Abstract. This lecture introduces the fundamental principles of real-time systems, focusing on task scheduling, timing guarantees, and resource management. After defining the main real-time task and system models, we will examine classical scheduling policies such as Fixed Priority Scheduling (FPS) and Earliest Deadline First (EDF), discussing their optimality properties, schedulability tests, processor utilization bounds, and response-time analysis techniques. The lecture will also address practical issues including shared-resource access, priority inversion, and synchronization protocols.

Bio. Enrico Bini is Associate Professor at Department of Computer Science, University of Turin. He also held positions at Scuola Superiore Sant'Anna (assistant professor, until 2016) and Lund University (Marie-Curie fellow, 2012-14). In 2004, he completed the PhD on Real-Time Systems at Scuola Superiore Sant'Anna (recipient of the "Spitali Award" for best PhD thesis of the whole university). In January 2010, he also completed a Master degree in Mathematics with a thesis on optimal sampling for linear control systems. He has published more than 100 papers (1 Test-of-Time award by the IEEE TCRTS, 4 best-paper and 1 outstanding-pape awards @RTNS @RTCSA @ICC) on real-time scheduling, operating systems, optimization methods for real-time and control systems, optimal management of distributed and parallel resources. Until 2026, he served the community as Associate Editor of IEEE Transactions on Computers, ACM Computing Surveys, and Springer's Real-Time Systems journal. Since 2026, he is Editor-in-Chief of Springer's Real-Time Systems journal. His service also includes chairing 12 events (including Program Chair of RTSS21, Program co-Chair of RTCSA2020, chair of the Embedded Software track of DAC2020) and the participation in about 100 Technical Program Committees.

Abstract. This lecture examines when tasks are scheduled on multicore platforms, where migrations, processor locality and resource contention must be taken into account. Two major approaches to multiprocessor scheduling (Partitioned Scheduling and Global Scheduling) will be illustrated. Also, we introduce the notions of fairness, lag, and progress to quantify the deviation between an ideal processor-sharing schedule and an actual implementation. Finally, we focus on Hierarchical Scheduling, to model the partial availability of processing capacity such as in virtualized environemnts. Resource models and server-based approaches will be introduced to characterize the temporal interface between applications and the computing infrastructure. The lecture will show how hierarchical scheduling enables compositional timing analysis and supports the deployment of real-time applications on virtualized, cloud, and shared computing environments where processor availability is limited or varies over time.

Bio. Enrico Bini is Associate Professor at Department of Computer Science, University of Turin. He also held positions at Scuola Superiore Sant'Anna (assistant professor, until 2016) and Lund University (Marie-Curie fellow, 2012-14). In 2004, he completed the PhD on Real-Time Systems at Scuola Superiore Sant'Anna (recipient of the "Spitali Award" for best PhD thesis of the whole university). In January 2010, he also completed a Master degree in Mathematics with a thesis on optimal sampling for linear control systems. He has published more than 100 papers (1 Test-of-Time award by the IEEE TCRTS, 4 best-paper and 1 outstanding-pape awards @RTNS @RTCSA @ICC) on real-time scheduling, operating systems, optimization methods for real-time and control systems, optimal management of distributed and parallel resources. Until 2026, he served the community as Associate Editor of IEEE Transactions on Computers, ACM Computing Surveys, and Springer's Real-Time Systems journal. Since 2026, he is Editor-in-Chief of Springer's Real-Time Systems journal. His service also includes chairing 12 events (including Program Chair of RTSS21, Program co-Chair of RTCSA2020, chair of the Embedded Software track of DAC2020) and the participation in about 100 Technical Program Committees.

Abstract. This lab explores how Linux can be adapted for real-time applications through PREEMPT_RT, real-time schedulers, and CPU isolation. It first presents the main sources of non-determinism in standard Linux kernels and explains how PREEMPT_RT improves predictability by making the kernel more preemptible, replacing spinlocks with priority-inheritance RT-mutexes, and moving interrupt handlers into schedulable threads. These concepts are then applied in a hands-on session by measuring latency with cyclictest, experimenting with SCHED_FIFO, SCHED_RR, and SCHED_DEADLINE, stressing the system with stress-ng, and configuring isolated CPUs.

Bio. Raffaele Della Corte is an Assistant Professor (tenure-track) within the Department of Electrical Engineering and Information Technologies at the University of Naples Federico II. He received the Ph.D. in Computer Engineering from the same institution in 2016. His research interests encompass data-driven failure analysis, online monitoring of software systems, and real-time computing. He is co-author of several peer-reviewed papers on these topics. He is involved in the Program Committee of several top-ranking conferences, including IEEE ISSRE and IEEE ICWS, and he served as an invited reviewer in many high-ranked journals, such as IEEE TSC and IEEE TSE. He has been principal investigator/research member in several research/industrial funded projects focused on the analysis and monitoring of critical systems.

Abstract. Virtualization is at the heart of cloud computing, as it allows to have multiple virtual machines (VMs) sharing the same hardware resources, improving hardware consolidation and utilization efficiency. Containers, as a form of OS-level lightweight virtualization, are widespread used too. But what are the implications for real-time services, on their timeliness and freedom from interferences from multiple VMs and/or containers sharing the same resources? In this lecture we will present the principles of virtualization and hypervisors and the main solutions adopted to deal with real-time constraints, including real-time schedulers in hypervisors, partitioning hypervisors, and real-time containers.

Bio. Luigi De Simone is a tenure-track Assistant Professor at Dipartimento di Ingegneria Elettrica e delle Tecnologie dell’Informazione (DIETI), Università degli Studi di Napoli Federico II, where he leads research activities on the dependability of critical software infrastructures for cloud-edge and cyber-physical systems. His current research focuses on reliable and predictable container orchestration, robustness and isolation assessment of virtualization backends and hypervisors, and fault-injection methodologies for AI-enabled edge/cloud systems.

Abstract. TBD

Bio. Luigi De Simone is a tenure-track Assistant Professor at Dipartimento di Ingegneria Elettrica e delle Tecnologie dell’Informazione (DIETI), Università degli Studi di Napoli Federico II, where he leads research activities on the dependability of critical software infrastructures for cloud-edge and cyber-physical systems. His current research focuses on reliable and predictable container orchestration, robustness and isolation assessment of virtualization backends and hypervisors, and fault-injection methodologies for AI-enabled edge/cloud systems.

Abstract.Xen Project is the leading Open Source static partitioning hypervisor for embedded, industrial, and automotive systems. Its core role is that of an enforcer: ensuring strict isolation between domains so that the execution of one domain remains unaffected by others. With a microkernel design and a core smaller than 50K lines of code, Xen is well suited for mixed-criticality deployments where a real-time workload runs side-by-side with a complex guest such as Linux or Android. This presentation offers a 360-degree view of Xen for critical systems. We will start with the architecture and the key features that make Xen a top-tier solution: dom0less for fast and deterministic boot, cache coloring to reduce interference and improve interrupt latency, real-time scheduling, and techniques for safe inter-domain communication. We will explain how to build a real-time, mixed-criticality system in which an RTOS such as Zephyr meets its deadlines alongside general-purpose workloads. Along the way, we will dig into the details of the work required to make Xen real-time, with a particular focus on cache isolation technologies. We will then turn to safety. Since 2023, AMD, together with the Xen community, has been working to make Xen safety-certifiable to ISO 26262 ASIL D and IEC 61508 SIL 3, across both x86 and ARM. After obtaining Safety Concept Approval from the assessors, the project is now progressing toward the final certification phase. We will cover MISRA C compliance, safety requirements, end-to-end traceability, automated requirements-based testing, code coverage, and FMEA. Finally, we will ground these capabilities in real industrial use cases, showing how Xen is deployed today to consolidate safe and non-safe functions onto a single platform. Attendees will leave with a clear picture of where Xen stands as a real-time, safety-certifiable hypervisor and what to expect next.

Bio.Stefano Stabellini is a Fellow at AMD, where he leads system software architecture and the virtualization team. Previously, at Aporeto, he created a virtualization-based security solution for containers and authored several security articles. As Senior Principal Software Engineer in Citrix, he led a small group of passionate engineers working on Open Source projects. Stefano has been involved in Xen development since 2007. He created libxenlight in November 2009 and started the Xen port to ARM with virtualization extensions in 2011. Today he is a Xen Project committer, and he maintains Xen on ARM and Xen support in Linux and QEMU.

Abstract. This presentation explores the challenge of achieving low and predictable latencies in real-time distributed applications using Kubernetes as the orchestration platform. It includes a dissect of the common sources of latencies and jitter in distributed environments with the goal to provide actionable strategies to design, deploy, and maintain robust distributed systems that meet stringent real-time performance requirements.

Bio. Fredrik Svensson is a researcher at Ericsson, specializing in cloud and distributed systems. He holds an M.Sc. in Computer Science from Blekinge Institute of Technology (Sweden). His research interests span embedded and distributed systems, with a particular focus on mission-critical and real-time applications using cloud technologies.

Abstract. Time-Sensitive Networking (TSN) comprises a set of IEEE 802.1 and related Ethernet standards that extend standard Ethernet with real-time capabilities, including time synchronization, frame preemption, fault tolerance, traffic shaping, and scheduled traffic. These enhancements make TSN attractive across domains with diverse requirements, from industrial automation, where scalability, extensibility, and dynamic reconfiguration are critical, to automotive and driver-assistance systems, where cost, flexibility, and platform integration are key concerns. For applications with stringent timing requirements, safety-critical scheduled traffic (ST) can be provided with deterministic latency and low jitter while coexisting with best-effort (BE) traffic in the same multi-hop switched Ethernet network. One central TSN mechanism enabling such guarantees is the Time-Aware Shaper, defined in IEEE 802.1Qbv, typically used together with clock synchronization through IEEE 802.1AS-Rev to provide a shared notion of time across the network. Other TSN mechanisms complement TAS by addressing different system needs, including credit-based and asynchronous traffic shaping for bounded-latency traffic, frame preemption for reducing interference from lower-priority frames, per-stream filtering and policing for fault containment, and Frame Replication and Elimination for Reliability (FRER, IEEE 802.1CB) for seamless redundancy. However, the flexibility of TSN also introduces significant challenges. There is no universal guideline for selecting the appropriate mechanism, or combination of mechanisms, for a given application domain or set of requirements. Some combinations offer limited practical benefit, while others are more natural choices depending on traffic characteristics, timing constraints, and system architecture. Although individual TSN mechanisms are conceptually straightforward, their interactions remain insufficiently understood and may lead to unexpected behavior. In addition, the configuration space can be extremely large. Some mechanisms are relatively simple to configure, whereas others, such as schedule synthesis for IEEE 802.1Qbv, are computationally challenging and may be difficult to apply in dynamic settings such as industrial automation. As a result, configuration is one of the central challenges in the practical deployment of TSN. This talk will examine the end-to-end scheduling problem in TSN, its computational complexity, its relation to task scheduling, and the applicability of different TSN mechanisms across application domains. We will also discuss worst-case timing analysis techniques, including network calculus and worst-case response-time analysis, and illustrate the configuration challenge using automotive multi-core and multi-SoC architectures. Finally, we will discuss the industrial adoption of TSN, with a particular focus on how addressing the configuration problem can enable faster, safer, and more practical deployment of TSN in conservative safety-critical domains.

Bio. Silviu S. Craciunas is a Senior Principal Scientist at NXP Semiconductors, working primarily in the automotive domain, and a part-time Associate Professor at the Technical University of Denmark, where he focuses on Time-Sensitive Networking, edge AI, and safety-critical systems. Prior to joining NXP, he held research positions at TTTech Computertechnik AG and TTTech Auto AG in Vienna, Austria, working on deterministic networking and real-time systems across the industrial, aerospace, and automotive domains. His research interests include real-time and safety-critical distributed systems, real-time scheduling algorithms, and deterministic networks. His recent work focuses on designing scalable methods and algorithms for complex scheduling problems in multi-core and multi-SoC platforms, spanning both application and network domains. He has received multiple best and outstanding paper awards at venues including RTNS, ETFA, ECRTS, and RTAS. He served as TPC co-chair for RTNS 2026 and as Industry Session Co-Chair for RTSS 2024 and WFCS 2024, and has been a member of the technical program committees of several leading conferences, including RTNS, ECRTS, and RTAS.

Abstract. Many cyber-physical systems are distributed computer systems, and a synchronized global time service provides any two non-faulty computers in such a system with a common view of the current point in real time (with some error margin we call the “precision”). A synchronized global time has many use cases, ranging from simple timestamping to time-triggered behavior, a pre-planned coordinated behavior of the distributed computers. In this lecture, we will explore the various use cases of a synchronized global time in light of industry-specific requirements. We will also study synchronization protocols that establish the synchronized global time service.

Bio. Wilfried Steiner is Director of TTTech Labs and Privatdozent (Habilitated Lecturer) in Computer Science at TU Wien, where he teaches dependable cyber-physical systems. His work focuses on dependable and deterministic distributed systems, with contributions to applications in automotive, aerospace, space, and industrial automation. He is a principal architect and co-designer of TTEthernet (SAE AS6802) and a contributor to IEEE 802.1 TSN, with key contributions to fault-tolerant communication and scheduling in mixed-criticality real-time systems. His work pioneered the use of SMT-based configuration synthesis for time-triggered networks, enabling strong correctness guarantees in both research and industrial practice. At TTTech, he has led major technology programs and advanced the adoption of rigorous dependability engineering, integrating formal methods and scheduling into production systems. He has collaborated with organizations such as NASA and SRI International and is an inventor on multiple patents in dependable real-time communication. His current work focuses on the design and verification of highly dependable, safety-critical distributed systems of systems, particularly in aerospace and space applications.

Abstract. Terrestrial cloud computing has transformed how data is processed, stored, and delivered. However, its core assumptions—abundant resources, stable connectivity, elastic infrastructure, and centralized data centers—do not directly apply to space systems. Satellites operate under strict constraints on power, computation, communication, timing, and maintainability. At the same time, future missions increasingly require more onboard data processing, autonomous decision-making, and cooperation across satellite constellations. This lecture discusses why a simple transfer of terrestrial cloud concepts to space is insufficient, and why space systems need their own real-time, distributed, and mission-aware edge-computing model. As a use case, the lecture considers a satellite constellation performing time-critical mission, where sensing, AI-based processing, inter-satellite communication, and selective downlink must be scheduled under real-time and resource constraints. The use case opens several research questions, including task allocation, deadline-aware scheduling, communication-computation trade-offs, resilience, and security in future space edge-cloud systems.

Bio. Dr. Zain Hammadeh is a researcher at the German Aerospace Center (DLR), Institute of Software Technology. He leads the research of the Real-Time Software and Security group, where his work focuses on dependable and secure software for safety-critical systems, with a particular emphasis on space applications. He works at the intersection of real-time computing and cybersecurity, with the goal of enabling secure, predictable, and resilient space software for future autonomous and distributed missions.