Les systèmes embarqués temps réel nécessitent une analyse temporelle pour valider leur comportement temporel.Afin de réduire le coût de développement, l’analyse doit être effectuée à une phase précoce au moment de la modélisationpour détecter les anomalies de conception. L’analyse d’ordonnançabilité est une des analyses temporelles qui permettentde s’assurer du bon fonctionnement du système conçu. Elle est issue de la théorie de l’ordonnancement temps réel.Plusieurs méthodes et tests analytiques ont été proposés par la communauté académique mais peu sont les tests adoptéspar les industriels. En effet, l’utilisation des tests d’analyse exige une large connaissance des travaux de recherches - misà jour régulièrement - afin de choisir la méthode la plus adaptée aux systèmes conçus pour les valider ou les dimensionnerpour le cas des systèmes qui sont en cours de conception.Cette thèse s’intéresse à cette utilisation minimaliste de la théorie de l’ordonnancement dans l’industrie, et propose dessolutions d’aide à la décision basées sur l’ingénierie dirigée par les modèles. Nos solutions visent à augmenter l’applicabilitéde la théorie de l’ordonnancement, à faciliter le choix des tests appropriés et à réduire le surdimensionnement qui peutêtre généré au moment de la conception. Ces solutions sont implémentées dans un Framework appelé MoSaRT offrantdes fonctionnalités pour les concepteurs (modeleurs et analystes) afin d’améliorer le processus de conception des systèmestemps réel en vue de leur ordonnançabilité. / Real-time embedded systems need to be analyzed at an early stage in order to detect temporal vulnerabilities. Indeed,software development costs are sharply impacted by wrong design choices made in the early stages of development, inparticular during the design phase, but often detected after the implementation. The schedulability analysis is one of themain analyses required to ensure the timing correctness of a real-time system. Indeed, the real-time scheduling theoryhas been devoted to propose different models providing several levels of expressiveness, and different analytical methodswith different levels of accuracy. The utilization of the real-time scheduling theory in practical cases could be profitable.Unfortunately, it is not sufficiently applied and research results have been exploited in an industrial context only to amodest extent to date. Actually, a difficulty faced by the real-time designers is to find the appropriate analysis testshelping to validate properly the system and also to reduce the over-dimensioning. This thesis is interested in the chasmexisting between real-time design community and real-time analysis community. The purpose of our work is to fill thegap between the modeling of real-time systems and the scheduling analysis. Then, we propose a decision aiding solutionbased on model-driven engineering. Our solution is dedicated (i) to increase the usage of the real-time scheduling theory,(ii) to facilitate the scheduling analysis tests choice and (iii) to reduce the pessimism during the design phase of real-timesystems.Our proposal is embodied as a framework unifying the designers and analysts efforts. The framework called MoSaRToffers a design language very close to the taxonomy of real-time scheduling theory. Moreover, MoSaRT provides ananalysis repository concept to enhance the applicability of the scheduling theory and also to improve the way designerscheck their designs.

As embedded software and platforms become more complicated, many safety properties are left to simulation and testing. MRICDF is a formal polychronous language used to guarantee certain safety properties and alleviate the burden of software development and testing. We propose real-time extensions to MRICDF so that temporal properties of embedded systems can also be proven. We adapt the extended precedence encoding technique of Prelude and expand upon current schedulability analysis techniques for multi-periodic real-time systems. / Master of Science

Cyber-Physical Systems (CPS) are widely deployed in critical applications which are subject to strict timing constraints. To ensure correct timing behavior, much of the effort has been dedicated to the development of validation and verification methods for CPS (e.g., system models and their timing and schedulability analysis). As CPS is becoming increasingly complex, there is an urgent need for efficient optimization techniques that can aid the design of large-scale systems. Specifically, techniques that can find good design options in a reasonable amount of time while meeting all the timing and other critical requirements are becoming vital. However, the current mindset is to use existing schedulability analysis and optimization techniques for the design optimization of time-critical CPS. This has resulted in two issues in today's CPS design: 1) Existing timing and schedulability analysis are very difficult and inefficient to be integrated into well-established optimization frameworks such as mathematical programming; 2) New system models and timing analysis are being developed in a way that is increasingly unfriendly to optimization. Due to these difficulties, existing practice for optimization mostly relies on meta or ad-hoc heuristics, which suffers either from sub-optimality or limited applicability. In this dissertation, we seek to address these issues and explore two new directions for developing optimization algorithms for time-critical CPS. The first is to develop {em optimization-oriented timing analysis}, that are efficient to formulate in mathematical programming framework. The second is a domain-specific optimization framework. The framework leverages domain-specific knowledge to provide methods that abstract timing analysis into a simple mathematical form. This allows to efficiently handle the complexity of timing analysis in optimization algorithms. The results on a number of case studies show that the proposed approaches have the potential to significantly improve upon scalability (several orders of magnitude faster) and solution quality, while being applicable to various system models, timing analysis techniques, and design optimization problems in time-critical CPS. / Doctor of Philosophy / Cyber-Physical Systems (CPS) tightly intertwine computing units and physical plants to accomplish complex tasks such as control and monitoring. They are often deployed in critical applications subject to strict timing constraints. For example, many control applications and tasks are required to finished within bounded latencies. To guarantee such timing correctness, much of the effort has been dedicated to studying methods for delay and latency estimation. These techniques are known as schedulability analysis/timing analysis. As CPS becomes increasingly complex, there is an urgent need for efficient optimization techniques that can aid the design of large-scale and correct CPS. Specifically, techniques that can find good design options in reasonable amount of time while meeting all the timing and other critical requirements are becoming vital. However, most of the existing schedulability analysis are either non-linear, non-convex, non-continuous or without closed form. This gives significant challenge for integrating these analysis into optimization. In this dissertation, we explore two new paradigm-shifting approaches for developing optimization algorithms for the design of CPS. Experimental evaluations on both synthetic and industrial case studies show that the new approaches significantly improve upon existing optimization techniques in terms of scalability and quality of solution.

No description available.

The requirements for real-time systems in safety-critical applications typically contain strict timing constraints. The design of such a system must be subject to extensive validation to guarantee that critical timing constraints will never be violated while the system operates. A mathematically rigorous technique to do so is to perform a schedulability analysis for formally verifying models of the computational workload. Different workload models allow to describe task activations at different levels of expressiveness, ranging from traditional periodic models to sophisticated graph-based ones. An inherent conflict arises between the expressiveness and analysis efficiency of task models. The more expressive a task model is, the more accurately it can describe a system design, reducing over-approximations and thus minimizing wasteful over-provisioning of system resources. However, more expressiveness implies higher computational complexity of corresponding analysis methods. Consequently, an ideal model provides the highest possible expressiveness for which efficient exact analysis methods exist. This thesis investigates the trade-off between expressiveness and analysis efficiency. A new digraph-based task model is introduced, which generalizes all previously proposed models that can be analyzed in pseudo-polynomial time without using any analysis-specific over-approximations. We develop methods allowing to efficiently analyze variants of the model despite their strictly increased expressiveness. A key contribution is the notion of path abstraction which enables efficient graph traversal algorithms. We demonstrate tractability borderlines for different classes of schedulers, namely static priority and earliest-deadline first schedulers, by establishing hardness results. These hardness proofs provide insights about the inherent complexity of developing efficient analysis methods and indicate fundamental difficulties of the considered schedulability problems. Finally, we develop a novel abstraction refinement scheme to cope with combinatorial explosion and apply it to schedulability and response-time analysis problems. All methods presented in this thesis are extensively evaluated, demonstrating practical applicability.

In this thesis, we study temperature-constrained multiprocessor real-time systems, where real-time guarantees must be met without exceeding safe temperature levels within the processors. We focus on Pfair scheduling algorithms, especially ERfair scheduling scheme (a work-conserving extension to Pfair scheduling) as our main multiprocessor real-time scheduling methodology. Then, we study the benefits of simple reactive speed scaling as described in the real-time multiprocessor systems. In this thesis, in support of the temperature-awareness, we extend the applicability of the reactive speed scaling to global scheduling schemes for multiprocessors. We propose temperature-aware scheduling and processor selection schemes motivated by existing (thermally non-optimal) ERfair scheduling in order to reduce thermal stress and therefore increase the processor utilization. Then, we show that the proposed algorithm and reactive scheme can enhance the processor utilization compared with any constant speed scheme on real-time multiprocessor systems. Additionally, we show how the maximum schedulable utilization (MSU) for partitioning heuristics can be determined on the temperature-constrained multiprocessor real-time systems.

Embedded applications have become more and more complex, increasing the demands on the communication network. For reasons such as safety and usability, there are real-time constraints that must be met. Also, to offer high performance, network protocols should offer efficient user services aimed at specific types of communication. At the same time, it is desirable to design and implement embedded networks with reduced cost and development time, which means using available hardware for standard networks. To that end, there is a trend towards using switched Ethernet for embedded systems because of its hight bit rate and low cost. Unfortunately, since switched Ethernet is not specifically designed for embedded systems, it has several limitations such as poor support for QoS because of FCFS queuing policy and high protocol overhead. This thesis contributes towards fulfilling these requirements by developing (i) real-time analytical frameworks for providing QoS guarantees in packet-switched networks and (II) packet-merging techniques to reduce the protocol overhead. We have developed two real-time analytical frameworks for networks with FCFS queuing in the switches, one for FCFS queuing in the source nodes and one for EDF queuing in the source nodes. The correctness and tightness of the real-time analytical frameworks for different network components in a singel-switch neetwork are given by strict theoretical proofs, and the performance of our end-to-end analyses is evaluated by simulations. In conjunction with this, we have compared our results to Network Calculus (NC), a commonly used analytical scheme for FCFS queuing. Our comparison study shows that our anlysis is more accurate than NC for singel-switch networks. To reduce the protocol overhead, we have proposed two active switched Ethernet approaches, one for real-time many-to-many communication and the other for the real-time short message traffic that is often present in embedded applications. A significant improvement in performance achieved by using our proposed active networks is demonstrated. Although our approaches are exemplified using switched Ethernet, the general approaches are not limited to switched Ethernet networks but can easily be moified to other similar packet-switched networks.

The ever-increasing functional and nonfunctional requirements in real-time safety-critical embedded systems call for new design flows that solve the specification, validation, and synthesis problems. Ensuring key properties, such as functional determinism and temporal predictability, has been the main objective of many embedded system design models. Dataflow models of computation (such as KPN, SDF, CSDF, etc.) are widely used to model stream-based embedded systems due to their inherent functional determinism. Since the introduction of the (C)SDF model, a considerable effort has been made to solve the static-periodic scheduling problem. Ensuring boundedness and liveness is the essence of the proposed algorithms in addition to optimizing some nonfunctional performance metrics (e.g. buffer minimization, throughput maximization, etc.). However, nowadays real-time embedded systems are so complex that real-time operating systems are used to manage hardware resources and host real-time tasks. Most of real-time operating systems rely on priority-driven scheduling algorithms (e.g. RM, EDF, etc.) instead of static schedules which are inflexible and difficult to maintain. This thesis addresses the real-time scheduling problem of dataflow graph specifications; i.e. transformation of the dataflow specification to a set of independent real-time tasks w.r.t. a given priority-driven scheduling policy such that the following properties are satisfied: (1) channels are bounded and overflow/underflow-free; (2) the task set is schedulable on a given uniprocessor (or multiprocessor) architecture. This problem requires the synthesis of scheduling parameters (e.g. periods, priorities, processor allocation, etc.) and channel capacities. Furthermore, the thesis considers two performance optimization problems: buffer minimization and throughput maximization.

La Tesis presenta el desarrollo de una metodología de análisis y diseño de sistemas distribuidos de tiempo real estricto, y su aplicación a una implementación práctica en lenguaje Ada.Se han optimizado los métodos existentes para la planificación y análisis de sistemas distribuidos de tiempo real mediante un algoritmo heurístico para la asignación de prioridades, y la aplicación del algoritmo de servidor esporádico a la planificación de redes de comunicación de tiempo real. También se ha ampliado el campo de aplicación del análisis a sistemas más complejos en los que existe sincronización por intercambio de eventos o paso de mensajes.Se ha demostrado que la metodología propuesta se puede implementar en sistemas de tiempo real prácticos, a través de su aplicación a sistemas distribuidos programados en lenguaje Ada. / The Thesis presents a methodology to analyze and design distributed real-time systems, and its application to a practical implementation.Existing methods for scheduling and analyzing distributed real-time systems have been optimized through a new heuristic algorithm for assigning priorities, and with the application of the sporadic server algorithm for scheduling real-time communication networks. The area of application of the analysis has been extended to more complex systems, like those with synchronization through event exchange or message passing.It has been demonstrated that the proposed methodology can be implemented in practical real-time systems, through the application to a distributed system programmed in the Ada language.

Se propone una metodología de desarrollo de aplicaciones de tiempo real estricto que van a ser ejecutadas en plataformas distribuidas abiertas. En esta metodología, el diseñador de la aplicación no conoce la carga de trabajo de la plataforma que será ejecutada concurrentemente junto con la aplicación que diseña. La metodología se basa en el paradigma de reserva de recursos, y utiliza como base el concepto de plataforma virtual, tanto para describir el uso de los recursos que una aplicación requiere, como para ejecutar la aplicación satisfaciendo sus requisitos temporales. La plataforma virtual es utilizada en el proceso de negociación con el servicio de reserva de recursos de la plataforma física, con objeto de obtener una configuración de la aplicación que haga compatible su ejecución con la carga de trabajo que ya se está ejecutando en dicha plataforma. La metodología aborda todas las fases del desarrollo de una aplicación: describe la información que debe asociarse al código de la aplicación para poder ser configurado, así como el proceso que permite analizar independientemente su planificabilidad en base a la plataforma virtual; especifica el proceso de despliegue de la aplicación y define la información que se utiliza para negociar su ejecución con el servicio de reserva de recursos de la plataforma física y para generar los datos de configuración que deben ser asignados al código cuando se ejecute. Todos estos procesos son dirigidos por modelos, por lo que la tesis aborda la definición de las transformaciones de modelos requeridas, así como la formulación de los metamodelos formales utilizados en ellas. Por otro lado, aunque la tecnología es independiente de la plataforma de ejecución, se especifica la funcionalidad que debe ofrecer el servicio de reserva de recursos presente en la misma para dar soporte a la metodología propuesta, y se analiza su compatibilidad con algunas implementaciones actualmente disponibles / This thesis proposes a methodology for the development of hard real-time applications that will be executed in open distributed platforms. When this methodology is applied, the application designer does not know the workload of the platform that will execute concurrently with the designed application. The methodology is based on the resource reservation paradigm, and relies on the concept of virtual platform, both to describe the resources usage required by an application to execute, and to run the application guaranteeing the fulfillment of the specified timing requirements. The virtual platform is also used on the negotiation process with the resource reservation service of the physical platform in order to obtain a configuration of the application that supports its execution together with the current workload running on that platform. The methodology deals with all the phases of the application design: it describes the information that must be associated to the application code in order to obtain a proper configuration, as well as the process that allows an independent schedulability analysis of the application based on its virtual platform; it specifies the application deployment process and defines the information that is used to negotiate the execution of the application with the resource reservation service of the physical platform, and to generate the configuration data that must be assigned to the code when it is executed. The methodology follows a model-driven perspective, so the thesis addresses the required models transformations, as well as the formulation of the metamodels used in them. Moreover, although the technology is independent from the execution platform, the functionality that must be provided by the resource reservation service to support the proposed technology is specified and its compatibility with other implementations is analyzed.