Reverse engineering (RE) a register transfer level (RTL) description allows an attacker to counterfeit intellectual property (IP) as well as introduce hardware trojans. To mitigate this risk, RTL obfuscation can be employed. Most of the existing obfuscation methods are targeted at gate-level and layout-level. In this work, we propose key based RTL obfuscation scheme at an early design phase during high-level synthesis (HLS). Given a control data flow graph (CDFG), obfuscation points are identified during scheduling and obfuscation logic is inserted during the datapath generation phase. In order to keep performance overhead low, such insertion is done only on noncritical paths. We implemented the proposed obfuscation technique in an in-house HLS system and the obfuscated RTL designs were synthesized to gate-level with Synopsys Design compiler targeting 90nm CMOS technology library. Based on the experimental results on four datapath intensive benchmarks, we demonstrate that proposed approach obfuscates the design with extremely low probability of reverse engineering. For a 32-bit obfuscation key, the average area, delay, and power overheads are 2.45%, 2.65%, and 2.61% respectively, which are reasonable.
The recent surge in hardware security is significant due to offshoring the proprietary Intellectual property (IP). One distinct dimension of the disruptive threat is malicious logic insertion, also known as Hardware Trojan (HT). HT subverts the normal operations of a device stealthily. The diversity in HTs activation mechanisms and their location in design brings no catch-all detection techniques. In this paper, we propose to leverage principle features of social network analysis to security analysis of Register Transfer Level (RTL) designs against HT. The approach is based on investigating design properties, and it extends the current detection techniques. In particular, we perform both node- and graph-level analysis to determine the direct and indirect interactions between nets in a design. This technique helps not only in finding vulnerable nets that can act as HT triggering signals but also their interactions to influence a particular net to act as HT payload signal. We experiment the technique on 420 combinational HT instances, and on average, we can detect both triggering and payload signals with accuracy up to 97.37%.
Advances in technology have enabled tremendous progress in the development of a highly connected ecosystem of ubiquitous computing devices collectively called the Internet of Things (IoT). Ensuring the security of IoT devices is a high priority due to the sensitive nature of the collected data. Physically Unclonable Functions (PUFs) have emerged as critical hardware primitive for ensuring the security of IoT nodes. Malicious modeling of PUF architectures has proven to be difficult due to the inherently stochastic nature of PUF architectures. Extant approaches to malicious PUF modeling assume that a priori knowledge and physical access to the PUF architecture is available for malicious attack on the IoT node. However, many IoT networks make the underlying assumption that the PUF architecture is sufficiently tamper-proof, both physically and mathematically. In this work, we show that knowledge of the underlying PUF structure is not necessary to clone a PUF. We present a novel non-invasive, architecture independent, machine learning attack for strong PUF designs with a cloning accuracy of 93.5% and improvements of up to 48.31% over an alternative, two-stage brute force attack model. We also propose a machine-learning based countermeasure, discriminator, which can distinguish cloned PUF devices and authentic PUFs with an average accuracy of 96.01%. The proposed discriminator can be used for rapidly authenticating millions of IoT nodes remotely from the cloud server.
As the design complexity increases, the attack space for malicious modifications in the design also increases. Attackers in untrusted phases during IC design cycle may embed a Hardware Trojan (HT). Identification of such HT is complex, due to rare triggering nature of the nodes involved in the HT. For analyzing HT susceptibility, it is essential to have knowledge about the high-level signal statistics (mean, variance etc.) of the inputs of a design as this information can identify nets having rarely changing signal values. A random vector based simulation approach fails to accurately model the toggling behavior of the nets in the design and hence cannot be used to study the compromising behavior of a subverted arithmetic module instance in a design. In this paper, we propose an empirical framework for HT susceptibility analysis based on spatially correlated input streams applied to arithmetic module (adder and multiplier) architectures in Digital Signal Processing (DSP) hardware. Empirical analysis for varying input statistics of several adder architectures show that the ripple-carry and carry-lookahead adder have the lowest toggle nets depending on an attacker selectable triggering threshold, hence higher resistance to HT. On the other hand, dadda multiplier outperforms other multiplier architectures. The results indicate that for an IP integrator, HT susceptibility analysis can help choose between alternative architectures for building HT-free designs based on both circuit parameters (e.g. power, performance, and area (PPA)) and toggle behavior.