Security is one of the most important features that a system must provide. Depending on the application of the target device, different threats should be considered at design time. However, the attack space is vast. Hence, it is difficult to decide what components to protect, what level of protection they require and how efficient they are in the field. This paper tries to close this validation gap for power based side channel attacks by providing a fast and reliable leakage assessment at design time that can be used to perform design space exploration for security. To accomplish our goal, we use Generative Adversarial Networks (GAN) to generate reliable power traces for hardware implementations at design time that are subsequently used to assess the leakage of the design. As a case study, we validated our framework against three AES implementations (i.e., unprotected, masked-protected, and balanced protected). In comparison to CAD-based scenarios, our findings show that the GAN model creates extremely reliable power traces in terms of attackability and leakage assessment. In addition, it is approximately 120 times quicker than CAD tools with respect to trace generation.

Hardware security is currently a very influential domain, where each year countless works are published concerning attacks against hardware and countermeasures. A significant number of them use machine learning, which is proven to be very effective in other domains. This survey, as one of the early attempts, presents the usage of machine learning in hardware security in a full and organized manner. Our contributions include classification and introduction to the relevant fields of machine learning, a comprehensive and critical overview of machine learning usage in hardware security, and an investigation of the hardware attacks against machine learning (neural network) implementations.

The massive deployment of Internet of Things (IoT) devices makes them vulnerable against physical tampering attacks, such as fault injection. These kind of hardware attacks are very popular as they typically do not require complex equipment or high expertise. Hence, it is important that IoT devices are protected against them. In this work, we present a novel fault injection attack detector with high flexibility and low overhead. Our solution is based on the reuse of a security primitive used in many IoT devices, i.e., ring oscillator (RO) physically unclonable function (PUF). Our results show that we obtain a high detection effectiveness and no false alarms against most popular fault injection attacks based on voltage and clock manipulations.

The National Institute of Standard and Technology (NIST) has recently started a competition with the objective to standardize lightweight cryptography (LWC). The winning schemes will be deployed in Internet-of-Things (IoT) devices, a key step for the current and future information and communication technology market. GIFT is an efficient lightweight cipher and it is used by one-fourth of the LWC candidates in the NIST LWC competition. Thus, its security evaluation is critical. One vital threat to the security are so-called logical side-channel attacks based on cache observations. In this work, we propose a novel cache attack on GIFT referred to as GRINCH. We analyzed the vulnerabilities of GIFT and exploited them in our attack. The results show that the attack is effective and that the full key could be recovered with less than 400 encryptions.

Spiking Neural Networks (SNNs) are a strong candidate to be used in future machine learning applications. SNNs can obtain the same accuracy of complex deep learning networks, while only using a fraction of its power. As a result, an increase in popularity of SNNs is expected in the near future for cyber physical systems, especially in the Internet of Things (IoT) segment. However, SNNs work very different than conventional neural network architectures. Consequently, applying SNNs in the field might introduce new unexpected security vulnerabilities. This paper explores and identifies potential sources of information leakage for the Izhikevich neuron, which is a popular neuron model used in digital implementations of SNNs. Simulations and experiments on FPGA implementation of the spiking neurons show that timing and power can be used to infer important information of the internal functionality of the network. Additionally, the paper demonstrates that is feasible to perform a reverse engineering attack using both power and timing leakage.

Internet of things (IoT) devices are appearing in all aspects of our digital life. As such, they have become prime targets for attackers and hackers. An adequate protection against attacks is only possible when the confidentiality and integrity of the data and applications of these devices are secured. State-of-the-art solutions mostly address software and network attacks, but overlook physical/hardware attacks. Such attacks can still exploit software vulnerabilities or even introduce them. In this paper, we present embedded memory security (EMS); it protects against physical tampering of the memory of IoT devices. As a case study, we have equipped a RISC-V based system-on-chip (SoC) with an EMS module. Our experimental results show that EMS successfully can protect the SoC against hardware tampering attacks, while having a low performance overhead.

Attificial neural networks are currently used for many tasks, including safety critical ones such as automated driving. Hence, it is very important to protect them against faults and fault attacks. In this work, we propose two fault injection attack detection mechanisms: one based on using output labels for a reference input, and the other on the activations of neurons. First, we calibrate our detectors during normal conditions. Thereafter, we verify them to maximize fault detection performance. To prove the effectiveness of our solution, we consider highly employed neural networks (AlexNet, GoogleNet, and VGG) with their associated dataset ImageNet. Our results show that for both detectors we are able to obtain a high rate of coverage against faults, typically above 96%. Moreover, the hardware and software implementations of our detector indicate an extremely low area and time overhead.

Asymmetric algorithms such as RSA are considered secure from an algorithmic point of view, yet their implementations are typically vulnerable as they are used by attackers to comprise the secret key. Many countermeasures have been proposed to thwart these attacks. However, they are typically broken as the key can be easily compromised when attackers succeed figuring out which part of the traces belong to the square and multiply operations. In this paper, a new countermeasure is proposed against side channel attacks, referred to as multi-bit blinding. The proposed method provides a constant execution behavior regardless of the key value without additional cost (i.e., dummy/extra operations). It realizes this by considering multiple bits of the key (i.e., two in this paper) simultaneously and always perform the same operations on them independent of the two-bit value. This makes attacks much harder as the attacker cannot retrieve the key simply by identifying the operations. Instead, the attackers need to guess the correct values of the operations as well. As a case study, the security of an RSA algorithm implementation based on the proposed method is evaluated. Our experimental results show that the new method is secure against profiled and non-profiled side channel attacks with less overhead than currently published countermeasures.

Side-channel attacks (SCAs) are powerful attacks that could be used to retrieve keys from electronic devices. Several physical leakage sources can be exploited in SCAs, such as power, time, heat, and so on. Heat is one of the side-channels that is not frequently analyzed by attackers in the literature due to the high noise associated with thermal traces. This article investigates the practicality of adapting power-based SCAs [i.e., correlation power analysis (CPA) and deep-learning-based power attacks (DL-based PA)] for thermal attacks and refer to them as correlation thermal attack (CTA) and DL-based thermal attack (DL-based TA). In addition, we introduce a new attack called progressive CTA (PCTA). We evaluate the different thermal SCAs against an unprotected and protected software implementation of Rivest–Shamir–Adleman (RSA). Our results show the practicality of the three attacks (i.e. CTA, DL-based TA, and PCTA) as a 100% key recovery is realized.

In the recent years, cache based side-channel attacks have become a serious threat for computers. To face this issue, researches have been looking at verifying the security policies. However, these approaches are limited to manual security verification and they typically work for a small subset of the attacks. Hence, an effective verification environment to automatically verify the cache security for all side-channel attacks is still missing. To address this shortcoming, we propose a security verification methodology that formally verifies cache designs against cache side-channel vulnerabilities. Results show that this verification template is a straightforward, automated method in verifying cache invulnerability.

Physically Unclonable Functions (PUFs) are security primitives that provide trustworthy hardware for key-generation and device authentication. Among them, in contrast to dedicated PUFs, intrinsic PUFs are created from existing hardware components that exploit their variability through software. In this work we focus on GPUs and present G-PUF, a PUF implemented entirely in software on CUDA and hence does not require hardware modifications. Our results show that G-PUF has comparable characteristics to SRAM and DRAM PUFs in terms of uniformity 55.61% and reliability 90.09%.

Cache attacks are one of the most wide-spread and dangerous threats to embedded computing systems' security. A promising approach to detect such attacks at runtime is to monitor the System-on-Chip (SoC) behavior. However, designing a secure SoC capable of detecting such attacks is very challenging: the monitors should be lightweight in order to avoid excessive power/energy and area costs and the attack behavior should be clearly known upfront. In this work, we present LiD-CAT, a lightweight and flexible hardware detector that is aware of leakage patterns that can be used by attackers to perform cache based attacks. LiD-CAT is a cache wrapper that implements a set of leakage properties derived from cache attacks and cache models using templates. These templates identify suspicious behavior that may lead to cache attacks. LiD-CAT is evaluated using two different cache architectures, one with a secure cache and one without. On each of them, SPEC2000 benchmarks are run together with malicious applications that execute cache attacks (i.e., Evict+Time, Prime+Probe, Flush+Reload and Flush+Flush). Results show that our lightweight detector successfully detects 99.99% of the attacks with less than 1% false-positives, has no timing penalties, and increases the area of a SoC with only 1.6%.

Physical fault injection attacks are becoming an important threat to computer systems, as fault injection equipment becomes more and more accessible. In this work, we propose a new strategy to detect fault attacks in cryptosystems. We use a recurrent neural network (RNN) to detect problems in the program flow caused by injected faults. Our neural network is trained using the instructions of non-faulty operations and therefore, it can protect against both current and future attacks. As a case study, we use two implementations of software RSA. To test the effectiveness of our detector, we propose a collection of fault injection models, where each model represents different types of faults in the instructions. Evaluation results show that we obtain a high detection accuracy in case injected faults lead to changes in the instruction flow and hence, making it difficult to steal secrete keys. Finally, we propose an efficient hardware implementation with only a 6% area overhead compared to a RISC-V processor.

Increasingly complex and powerful Systems-on-Chips (SoCs), connected through a 5G network, form the basis of the Internet-of-Things (IoT). These technologies will drive the digitalization in all domains, e.g. industry automation, automotive, avionics, and healthcare. A major requirement for all above domains is the long-term (10 to 30 years) secure communication between the SoCs and the cloud over public 5G networks. The foreseeable breakthrough of quantum computers represents a risk for all communication. In order to prepare for such an event, SoCs must integrate secure quantum-computer-resistant cryptography which is reliable and protected against SW and HW attacks. Empowering SoCs with such strong security poses a challenging problem due to limited resources, tight performance requirements and long-term life-cycles. While current works are focused on efficient implementations of post-quantum cryptography, implementation-security and reliability aspects for SoCs are still largely unexplored. To this end, we present three contributions. First, we present a RISC-V co-processor for post-quantum security, able to support lattice-based cryptography. Second, we use HW/SW co-design techniques to accelerate the NTT transformation and hash generation. Third, we perform the fault analysis of the implementation. We show that our coprocessor achieves high reliability and security capabilities while preserving good performance.

Multi-Processor System-on-Chip is one of the todays standard platforms which has being used in several applications, including time critical. In order to meet safety, thus attending real-time constraints, security may be put aside during the design stage. This is the case of the Priority-Preemptive NoCs, a widely used real-time interconnection structure. Their explicit behavior while dealing with communication flows constrained by tight deadlines creates security flaws. To this end, this work presents three contributions. First, we demonstrate for the first time an attack that exploits preemptive NoC-based MPSoCs. Second, we integrate security countermeasures that avoid these attacks while meeting hard deadlines. Third, we evaluate the impact of the attacks and the protected system. Results show that preemptive NoCs must be protected and that it is possible to effectively and efficiently mitigate the vulnerabilities while keeping the deterministic behavior required for time-critical applications.

Multi-Processor Systems-on-Chips (MPSoCs) are a platform for a wide variety of applications and use-cases. The high on-chip connectivity, the programming flexibility, and the reuse of IPs, however, also introduce security concerns. Problems arise when applications with different trust and protection levels share resources of the MPSoC, such as processing units, cache memories and the Network-on-Chip (NoC) communication structure. If a program gets compromised, an adversary can observe the use of these resources and infer (potentially secret) information from other applications. In this work, we explore the cache-based attack by Bogdanov et al., which infers the cache activity of a target program through timing measurements and exploits collisions that occur when the same cache location is accessed for different program inputs. We implement this differential cache-collision attack on the MPSoC Glass and introduce an optimized variant of it, the Earthquake Attack, which leverages the NoC-based communication to increase attack efficiency. Our results show that Earthquake performs well under different cache line and MPSoC configurations, illustrating that cache-collision attacks are considerable threats on MPSoCs.

Neural networks and deep learning are promising techniques for bringing brain inspired computing into embedded platforms. They pave the way to new kinds of associative memories, classifiers, data-mining, machine learning or search engines, which can be the basis of critical and sensitive applications such as autonomous driving. Emerging non-volatile memory technologies integrated in the so called Multi-Processor System-on-Chip (MPSoC) architectures enable the realization of such computational paradigms. These architectures take advantage of the Network-on-Chip concept to efficiently carry out communications with dedicated distributed memories and processing elements. However, current MPSoC-based neuromorphic architectures are deployed without taking security into account. The growing complexity and the hyper-sharing of hardware resources of MPSoCs may become a threat, thus increasing the risk of malware infections and Trojans introduced at design time. Specially, MPSoC microarchitectural side-channels and fault injection attacks can be exploited to leak sensitive information and to cause malfunctions. In this work we present three contributions to that issue: i) first analysis of security issues in MPSoC-based neuromorphic architectures; ii) discussion of the threat model of the neuromorphic architectures; ii) demonstration of the correlation between SNN input and the neural computation.

The exigency for heterogeneous many-core systems has brought an exponential growth in the complexity of their interconnections. In this manner, other Network-on-Chip (NoC) alternatives are being sought to attend the requirements in terms of power consumption and performance. Nevertheless, several of these proposals present very complex architectures, with virtual channels, tables and extra controls. In this paper we propose the junction of two advantageous strategies: hierarchical topology with adaptability. The use of these two techniques is novel in the literature and it allows ensuring high performance even when the application has their communication rates altered. The gains in power and in performance are possible due to the use of low cost components in a hierarchical structure.

Multi-Processors Systems-on-Chip (MPSoCs) are demanding for high performance, low power and high density, and therefore, three-dimensional integrated circuits (3DIC) emerge as a solution to integrate these systems. In order to appropriately interconnect the layers of these systems in terms of flexibility and scalability, a Network-on-Chip (NoC) is typically employed. In this paper, we argue about the scenario of 3D designs, covering all important issues about this new concept. In agreement with all features discussed in this paper, we have proposed a hierarchical 3D topology that meets well the reality of these designs. Experimental results analyze different topologies and show the large benefits in area and power of our proposal.

Networks-on-chip has been seen as an interconnect solution for complex systems. However, performance and energy issues still represent limiting factors for Multi-Processors System-on-Chip (MPSoC). Complex router architectures can be prohibitive for the embedded domain, once they dissipate too much power and energy. In this paper we propose a low power hierarchical network topology with GALS interfaces, allowing each cluster operates in a specific frequency. The clusters are composed by crossbar devices and the number of cores allocated for each cluster is defined considering floorplan information. Experimental results show that our strategy can reduce the power dissipation in up to 58% and the latency in up to 56% for the benchmarks analyzed when compared with a packet-switched mesh network-on-chip.

This research work presents and compares two NoC power estimation models, one based on the volume of information transmitted in the network, and another based on the transmission rates of each router.