Nikolay Stoimenov

Real-time interfaces for composing real-time systems

Recently, a number of frameworks were proposed to extend interface theory to the domains of single-processor and distributed real-time systems. This paper unifies some of these approaches and proves properties like refinement and independent implementability. We also explicitly state the requirements to a framework for these properties to be fulfilled. Further, a new notion of adaptive interfaces is introduced that supports the design by providing mechanisms for propagating system constraints, such as (end-to-end) delays, available computing and communication resources, buffer spaces, and energy. Guarantees and assumptions on interfaces are not any longer static but adapt according to the system environment. This can be used to answer synthesis questions at design time or to adapt system parameters to changing environment requirements at run-time. The applicability of the presented framework is proven by adapting it to a number of different real-time analysis models.

Scheduling of mixed-criticality applications on resource-sharing multicore systems

A common trend in real-time safety-critical embedded systems is to integrate multiple applications on a single platform. Such systems are known as mixed-criticality (MC) systems as the applications are usually characterized by different criticality levels (CLs). Nowadays, multicore platforms are promoted due to cost and performance benefits. However, certification of multicore MC systems is challenging because concurrently executed applications with different CLs may block each other when accessing shared platform resources. Most of the existing research on multicore MC scheduling ignores the effects of resource sharing on the execution times of applications. This paper proposes a MC scheduling strategy which explicitly accounts for these effects. Applications are executed by a flexible time-triggered criticality-monotonic scheduling scheme. Schedulers on different cores are dynamically synchronized such that only a statically known subset of applications of the same CL can interfere on shared resources, e. g.,memories, buses. Therefore, the timing effects of resource sharing are bounded and we quantify them at design time. We combine this scheduling strategy with a mapping optimization technique for achieving better resource utilization. The efficiency of the approach is demonstrated through extensive simulations as well as comparisons with traditional temporal partitioning and state-of-the-art scheduling algorithms. It is also validated on a real-world avionics system.

Reliable mode changes in real-time systems with fixed priority or EDF scheduling

Many application domains require adaptive real-time embedded systems that can change their functionality over time. In such systems it is not only necessary to guarantee timing constraints in every operating mode, but also during the transition between different modes. Known approaches that address the problem of timing analysis over mode changes are restricted to fixed priority scheduling policies. In addition, most of them are also limited to simple periodic event stream models and therefore, they can not faithfully abstract the bursty timing behavior which can be observed in embedded systems. In this paper, we propose a new method for the design and analysis of adaptive multi-mode systems that supports any event stream model and can handle earliest deadline first (EDF) as well as fixed priority (FP) scheduling of tasks. We embed the analysis method into a well-established modular performance analysis framework based on Real-Time Calculus and prove its applicability by analyzing a case study.

Mixed-criticality scheduling on cluster-based manycores with shared communication and storage resources

The embedded system industry is facing an increasing pressure for migrating from single-core to multi- and many-core platforms for size, performance and cost purposes. Real-time embedded system design follows this trend by integrating multiple applications with different safety criticality levels into a common platform. Scheduling mixed-criticality applications on today’s multi/many-core platforms and providing safe worst-case response time bounds for the real-time applications is challenging given the shared platform resources. For instance, sharing of memory buses introduces delays due to contention, which are non-negligible. Bounding these delays is not trivial, as one needs to model all possible interference scenarios. In this work, we introduce a combined analysis of computing, memory and communication scheduling in a mixed-criticality setting. In particular, we propose: (1) a mixed-criticality scheduling policy for cluster-based many-core systems with two shared resource classes, i.e., a shared multi-bank memory within each cluster, and a network-on-chip for inter-cluster communication and access to external memories; (2) a response time analysis for the proposed scheduling policy, which takes into account the interferences from the two classes of shared resources; and (3) a design exploration framework and algorithms for optimizing the resource utilizations under mixed-criticality timing constraints. The considered cluster-based architecture model describes closely state-of-the-art many-core platforms, such as the Kalray MPPA®-256. The applicability of the approach is demonstrated with a real-world avionics application. Also, the scheduling policy is compared against state-of-the-art scheduling policies based on extensive simulations with synthetic task sets.

Modular performance analysis of cyclic dataflow graphs

Applications for parallel and distributed embedded systems are often specified as dataflow graphs with dependency cycles. Examples of corresponding models of computation are marked graphs or synchronous dataflow (SDF) graphs. Performance analysis is often used in the exploration of different implementation alternatives or in order to provide guarantees on the timing behavior. This paper describes a new approach to the modular performance analysis of cyclic dataflow graphs such as SDF graphs as existing component-based analysis methods are not able to faithfully deal with cycles in the event flow. The new method results in tight bounds on essential quantities like buffer sizes, end-to-end delays and throughput. Because of the generality of the approach, one can analyze not only systems that can be modeled as marked graphs but also implementations that contain buffers with finite sizes, that produce system-wide back-pressure caused by blocking write semantics. The embedding of the novel approach into a modular performance analysis method allows the analysis of distributed implementations that use resource sharing mechanisms such as fixed-priority scheduling and time division multiple access (TDMA). The paper presents the new models and methods as well as experimental results.

Mapping mixed-criticality applications on multi-core architectures

A common trend in real-time embedded systems is to integrate multiple applications on a single platform. Such systems are known as mixed-criticality (MC) systems when the applications are characterized by different criticality levels. Nowadays, multicore platforms are promoted due to cost and performance benefits. However, certification of multicore MC systems is challenging as concurrently executed applications of different criticalities may block each other when accessing shared platform resources. Most of the existing research on multicore MC scheduling ignores the effects of resource sharing on the response times of applications. Recently, a MC scheduling strategy was proposed, which explicitly accounts for these effects. This paper discusses how to combine this policy with an optimization method for the partitioning of tasks to cores as well as the static mapping of memory blocks, i.e., task data and communication buffers, to the banks of a shared memory architecture. Optimization is performed at design time targeting at minimizing the worst-case response times of tasks and achieving efficient resource utilization. The proposed optimization method is evaluated using an industrial application.

Service adaptions for mixed-criticality systems

Complex embedded systems are typically mixed-critical, where heterogeneous guarantees must be provided for functionalities of different criticalities. We study in this paper the reconfiguration of services provided to low criticality tasks in reaction to the overruns of high criticality tasks. We further investigate the quantification of the resetting time of the system services. For both service reconfiguration and resetting, we derive tight analysis results under Earliest Deadline First (EDF) scheduling.

Timed model checking with abstractions: towards worst-case response time analysis in resource-sharing manycore systems

Multicore architectures are increasingly used nowadays in embedded real-time systems. Parallel execution of tasks feigns the possibility of a massive increase in performance. However, this is usually not achieved because of contention on shared resources. Concurrently executing tasks mutually block their accesses to the shared resource, causing non-deterministic delays. Timing analysis of tasks in such systems is then far from trivial. Recently, several analytic methods have been proposed for this purpose, however, they cannot model complex arbitration schemes such as FlexRay which is a common bus arbitration protocol in the automotive industry. This paper considers real-time tasks composed of superblocks, i.e., sequences of computation and resource accessing phases. Resource accesses such as accesses to memories and caches are synchronous, i.e., they cause execution on the processing core to stall until the access is served. For such systems, the paper presents a state-based modeling and analysis approach based on Timed Automata which can model accurately arbitration schemes of any complexity. Based on it, we compute safe bounds on the worst-case response times of tasks. The scalability of the approach is increased significantly by abstracting several cores and their tasks with one arrival curve, which represents their resource accesses and computation times. This curve is then incorporated into the Timed Automata model of the system. The accuracy and scalability of the approach are evaluated with a real-world application from the automotive industry and benchmark applications.

Feasibility Analysis of On-Line DVS Algorithms for Scheduling Arbitrary Event Streams

Performance boosting of modern computing systems has been constrained by the significant chip/circuit power dissipation. Dynamic voltage scaling (DVS) has been applied in the past decade for reducing the energy consumption by dynamically changing the supply voltage. On-line scheduling algorithms for DVS systems usually guarantee the real-time constraints of the system based on the condition that they can select any system speed that is sufficiently high to allow processing of all events within their deadlines. However, practical systems have a maximum available system speed and the feasibility of using on-line DVS algorithms needs to be verified during design time, i.e., they will never require during runtime a speed higher than the maximum available. This paper presents feasibility analysis of two on-line DVS algorithms that can compute in advance an upper bound on the system speed that these algorithms may require given that there is a single input event stream described by the worst-case event arrivals in interval domain. Moreover, we also present new results on the competitive ratios of the resulting schedules for energy consumption minimization with comparison to the off-line optimal solutions to show the effectiveness of the two algorithms. At the end, the performance of the different algorithms is evaluated.

Designing energy-efficient NoC for real-time embedded systems through slack optimization

Hard real-time embedded systems impose a strict latency requirement on interconnection subsystems. In the case of network-on-chip (NoC), this means each packet of a traffic stream has to be delivered within a time interval. In addition, with the increasing complexity of NoC, it consumes a significant portion of total chip power, which boosts the power footprint of such chips. In this work, we propose a methodology to minimize the energy consumption of NoC without violating the prespecified latency deadlines of real-time applications. First, we develop a formal approach based on network calculus to obtain the worst-case delay bound of all packets, from which we derive a safe estimate of the number of cycles that a packet can be further delayed in the network without violating its deadline- the worst-case slack. With this information, we then develop an optimization algorithm that trades the slacks for lower NoC energy. Our algorithm recognizes the distribution of slacks for different traffic streams, and assigns different voltages and frequencies to different routers to achieve NoC energy-efficiency, while meeting the deadlines for all packets.