motivated by the our analysis, we propose several alternate BFD-TRNG designs which could achieve improved performance. Various post-processing methods which are specific to the proposed designs are also studied.
University of Minnesota Ph.D. dissertation. July 2015. Major: Electrical Engineering. Advisor: Keshab Parhi. 1 computer file (PDF); 202 pages.
Authentication and Obfuscation of Digital Signal Processing Integrated Circuits.
Retrieved from the University of Minnesota Digital Conservancy,
Content distributed via the University of Minnesota's Digital Conservancy may be subject to additional license and use restrictions applied by the depositor.