Organizer
Andy D. Pimentel, University of Amsterdam
Biography: Andy Pimentel is full professor at the University of Amsterdam where he chairs the Parallel Computing Systems group. His research centers around system-level modeling, simulation, and exploration of (embedded) multicore and manycore computer systems with the purpose of efficiently and effectively designing and programming these systems. He has an MSc and PhD in computer science, both from the University of Amsterdam. He is a co-founder of the International Conference on Embedded Computer Systems: Architectures, Modeling, and Simulation (SAMOS). He has (co)authored more than 110 scientific publications and is an Associate Editor of Elsevier’s Simulation Modelling Practice and Theory as well as Springers Journal of Signal Processing Systems. He served as the General Chair of HIPEAC’15, as Local Organization Co-Chair of ESWeek’15, and as Program (Vice-)Chair of CODES+ISSS in 2016 and 2017. Furthermore, he has served on the TPC of many leading (embedded) computer systems design conferences, such as DAC, DATE, CODES+ISSS, ICCD, ICCAD, FPL, SAMOS, and ESTIMedia.
Abstract
Due to the increasing performance demands of mission- and safety-critical Cyber Physical Systems (CPS), these systems exhibit a rapidly growing complexity, manifested by an increasing number of (distributed) computational cores and application components connected via complex networks. However, with the growing complexity and interconnectivity of these systems, the chances of hardware failures as well as disruptions due to cyber-attacks will also quickly increase. System adaptivity, for example in the form of dynamically remapping of application components to processing cores, represents a promising technique to handle this challenging scenario. In this session, we address the (consequences of the) idea of deploying runtime adaptivity to mission and safety-critical CPS, yielding dynamically morphing systems, to establish robustness against computational hurdles, component failures, and cyber-attacks.
Presentations
The TeamPlay Coordination Language for Mission-critical Cyber-physical Systems
Clemens Grelck, University of Amsterdam
Abstract: Software engineering for mission-critical cyber-physical systems is a challenging task as the need for functional correctness of code is complemented by the additional requirements of meeting real-time deadlines, controlling energy-consumption and being prepared for temporarily or permanently failing hardware units. Intertwining all these requirements within a code base inevitably leads to a software mess that is hard to manage and maintain but also limits trust for its functional and non-functional correctness.
Exogeneous coordination is a concept that aims at strict separation of concerns between the implementation of software components with both functional and non-functional properties on the one hand and their orderly interaction as a fully- edged application with compound functional and non-functional properties on the other hand. We present the TeamPlay coordination language for the high-level specification of energy- and time-aware cyber-physical systems that are robust against hardware failure. The TeamPlay approach to resilience trades computing resources of modern heterogeneous parallel hardware for fault-tolerance in a declarative style. We present the underlying coordination model and domain-specific language (DSL) and sketch out the corresponding compiler and runtime system technology.
Speaker’s biography: Clemens Grelck is an Associate Professor for Compiler Design in the Parallel Computing Systems group at the University of Amsterdam, Netherlands. Prior to joining the University of Amsterdam he held academic positions at the Universities of Kiel and Lübeck, Germany, and the University of Hertfordshire, United Kingdom. His research interests revolve around high-level programming models for parallel and heterogeneous computing and the corresponding compilation and runtime system technologies. Besides the productivity/ performance trade-off Clemens works on resource-aware programming abstractions that treat time, energy, security and resilience in cyber-physical systems as first-class citizens. He has (co)authored more than 110 scientific publications and has been involved in a variety of EU collaborative projects as well as Erasmus+ strategic partnerships and COST Actions.
Scheduling Adaptive Dataflow Applications in Safety-critical Cyber-physical Systems
Florian Haas, University of Augsburg
Abstract: The representation of an application as a directed acyclic graph (DAG), consisting of immutable data partitions and actor nodes in between, simplifies the schedulability analysis, enhances the robustness, and facilitates the adaptivity of mission- and safety-critical cyber-physical systems. The actor nodes represent the executable code of the tasks of the system, while the data dependencies are modelled as the edges between, explicitly describing the data ow of the system. We present an adaptivity-aware runtime environment for such data ow applications that allows to dynamically adjust the mapping of tasks to cores, as well as the individual level of redundancy of each task to adapt to temporal or permanent failures of individual processing nodes. However, the real-time requirements of the system still have to be satisfied, even under the presence of errors or attacks. The graph scheduling algorithm regards a specified fault model and thus ensures a correct timing behaviour, and allows for graceful degradation in case of excessive errors. Given the deadlines, the fault model, and the demanded minimum quality of service, the hardware architecture to satisfy all the requirements of the overall system can be found by means of a design space exploration.
Speaker’s biography: Florian Haas is a postdoctoral researcher at the Chair for Embedded Systems at the University of Augsburg. His research interests are parallel embedded systems under real-time constraints with particular demands on performance and fault tolerance. He received his PhD degree in computer science from the University of Augsburg on fault tolerance of parallel applications on multi-core processors.
Designing Adaptive Avionics Embedded Systems
Stefanos Skalistis, Collins Aerospace
Abstract: Traditional avionics embedded systems are loosely coupled, requiring minimal interaction and data exchange with each other. As a result, most avionics system can be implemented with single-core embedded devices, while, at the same time, faults can be easily contained at device-level and addressed with redundancy techniques, such as triple-modular redundancy. Yet, the recent advances in machine learning techniques, on the one hand have enabled the development of even more autonomous systems, but on the other have increased data processing demands and are far more interconnected. Consequently, existing techniques fall short on meeting the reliability standards of the avionics domain for avionics multi-core systems. In this talk, contemporary challenges of multi-core avionics systems will be discussed and how adaptivity can raise up to those challenges, work that is carried out in H2020 AdMorph project. A particular focus will be given on how adaptivity can be modelled and analysed, such that it could be put against the strict certification requirements of the avionics domain.
Speaker’s biography: Stefanos Skalistis is a research scientist at the Applied Research & Technology of Collins Aerospace focusing on real-time systems, fault-tolerance, timing analysis and correct-by-construction techniques. He acquired his PhD at EPFL (Switzerland) on efficient and adaptive real-time systems, with a particular focus on analysing and managing timing interference (contention) on multi-core architectures. He has also worked as a researcher in multiple research institutions – latest INRIA (France) – and has published his works, as well as served as a reviewer, in several leading real-time and embedded systems conferences, such as RTSS, DATE, ECRTS and RTAS.