Visible to the public Biblio

Found 103 results

Filters: Keyword is Embedded systems  [Clear All Filters]
2014-09-17
Chasaki, D., Wolf, T..  2012.  Attacks and Defenses in the Data Plane of Networks. Dependable and Secure Computing, IEEE Transactions on. 9:798-810.

Security issues in computer networks have focused on attacks on end systems and the control plane. An entirely new class of emerging network attacks aims at the data plane of the network. Data plane forwarding in network routers has traditionally been implemented with custom-logic hardware, but recent router designs increasingly use software-programmable network processors for packet forwarding. These general-purpose processing devices exhibit software vulnerabilities and are susceptible to attacks. We demonstrate-to our knowledge the first-practical attack that exploits a vulnerability in packet processing software to launch a devastating denial-of-service attack from within the network infrastructure. This attack uses only a single attack packet to consume the full link bandwidth of the router's outgoing link. We also present a hardware-based defense mechanism that can detect situations where malicious packets try to change the operation of the network processor. Using a hardware monitor, our NetFPGA-based prototype system checks every instruction executed by the network processor and can detect deviations from correct processing within four clock cycles. A recovery system can restore the network processor to a safe state within six cycles. This high-speed detection and recovery system can ensure that network processors can be protected effectively and efficiently from this new class of attacks.

2015-04-30
Weyrich, M., Schmidt, J.-P., Ebert, C..  2014.  Machine-to-Machine Communication. Software, IEEE. 31:19-23.

Although wireless communication is integral to our daily lives, there are numerous crucial questions related to coverage, energy consumption, reliability, and security when it comes to industrial deployment. The authors provide an overview of wireless machine-to-machine (M2M) technologies in the context of a smart factory.

Muller, K., Sigl, G., Triquet, B., Paulitsch, M..  2014.  On MILS I/O Sharing Targeting Avionic Systems. Dependable Computing Conference (EDCC), 2014 Tenth European. :182-193.

This paper discusses strategies for I/O sharing in Multiple Independent Levels of Security (MILS) systems mostly deployed in the special environment of avionic systems. MILS system designs are promising approaches for handling the increasing complexity of functionally integrated systems, where multiple applications run concurrently on the same hardware platform. Such integrated systems, also known as Integrated Modular Avionics (IMA) in the aviation industry, require communication to remote systems located outside of the hosting hardware platform. One possible solution is to provide each partition, the isolated runtime environment of an application, a direct interface to the communication's hardware controller. Nevertheless, this approach requires a special design of the hardware itself. This paper discusses efficient system architectures for I/O sharing in the environment of high-criticality embedded systems and the exemplary analysis of Free scale's proprietary Data Path Acceleration Architecture (DPAA) with respect to generic hardware requirements. Based on this analysis we also discuss the development of possible architectures matching with the MILS approach. Even though the analysis focuses on avionics it is equally applicable to automotive architectures such as Auto SAR.

Zheng, J.X., Dongfang Li, Potkonjak, M..  2014.  A secure and unclonable embedded system using instruction-level PUF authentication. Field Programmable Logic and Applications (FPL), 2014 24th International Conference on. :1-4.

In this paper we present a secure and unclonable embedded system design that can target either an FPGA or an ASIC technology. The premise of the security is that the executed machine code and the executing environment (the embedded processor) will authenticate each other at a per-instruction basis using Physical Unclonable Functions (PUFs) that are built into the processor. The PUFs ensure that the execution of the binary code may only proceed if the binary is compiled with the correct intrinsic knowledge of the PUFs, and that such intrinsic knowledge is virtually unique to each processor and therefore unclonable. We will explain how to implement and integrate the PUFs into the processor's execution environment such that each instruction is authenticated and de-obfuscated on-demand and how to transform an ordinary binary executable into PUF-aware, obfuscated binaries. We will also present a prototype system on a Xilinx Spartan6-based FPGA board.

2015-05-01
Dua, Akshay, Bulusu, Nirupama, Feng, Wu-Chang, Hu, Wen.  2014.  Combating Software and Sybil Attacks to Data Integrity in Crowd-Sourced Embedded Systems. ACM Trans. Embed. Comput. Syst.. 13:154:1–154:19.

Crowd-sourced mobile embedded systems allow people to contribute sensor data, for critical applications, including transportation, emergency response and eHealth. Data integrity becomes imperative as malicious participants can launch software and Sybil attacks modifying the sensing platform and data. To address these attacks, we develop (1) a Trusted Sensing Peripheral (TSP) enabling collection of high-integrity raw or aggregated data, and participation in applications requiring additional modalities; and (2) a Secure Tasking and Aggregation Protocol (STAP) enabling aggregation of TSP trusted readings by untrusted intermediaries, while efficiently detecting fabricators. Evaluations demonstrate that TSP and STAP are practical and energy-efficient.

2015-05-04
Yuying Wang, Xingshe Zhou.  2014.  Spatio-temporal semantic enhancements for event model of cyber-physical systems. Signal Processing, Communications and Computing (ICSPCC), 2014 IEEE International Conference on. :813-818.

The newly emerging cyber-physical systems (CPS) discover events from multiple, distributed sources with multiple levels of detail and heterogeneous data format, which may not be compare and integrate, and turn to hardly combined determination for action. While existing efforts have mainly focused on investigating a uniform CPS event representation with spatio-temporal attributes, in this paper we propose a new event model with two-layer structure, Basic Event Model (BEM) and Extended Information Set (EIS). A BEM could be extended with EIS by semantic adaptor for spatio-temporal and other attribution enhancement. In particular, we define the event process functions, like event attribution extraction and composition determination, for CPS action trigger exploit the Complex Event Process (CEP) engine Esper. Examples show that such event model provides several advantages in terms of extensibility, flexibility and heterogeneous support, and lay the foundations of event-based system design in CPS.
 

Jantsch, A., Tammemae, K..  2014.  A framework of awareness for artificial subjects. Hardware/Software Codesign and System Synthesis (CODES+ISSS), 2014 International Conference on. :1-3.

A small battery driven bio-patch, attached to the human body and monitoring various vital signals such as temperature, humidity, heart activity, muscle and brain activity, is an example of a highly resource constrained system, that has the demanding task to assess correctly the state of the monitored subject (healthy, normal, weak, ill, improving, worsening, etc.), and its own capabilities (attached to subject, working sensors, sufficient energy supply, etc.). These systems and many other systems would benefit from a sense of itself and its environment to improve robustness and sensibility of its behavior. Although we can get inspiration from fields like neuroscience, robotics, AI, and control theory, the tight resource and energy constraints imply that we have to understand accurately what technique leads to a particular feature of awareness, how it contributes to improved behavior, and how it can be implemented cost-efficiently in hardware or software. We review the concepts of environment- and self-models, semantic interpretation, semantic attribution, history, goals and expectations, prediction, and self-inspection, how they contribute to awareness and self-awareness, and how they contribute to improved robustness and sensibility of behavior.

2015-05-05
Kurian, N.A., Thomas, A., George, B..  2014.  Automated fault diagnosis in Multiple Inductive Loop Detectors. India Conference (INDICON), 2014 Annual IEEE. :1-5.

Multiple Inductive Loop Detectors are advanced Inductive Loop Sensors that can measure traffic flow parameters in even conditions where the traffic is heterogeneous and does not conform to lanes. This sensor consists of many inductive loops in series, with each loop having a parallel capacitor across it. These inductive and capacitive elements of the sensor may undergo open or short circuit faults during operation. Such faults lead to erroneous interpretation of data acquired from the loops. Conventional methods used for fault diagnosis in inductive loop detectors consume time and effort as they require experienced technicians and involve extraction of loops from the saw-cut slots on the road. This also means that the traffic flow parameters cannot be measured until the sensor system becomes functional again. The repair activities would also disturb traffic flow. This paper presents a method for automating fault diagnosis for series-connected Multiple Inductive Loop Detectors, based on an impulse test. The system helps in the diagnosis of open/short faults associated with the inductive and capacitive elements of the sensor structure by displaying the fault status conveniently. Since the fault location as well as the fault type can be precisely identified using this method, the repair actions are also localised. The proposed system thereby results in significant savings in both repair time and repair costs. An embedded system was developed to realize this scheme and the same was tested on a loop prototype.
 

2015-05-06
Tsoutsos, N.G., Maniatakos, M..  2014.  Fabrication Attacks: Zero-Overhead Malicious Modifications Enabling Modern Microprocessor Privilege Escalation. Emerging Topics in Computing, IEEE Transactions on. 2:81-93.

The wide deployment of general purpose and embedded microprocessors has emphasized the need for defenses against cyber-attacks. Due to the globalized supply chain, however, there are several stages where a processor can be maliciously modified. The most promising stage, and the hardest during which to inject the hardware trojan, is the fabrication stage. As modern microprocessor chips are characterized by very dense, billion-transistor designs, such attacks must be very carefully crafted. In this paper, we demonstrate zero overhead malicious modifications on both high-performance and embedded microprocessors. These hardware trojans enable privilege escalation through execution of an instruction stream that excites the necessary conditions to make the modification appear. The minimal footprint, however, comes at the cost of a small window of attack opportunities. Experimental results show that malicious users can gain escalated privileges within a few million clock cycles. In addition, no system crashes were reported during normal operation, rendering the modifications transparent to the end user.
 

Xinhai Zhang, Persson, M., Nyberg, M., Mokhtari, B., Einarson, A., Linder, H., Westman, J., DeJiu Chen, Torngren, M..  2014.  Experience on applying software architecture recovery to automotive embedded systems. Software Maintenance, Reengineering and Reverse Engineering (CSMR-WCRE), 2014 Software Evolution Week - IEEE Conference on. :379-382.

The importance and potential advantages with a comprehensive product architecture description are well described in the literature. However, developing such a description takes additional resources, and it is difficult to maintain consistency with evolving implementations. This paper presents an approach and industrial experience which is based on architecture recovery from source code at truck manufacturer Scania CV AB. The extracted representation of the architecture is presented in several views and verified on CAN signal level. Lessons learned are discussed.
 

2015-11-12
Emfinger, W., Karsai, G..  2015.  Modeling Network Medium Access Protocols for Network Quality of Service Analysis. Real-Time Distributed Computing (ISORC), 2015 IEEE 18th International Symposium on. :292-295.

Design-time analysis and verification of distributed real-time embedded systems necessitates the modeling of the time-varying performance of the network and comparing that to application requirements. Earlier work has shown how to build a system network model that abstracted away the network's physical medium and protocols which govern its access and multiplexing. In this work we show how to apply a network medium channel access protocol, such as Time-Division Multiple Access (TDMA), to our network analysis methods and use the results to show that the abstracted model without the explicit model of the protocol is valid.

2017-02-14
L. Rivière, J. Bringer, T. H. Le, H. Chabanne.  2015.  "A novel simulation approach for fault injection resistance evaluation on smart cards". 2015 IEEE Eighth International Conference on Software Testing, Verification and Validation Workshops (ICSTW). :1-8.

Physical perturbations are performed against embedded systems that can contain valuable data. Such devices and in particular smart cards are targeted because potential attackers hold them. The embedded system security must hold against intentional hardware failures that can result in software errors. In a malicious purpose, an attacker could exploit such errors to find out secret data or disrupt a transaction. Simulation techniques help to point out fault injection vulnerabilities and come at an early stage in the development process. This paper proposes a generic fault injection simulation tool that has the particularity to embed the injection mechanism into the smart card source code. By its embedded nature, the Embedded Fault Simulator (EFS) allows us to perform fault injection simulations and side-channel analyses simultaneously. It makes it possible to achieve combined attacks, multiple fault attacks and to perform backward analyses. We appraise our approach on real, modern and complex smart card systems under data and control flow fault models. We illustrate the EFS capacities by performing a practical combined attack on an Advanced Encryption Standard (AES) implementation.

2017-02-23
T. Long, G. Yao.  2015.  "Verification for Security-Relevant Properties and Hyperproperties". 2015 IEEE 12th Intl Conf on Ubiquitous Intelligence and Computing and 2015 IEEE 12th Intl Conf on Autonomic and Trusted Computing and 2015 IEEE 15th Intl Conf on Scalable Computing and Communications and Its Associated Workshops (UIC-ATC-ScalCom). :490-497.

Privacy analysis is essential in the society. Data privacy preservation for access control, guaranteed service in wireless sensor networks are important parts. In programs' verification, we not only consider about these kinds of safety and liveness properties but some security policies like noninterference, and observational determinism which have been proposed as hyper properties. Fairness is widely applied in verification for concurrent systems, wireless sensor networks and embedded systems. This paper studies verification and analysis for proving security-relevant properties and hyper properties by proposing deductive proof rules under fairness requirements (constraints).

2017-02-27
Cómbita, L. F., Giraldo, J., Cárdenas, A. A., Quijano, N..  2015.  Response and reconfiguration of cyber-physical control systems: A survey. 2015 IEEE 2nd Colombian Conference on Automatic Control (CCAC). :1–6.

The integration of physical systems with distributed embedded computing and communication devices offers advantages on reliability, efficiency, and maintenance. At the same time, these embedded computers are susceptible to cyber-attacks that can harm the performance of the physical system, or even drive the system to an unsafe state; therefore, it is necessary to deploy security mechanisms that are able to automatically detect, isolate, and respond to potential attacks. Detection and isolation mechanisms have been widely studied for different types of attacks; however, automatic response to attacks has attracted considerably less attention. Our goal in this paper is to identify trends and recent results on how to respond and reconfigure a system under attack, and to identify limitations and open problems. We have found two main types of attack protection: i) preventive, which identifies the vulnerabilities in a control system and then increases its resiliency by modifying either control parameters or the redundancy of devices; ii) reactive, which responds as soon as the attack is detected (e.g., modifying the non-compromised controller actions).

2017-03-07
Ugwoke, F. N., Okafor, K. C., Chijindu, V. C..  2015.  Security QoS profiling against cyber terrorism in airport network systems. 2015 International Conference on Cyberspace (CYBER-Abuja). :241–251.

Attacks on airport information network services in the form of Denial of Service (DoS), Distributed DoS (DDoS), and hijacking are the most effective schemes mostly explored by cyber terrorists in the aviation industry running Mission Critical Services (MCSs). This work presents a case for Airport Information Resource Management Systems (AIRMS) which is a cloud based platform proposed for the Nigerian aviation industry. Granting that AIRMS is susceptible to DoS attacks, there is need to develop a robust counter security network model aimed at pre-empting such attacks and subsequently mitigating the vulnerability in such networks. Existing works in literature regarding cyber security DoS and other schemes have not explored embedded Stateful Packet Inspection (SPI) based on OpenFlow Application Centric Infrastructure (OACI) for securing critical network assets. As such, SPI-OACI was proposed to address the challenge of Vulnerability Bandwidth Depletion DDoS Attacks (VBDDA). A characterization of the Cisco 9000 router firewall as an embedded network device with support for Virtual DDoS protection was carried out in the AIRMS threat mitigation design. Afterwards, the mitigation procedure and the initial phase of the design with Riverbed modeler software were realized. For the security Quality of Service (QoS) profiling, the system response metrics (i.e. SPI-OACI delay, throughput and utilization) in cloud based network were analyzed only for normal traffic flows. The work concludes by offering practical suggestion for securing similar enterprise management systems running on cloud infrastructure against cyber terrorists.

2017-03-08
Bertino, E., Hartman, N. W..  2015.  Cybersecurity for product lifecycle management a research roadmap. 2015 IEEE International Conference on Intelligence and Security Informatics (ISI). :114–119.

This paper introduces a research agenda focusing on cybersecurity in the context of product lifecycle management. The paper discusses research directions on critical protection techniques, including protection techniques from insider threat, access control systems, secure supply chains and remote 3D printing, compliance techniques, and secure collaboration techniques. The paper then presents an overview of DBSAFE, a system for protecting data from insider threat.

2017-03-13
Kamoona, M., El-Sharkawy, M..  Submitted.  FlexiWi-Fi Security Manager Using Freescale Embedded System. 2015 2nd International Conference on Information Science and Security (ICISS). :1–4.

Among the current Wi-Fi two security models (Enterprise and Personal), while the Enterprise model (802.1X) offers an effective framework for authenticating and controlling the user traffic to a protected network, the Personal model (802.11) offers the cheapest and the easiest to setup solution. However, the drawback of the personal model implementation is that all access points and client radio NIC on the wireless LAN should use the same encryption key. A major underlying problem of the 802.11 standard is that the pre-shared keys are cumbersome to change. So if those keys are not updated frequently, unauthorized users with some resources and within a short timeframe can crack the key and breach the network security. The purpose of this paper is to propose and implement an effective method for the system administrator to manage the users connected to a router, update the keys and further distribute them for the trusted clients using the Freescale embedded system, Infrared and Bluetooth modules.

2017-04-20
Agarwal, N., Paul, K..  2016.  XEBRA: XEn Based Remote Attestation. 2016 IEEE Region 10 Conference (TENCON). :2383–2386.

Modern computing environments are increasingly getting distributed with one machine executing programs on the other remotely. Often, multiple machines work together to complete a task. Its important for collaborating machines to trust each other in order to perform properly. Such scenarios have brought up a key security issue of trustably and securely executing critical code on remote machines. We present a purely software based remote attestation technique XEBRA(XEn Based Remote Attestation) that guarantees the execution of correct code on a remote host, termed as remote attestation. XEBRA can be used to establish dynamic root of trust in a remote computing device using virtualization. We also show our approach to be feasible on embedded platforms by implementing it on an Intel Galileo board.

Tan, B., Biglari-Abhari, M., Salcic, Z..  2016.  A system-level security approach for heterogeneous MPSoCs. 2016 Conference on Design and Architectures for Signal and Image Processing (DASIP). :74–81.

Embedded systems are becoming increasingly complex as designers integrate different functionalities into a single application for execution on heterogeneous hardware platforms. In this work we propose a system-level security approach in order to provide isolation of tasks without the need to trust a central authority at run-time. We discuss security requirements that can be found in complex embedded systems that use heterogeneous execution platforms, and by regulating memory access we create mechanisms that allow safe use of shared IP with direct memory access, as well as shared libraries. We also present a prototype Isolation Unit that checks memory transactions and allows for dynamic configuration of permissions.

2017-05-16
Oswald, David F..  2016.  Wireless Attacks on Automotive Remote Keyless Entry Systems. Proceedings of the 6th International Workshop on Trustworthy Embedded Devices. :43–44.

Modern vehicles rely on a variety of electronic systems and components. One of those components is the vehicle key. Today, a key typically implements at least three functions: mechanical locking with a key blade, the electronic immobilizer to autorise the start of the engine, and the remote keyless entry (RKE) system that allows to wirelessly (un)lock the doors and disable the alarm system. These main components of a vehicle key are shown in Figure 1. For the mechanical part of the vehicle key, it is well known that the key blade can be easily copied and that the locking cylinder can be bypassed with other means (using so-called "decoders" or simply a screwdriver). In contrast, immobilizer and RKE rely on wireless protocols to cryptographically authenticate the vehicle key to the car. Immobilizers employ radio frequency identification (RFID) transponders to carry out a challenge-response protocol over a low-range bidirectional link at a frequency of 125 kHz. In the past, researchers have revealed severe aws in the cryptography and protocols used by immobilizers, leading to the break of the major systems Megamos, Hitag2, and DST40 [7, 6, 1]. In contrast to the immobilizer, the RKE part uses unidirectional communication (the vehicle only receives, the key fob only transmits) over a high-range wireless link with operating distances of tens to one hundred meters. These systems are based on rolling codes, which essentially transmit a counter (that is incremented on each button press) in a cryptographically authenticated manner. Until recently, the security of automotive RKE had been scrutinized to a lesser degree than that of immobilizers, even though vulnerabilities in similar systems have been known since 2008 with the attacks on KeeLoq [3]. Other results reported in the literature include an analytical attack on a single, outdated vehicle [2] and the so-called "RollJam" technique [5], which is based on a combination of replay and selective jamming. In 2016, it was shown that severe aws exist in the RKE systems of major automotive manufacturers [4]. On the one hand, the VWgroup (Volkswagen, Seat, Skoda, Audi) based the security of their RKE system on a few global cryptographic keys, potentially affecting hundreds of million vehicles world-wide. By extracting these global keys from the firmware of electronic controls units (ECUs) once, an adversary is able to create a duplicate of the owner's RKE fob by eavesdropping a single rolling code. The second case study in [4] exposes new cryptographic weaknesses in the Hitag2 cipher when used for RKE. Applying a correlation-based attack, an adversary can recover the 48-bit cryptographic key by eavesdropping four to eight rolling codes and performing a one-minute computation on a standard laptop. Again, this attack affects millions of vehicle world-wide. Manufacturers that used Hitag2 in their RKE system include Alfa Romeo, Peugeot, Lancia, Opel, Renault, and Ford among others. In this keynote talk, we will present the results of [4] and put them in into a broader context by revisiting the history of attacks on RKE systems and automotive electronics.

2017-09-05
Barbot, Benoît, Kwiatkowska, Marta, Mereacre, Alexandru, Paoletti, Nicola.  2016.  Building Power Consumption Models from Executable Timed I/O Automata Specifications. Proceedings of the 19th International Conference on Hybrid Systems: Computation and Control. :195–204.

We develop a novel model-based hardware-in-the-loop (HIL) framework for optimising energy consumption of embedded software controllers. Controller and plant models are specified as networks of parameterised timed input/output automata and translated into executable code. The controller is encoded into the target embedded hardware, which is connected to a power monitor and interacts with the simulation of the plant model. The framework then generates a power consumption model that maps controller transitions to distributions over power measurements, and is used to optimise the timing parameters of the controller, without compromising a given safety requirement. The novelty of our approach is that we measure the real power consumption of the controller and use thus obtained data for energy optimisation. We employ timed Petri nets as an intermediate representation of the executable specification, which facilitates efficient code generation and fast simulations. Our framework uniquely combines the advantages of rigorous specifications with accurate power measurements and methods for online model estimation, thus enabling automated design of correct and energy-efficient controllers.

Ghanei, Farshad, Tipnis, Pranav, Marcus, Kyle, Dantu, Karthik, Ko, Steve, Ziarek, Lukasz.  2016.  OS-based Resource Accounting for Asynchronous Resource Use in Mobile Systems. Proceedings of the 2016 International Symposium on Low Power Electronics and Design. :296–301.

One essential functionality of a modern operating system is to accurately account for the resource usage of the underlying hardware. This is especially important for computing systems that operate on battery power, since energy management requires accurately attributing resource uses to processes. However, components such as sensors, actuators and specialized network interfaces are often used in an asynchronous fashion, and makes it difficult to conduct accurate resource accounting. For example, a process that makes a request to a sensor may not be running on the processor for the full duration of the resource usage; and current mechanisms of resource accounting fail to provide accurate accounting for such asynchronous uses. This paper proposes a new mechanism to accurately account for the asynchronous usage of resources in mobile systems. Our insight is that by accurately relating the user requests with kernel requests to device and corresponding device responses, we can accurately attribute resource use to the requesting process. Our prototype implemented in Linux demonstrates that we can account for the usage of asynchronous resources such as GPS and WiFi accurately.

2017-11-20
Costin, Andrei.  2016.  Security of CCTV and Video Surveillance Systems: Threats, Vulnerabilities, Attacks, and Mitigations. Proceedings of the 6th International Workshop on Trustworthy Embedded Devices. :45–54.

Video surveillance, closed-circuit TV and IP-camera systems became virtually omnipresent and indispensable for many organizations, businesses, and users. Their main purpose is to provide physical security, increase safety, and prevent crime. They also became increasingly complex, comprising many communication means, embedded hardware and non-trivial firmware. However, most research to date focused mainly on the privacy aspects of such systems, and did not fully address their issues related to cyber-security in general, and visual layer (i.e., imagery semantics) attacks in particular. In this paper, we conduct a systematic review of existing and novel threats in video surveillance, closed-circuit TV and IP-camera systems based on publicly available data. The insights can then be used to better understand and identify the security and the privacy risks associated with the development, deployment and use of these systems. We study existing and novel threats, along with their existing or possible countermeasures, and summarize this knowledge into a comprehensive table that can be used in a practical way as a security checklist when assessing cyber-security level of existing or new CCTV designs and deployments. We also provide a set of recommendations and mitigations that can help improve the security and privacy levels provided by the hardware, the firmware, the network communications and the operation of video surveillance systems. We hope the findings in this paper will provide a valuable knowledge of the threat landscape that such systems are exposed to, as well as promote further research and widen the scope of this field beyond its current boundaries.

2017-12-04
Johnston, B., Lee, B., Angove, L., Rendell, A..  2017.  Embedded Accelerators for Scientific High-Performance Computing: An Energy Study of OpenCL Gaussian Elimination Workloads. 2017 46th International Conference on Parallel Processing Workshops (ICPPW). :59–68.

Energy efficient High-Performance Computing (HPC) is becoming increasingly important. Recent ventures into this space have introduced an unlikely candidate to achieve exascale scientific computing hardware with a small energy footprint. ARM processors and embedded GPU accelerators originally developed for energy efficiency in mobile devices, where battery life is critical, are being repurposed and deployed in the next generation of supercomputers. Unfortunately, the performance of executing scientific workloads on many of these devices is largely unknown, yet the bulk of computation required in high-performance supercomputers is scientific. We present an analysis of one such scientific code, in the form of Gaussian Elimination, and evaluate both execution time and energy used on a range of embedded accelerator SoCs. These include three ARM CPUs and two mobile GPUs. Understanding how these low power devices perform on scientific workloads will be critical in the selection of appropriate hardware for these supercomputers, for how can we estimate the performance of tens of thousands of these chips if the performance of one is largely unknown?

2017-12-12
Abdi, Fardin, Tabish, Rohan, Rungger, Matthias, Zamani, Majid, Caccamo, Marco.  2017.  Application and System-level Software Fault Tolerance Through Full System Restarts. Proceedings of the 8th International Conference on Cyber-Physical Systems. :197–206.

Due to the growing performance requirements, embedded systems are increasingly more complex. Meanwhile, they are also expected to be reliable. Guaranteeing reliability on complex systems is very challenging. Consequently, there is a substantial need for designs that enable the use of unverified components such as real-time operating system (RTOS) without requiring their correctness to guarantee safety. In this work, we propose a novel approach to design a controller that enables the system to restart and remain safe during and after the restart. Complementing this controller with a switching logic allows the system to use complex, unverified controller to drive the system as long as it does not jeopardize safety. Such a design also tolerates faults that occur in the underlying software layers such as RTOS and middleware and recovers from them through system-level restarts that reinitialize the software (middleware, RTOS, and applications) from a read-only storage. Our approach is implementable using one commercial off-the-shelf (COTS) processing unit. To demonstrate the efficacy of our solution, we fully implement a controller for a 3 degree of freedom (3DOF) helicopter. We test the system by injecting various types of faults into the applications and RTOS and verify that the system remains safe.