Article
Computer Science, Hardware & Architecture
Chong Xiao, Ming Tang, Sylvain Guilley
Summary: This article discusses the issue of information leakage caused by microarchitectural optimizations and prefetching, and proposes a side-channel attack method based on IP-based stride prefetcher, successfully recovering the secret key of AES-128.
JOURNAL OF SYSTEMS ARCHITECTURE
(2023)
Article
Computer Science, Hardware & Architecture
S. L. P. S. K. Patanjali, Sandip Ray, Swarup Bhunia
Summary: The adoption of non-traditional design techniques by chip designers is driven by the need to meet growing user requirements and manufacturing challenges at lower technology nodes. Popular in recent years, 2.5D/3DIC stacking allows chip manufacturers to integrate complex IPs without design penalties. However, a completely untrusted supply chain poses challenges for verification and testing, making trust assurance of these designs difficult.
IEEE TRANSACTIONS ON COMPUTERS
(2023)
Article
Computer Science, Information Systems
Zihao Wang, Shuanghe Peng, Wenbin Jiang, Xinyue Guo
Summary: Hardware prefetching can interfere with Flush+Reload cache side channel attack, but an improved method provided in this paper minimizes this impact effectively.
Proceedings Paper
Computer Science, Hardware & Architecture
Leon Li, Alex Orailoglu
Summary: Logic obfuscation protects semiconductor IPs by concealing implementation details with tamper-proof key. With key recovery attacks becoming more sophisticated, this work proposes a keyless obfuscation approach that relies on internally-generated and constantly-changing secrets. The proposed methodology achieves reverse engineering resilience and is proven to be practical and secure through experimental results.
2023 IEEE 41ST VLSI TEST SYMPOSIUM, VTS
(2023)
Article
Engineering, Electrical & Electronic
Pengfei Qiu, Qiang Gao, Chang Liu, Dongsheng Wang, Yongqiang Lyu, Xiaoyong Li, Chunlu Wang, Gang Qu
Summary: Performance Monitor Unit (PMU), an important hardware module in mainstream processors, is capable of recording some events triggered in transient executions, leading to a hardware vulnerability. We propose a new kind of side channel attack utilizing this vulnerability, which enables attackers to maliciously leak secret data. Through thorough study on PMU counters of five Intel processors, we find that 112 vulnerable PMU counters can be utilized in the attack to leak secret data protected by Intel Software Guard Extensions (SGX), with a throughput of up to 291.2 bytes per second (Bps) and an average error rate of 2.45%.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I-REGULAR PAPERS
(2023)
Review
Chemistry, Multidisciplinary
Maria Mendez Real, Ruben Salvador
Summary: This paper surveys the state-of-the-art physical SCA attacks on embedded DNN implementations, providing a detailed classification and analysis, including attack types and future research directions.
APPLIED SCIENCES-BASEL
(2021)
Article
Computer Science, Hardware & Architecture
Yuzhe Gu, Ming Tang, Quancheng Wang, Han Wang, Haili Ding
Summary: This paper proposes a cache design called ExtendCache, which defends against information leakage by focusing on cache sets that may cause vulnerability. It extends the cache set by borrowing cache lines from another set, preventing attackers from controlling the state of the target set and observing effective victim behavior. Evaluation using SPEC2017 benchmarks shows minimal impact on performance, with an average performance loss of only 0.48%.
JOURNAL OF SYSTEMS ARCHITECTURE
(2023)
Article
Engineering, Electrical & Electronic
Alan Rodrigo Diaz Rizo, Julian Leonhard, Hassan Aboushady, Haralampos-G Stratigopoulos
Summary: This paper demonstrates system-level locking for RF transceivers as an anti-piracy security technique for the first time. By leveraging a state-of-the-art logic locking technique to obfuscate digital blocks in the signal path, the locking strategy makes RF performance key-dependent. The technique offers several advantages, including general applicability, attack resilience, transparency when the correct key is used, and minimum overheads. A proof-of-concept is shown through hardware measurements.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS
(2022)
Article
Computer Science, Hardware & Architecture
Lilas Alrahis, Satwik Patnaik, Muhammad Abdullah Hanif, Muhammad Shafique, Ozgur Sinanoglu
Summary: The article highlights the success of graph neural networks (GNNs) in detecting intellectual property (IP) piracy and hardware Trojans (HTs), while also addressing the susceptibility of GNNs to data poisoning attacks. The authors propose PoisonedGNN as the first backdoor attack on GNNs in the context of hardware design and demonstrate its effectiveness through experiments.
IEEE TRANSACTIONS ON COMPUTERS
(2023)
Article
Computer Science, Hardware & Architecture
Abdulrahman Alaql, Saranyu Chattopadhyay, Prabuddha Chakraborty, Tamzidul Hoque, Swarup Bhunia
Summary: The article introduces LeGO, a learning-guided obfuscation framework that overcomes known vulnerabilities in a scalable and systematic manner, leading to a robust and lightweight locking mechanism. The proposed framework is guided by a security evaluation process that performs a thorough assessment of an obfuscated IP against various attacks and identifies vulnerabilities. The evaluation of the LeGO framework shows promising results in terms of robustness against diverse attacks with manageable overhead in terms of area, power, and delay.
IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS
(2022)
Article
Computer Science, Information Systems
Nemanja Kajtez, Yue Zhang, Basel Halak
Summary: The rising cost of manufacturing nanoscale integrated circuits has led to outsourcing by IC design companies, resulting in security threats in the multinational hardware supply chain. Logic locking techniques have been proposed to combat these threats, but adoption has been slow due to integration and efficacy issues.
Article
Computer Science, Hardware & Architecture
Maxime Montoya, Simone Bacles-Min, Anca Molnos, Jacques J. A. Fournier
Summary: With the development of the Internet of Things, the need for encryption of sensitive data has increased. However, low-cost hardware attacks pose a threat to the implementation of cryptographic algorithms. Existing countermeasures have limitations in protecting against both fault analysis and side-channel analysis. In this paper, a new joint countermeasure called dynamic encoding is proposed, which has a smaller power and area overhead compared to existing countermeasures. It provides efficient protection against side-channel analysis and inherent fault detection.
MICROPROCESSORS AND MICROSYSTEMS
(2023)
Article
Computer Science, Hardware & Architecture
Abdullah Aljuffri, Marc Zwalua, Cezar Rodolfo Wedig Reinbrecht, Said Hamdioui, Mottaqiallah Taouil
Summary: This article explores side-channel attacks using heat, introduces correlation thermal attack (CTA) and DL-based thermal attack (DL-based TA), as well as a new attack called progressive CTA (PCTA). The study shows that all three attacks are practical and achieve 100% key recovery, demonstrating the effectiveness of thermal side-channel attacks in retrieving keys from electronic devices.
IEEE TRANSACTIONS ON VERY LARGE SCALE INTEGRATION (VLSI) SYSTEMS
(2021)
Article
Computer Science, Information Systems
Rajit Karmakar, Harshit Kumar, Santanu Chattopadhyay
Summary: Logic encryption is a potential solution to protect Intellectual Property (IP) from piracy and counterfeiting, but recent attacks have raised concerns. A new dynamic obfuscation scheme proposed in this paper aims to protect against SAT attacks by safeguarding the Design-for-Testability (DfT) infrastructure and preventing key leakage through weak gate locations, offering protection against various attacks while maintaining testability. Unlike other SAT preventive schemes, the proposed method does not suffer from output corruption, meeting the fundamental requirements of a logic encryption scheme.
IEEE TRANSACTIONS ON EMERGING TOPICS IN COMPUTING
(2021)
Article
Computer Science, Hardware & Architecture
Carlton Shepherd, Benjamin Semal, Konstantinos Markantonakis
Summary: This paper presents new methods and results for recognizing black-box program functions using hardware performance counters (HPC). It shows that HPC can be used to classify hardware events and recognize benchmarking and cryptographic library functions. The paper also demonstrates that known OpenSSL vulnerabilities and standardized cryptographic functions within ARM TrustZone TEE applications can be detected using HPC differences. Mitigations, challenges, and future research directions are discussed.
IEEE TRANSACTIONS ON COMPUTERS
(2023)
Article
Computer Science, Hardware & Architecture
Mohammed Shayan, Sukanta Bhattacharjee, Robert Wille, Krishnendu Chakrabarty, Ramesh Karri
Summary: A digital microfluidic biochip (DMFB) is a miniaturized laboratory capable of implementing biochemical protocols, with a network connection, commonly used in medical diagnosis, drug development, and personalized medicine. Ensuring the security of DMFB CPS is crucial for their successful deployment, and recent research has focused on developing defense mechanisms such as checkpoints to enhance security.
IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS
(2021)
Article
Computer Science, Hardware & Architecture
Mojan Javaheripi, Mohammad Samragh, Bita Darvish Rouhani, Tara Javidi, Farinaz Koushanfar
Summary: CuRTAIL is a novel end-to-end computing framework designed to prevent adversarial attacks and has shown effectiveness against adversarial samples. The framework employs an unsupervised approach to validate the legitimacy of input samples, ensuring effective defense against generic attacks.
IEEE TRANSACTIONS ON DEPENDABLE AND SECURE COMPUTING
(2021)
Article
Computer Science, Information Systems
Siam U. Hussain, M. Sadegh Riazi, Farinaz Koushanfar
Summary: Designing custom secure function evaluation compilers has been a popular research area, yet integrated circuit synthesis tools with intelligent adaptation outperform them. It is time for custom compilers to adopt this trend.
IEEE SECURITY & PRIVACY
(2021)
Article
Computer Science, Information Systems
Malhar Jere, Tyler Farnan, Farinaz Koushanfar
Summary: Federated learning is a privacy-focused framework for training deep neural networks from decentralized data sources, but it is vulnerable to numerous attacks. A more robust threat modeling is needed to enhance security in federated learning environments.
IEEE SECURITY & PRIVACY
(2021)
Article
Engineering, Electrical & Electronic
Priyanka Mahesh, Akash Tiwari, Chenglu Jin, Panganamala R. Kumar, A. L. Narasimha Reddy, Satish T. S. Bukkapatanam, Nikhil Gupta, Ramesh Karri
Summary: The Industry 4.0 concept advocates a digital manufacturing paradigm to enhance production efficiency and quality through the integration of various manufacturing technologies and systems.
PROCEEDINGS OF THE IEEE
(2021)
Article
Computer Science, Hardware & Architecture
Rana Elnaggar, Ramesh Karri, Krishnendu Chakrabarty
Summary: The study introduces two security architectures for protecting IJTAG, one providing protection against data alteration but not against data-sniffing attacks, and the other addressing both data alteration and data-sniffing attacks at the cost of timing overhead.
IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS
(2021)
Article
Engineering, Electrical & Electronic
Samrat Acharya, Yury Dvorkin, Ramesh Karri
Summary: Power utilities and third-party aggregators are using AI to learn energy usage patterns of consumers and design optimal DR incentives, but this approach is vulnerable to data integrity attacks.
IEEE TRANSACTIONS ON SMART GRID
(2021)
Article
Computer Science, Hardware & Architecture
Virinchi Roy Surabhi, Prashanth Krishnamurthy, Hussam Amrouch, Joerg Henkel, Ramesh Karri, Farshad Khorrami
Summary: This study explores a non-destructive method for detecting Trojans in circuits using FinFET technology. The method utilizes short-term aging effects and circuit overclocking to induce bit errors at the circuit outputs, while employing Machine Learning tools to learn Trojan-free behavior. The results demonstrate the effectiveness of the method in FinFET technology, with robust Trojan detection capabilities across different chips.
IEEE TRANSACTIONS ON COMPUTERS
(2022)
Article
Engineering, Electrical & Electronic
Samrat Acharya, Robert Mieth, Charalambos Konstantinou, Ramesh Karri, Yury Dvorkin
Summary: This paper proposes cyber insurance for public Electric Vehicle Charging Stations (EVCSs) to hedge the economic loss caused by cyberattacks and develops a data-driven model to determine the optimal insurance premium. The study emphasizes the importance of risk assessment in designing insurance premiums and suggests deploying state-of-the-art defense mechanisms to reduce insurance costs.
IEEE TRANSACTIONS ON SMART GRID
(2022)
Article
Computer Science, Hardware & Architecture
Mojan Javaheripi, Jung-Woo Chang, Farinaz Koushanfar
Summary: AccHashtag is the first framework for high-accuracy detection of fault-injection attacks on DNNs with provable bounds on detection performance. It extracts a unique signature from the benign DNN prior to deployment to validate the model's integrity and verify the inference output on the fly. AccHashtag identifies the most vulnerable DNN layers to the fault-injection attack using a novel sensitivity analysis.
ACM JOURNAL ON EMERGING TECHNOLOGIES IN COMPUTING SYSTEMS
(2023)
Article
Computer Science, Hardware & Architecture
Huili Chen, Xinqiao Zhang, Ke Huang, Farinaz Koushanfar
Summary: This paper proposes AdaTest, a novel adaptive test pattern generation framework for efficient and reliable Hardware Trojan (HT) detection. AdaTest improves the scalability and accuracy of HT detection techniques, especially for small Trojans in the presence of noise and variations. It leverages Reinforcement Learning (RL) for high trigger coverage and employs adaptive sampling to prioritize test samples that provide more information for HT detection. AdaTest's optimized on-chip architecture minimizes hardware overhead and achieves significant speedup and test set size reduction compared to prior works, while maintaining a high Trojan detection rate.
ACM TRANSACTIONS ON EMBEDDED COMPUTING SYSTEMS
(2023)
Article
Computer Science, Theory & Methods
Florian Frank, Wenjie Xiong, Nikolaos Athanasios Anagnostopoulos, Andre Schaller, Tolga Arul, Farinaz Koushanfar, Stefan Katzenbeisser, Ulrich Ruehrmair, Jakub Szefer
Summary: This paper details a new attack method that exploits the decay properties of off-the-shelf DRAM modules to spy on the temperature in the vicinity of the IoT device. It shows that this attack can be performed remotely and purely digitally by compromising the device's software. The presented attack works even in devices without a dedicated temperature sensor onboard.
IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY
(2023)
Proceedings Paper
Computer Science, Information Systems
Thien Duc Nguyen, Phillip Rieger, Huili Chen, Hossein Yalame, Helen Moellering, Hossein Fereidooni, Samuel Marchal, Markus Miedinen, Azalia Mirhoseini, Shaza Zeitouni, Farinaz Koushanfar, Ahmad-Reza Sadeghi, Thomas Schneider
Summary: Federated Learning (FL) allows collaborative model training without sharing private data, but is vulnerable to backdoor attacks. Existing defenses either focus on specific attacker models or degrade model performance. FLAME introduces a defense framework that estimates noise injection to eliminate backdoors while maintaining benign performance. Evaluation on various datasets demonstrates FLAME's effectiveness with minimal impact on model performance.
PROCEEDINGS OF THE 31ST USENIX SECURITY SYMPOSIUM
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Paarth Neekhara, Shehzeen Hussain, Jinglong Du, Shlomo Dubnov, Farinaz Koushanfar, Julian McAuley
Summary: With adversarial reprogramming, pre-trained image classification networks can be repurposed for Natural Language Processing (NLP) and other sequence classification tasks without modifying the network architecture or parameters.
2022 IEEE WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV 2022)
(2022)
Proceedings Paper
Computer Science, Hardware & Architecture
Hussam Amrouch, Animesh Basak Chowdhury, Wentian Jin, Ramesh Karri, Farshad Khorrami, Prashanth Krishnamurthy, Ilia Polian, Victor M. van Santen, Benjamin Tan, Sheldon X-D Tan
Summary: With technology scaling to atomic levels, IC test and diagnosis of complex SoCs are becoming increasingly challenging. Machine learning plays a crucial role in IC test and reliability, addressing issues such as transistor aging and self-heating effects, and aiding in the development of emerging technologies.
2021 IEEE 39TH VLSI TEST SYMPOSIUM (VTS)
(2021)