Emmanuel Grolleau

Scheduling Dependent Periodic Tasks without Synchronization Mechanisms

This article studies the scheduling of critical embedded systems, which consist of a set of communicating periodic tasks with constrained deadlines. Currently, tasks are usually sequenced manually, partly because available scheduling policies do not ensure the determinism of task communications. Ensuring this determinism requires scheduling policies supporting task precedence constraints (which we call dependent tasks), which are used to force the order in which communicating tasks execute. We propose fixed priority scheduling policies for different classes of dependent tasks: with simultaneous or arbitrary release times, with simple precedences (between tasks of the same period) or extended precedences (between tasks of different periods). We only consider policies that do not require synchronization mechanisms (like semaphores). This completely prevents deadlocks or scheduling anomalies without requiring further proofs.

Dynamic priority scheduling of periodic tasks with extended precedences

The software architecture of a critical embedded control system generally consists of a set of multi-periodic communicating tasks. In order to be able to describe such a system, we define the notion of semaphore precedence constraint, which supports multi-rate communications that follow regular repetitive patterns. We propose a feasibility test for EDF and we study three implementations, for periodic task sets related by such extended precedences on monoprocessor architectures.

Off-Line Computation of Real-Time Schedules by Means of Petri nets

We present an off-line methodology of analysis of real-time systems, composed of periodic, precedence and resource constrained real-time tasks. As there is no polynomial optimal scheduling technique for such tasks sets, we present an enumerative method based on the construction of the state graph of a Petri net. The time is modeled by the Petri net through the earliest firing rule.

Periodicity of real-time schedules for dependent periodic tasks on identical multiprocessor platforms

This paper gives and proves correct a simulation interval for any schedule generated by a deterministic and memoryless scheduler (i.e., one where the scheduling decision is the same and unique for any two identical system states) for identical multiprocessor platforms. We first consider independent periodic tasks, then generalize the simulation interval to tasks sharing critical resources, and subject to precedence constraints or self-suspension. The simulation interval is based only on the periods, release times and deadlines, and is independent from any other parameters. It is proved large enough to cover any feasible schedule produced by any deterministic and memoryless scheduler on multiprocessor platforms, including non conservative schedulers. To the best of our knowledge, this simulation interval covers the largest class of task systems and scheduling algorithms on identical multiprocessor platforms ever studied. This simulation interval is used to derive a simulation algorithm using a linear space complexity. Finally, a generic exact schedulability test based on simulation is presented. This test can be applied only when sustainability is consistent with online variability of the tasks’ parameters.

Reducing the gap between design and scheduling

Model-based design techniques for real-time systems have limited real-time expressiveness, focusing their abilities to classic scheduling models (like the classic sporadic model) and to a reduced set of temporal analysis techniques (like the Rate Monotonic Analysis). Moreover, to perform analysis with the real-time scheduling theory, the system designers must check that their models are compliant with the assumptions of this theory. This article introduces an open meta-model, based on model-driven engineering, which aims to cover new real-time scheduling models and techniques. Therefore, it will be possible to connect several independent schedulability analysis tools, following closely the advances in real-time scheduling theory, dealing with a temporal model that will be covered by our meta-model. This connection can be done at different stages of the design (early for sensitivity analysis, at a later stage for temporal validation) to create a temporal model of the designed system, and to assist the designer who is not necessarily an expert in scheduling theory. This paper can be considered as an attempt to motivate the real-time community to have a taxonomy of real-time scheduling models, problems and analysis techniques.

Characterization and Analysis of Tasks with Offsets: Monotonic Transactions

This article introduces the concept of monotonic transactions. A monotonic transaction is a particular case of transactions for which the load arrival pattern is (or can be by rotation) localized at the beginning of the transaction. In the general context of tasks with offsets (general transactions) only exponential methods are known to calculate the worst-case response time. The pseudo-polynomial methods known give an upper bound of the worst-case response time. The method of analysis suggested in this article gives the real worst-case response time: moreover; this method has a complexity lower than that of the existing methods of approximation. There are two main steps in the application of this method: grouping the tasks of the transaction in a normal form and seeking a monotonic pattern

An fptas for Response Time Analysis of Fixed Priority Real-Time Tasks with Resource Augmentation

Response time analysis is required both for on-line admission of applications in dynamic systems and as an integral part of design tools for complex distributed real-time systems. We consider sporadic tasks with fixed-priorities and arbitrary deadlines to be executed upon a uniprocessor platform. Pseudo-polynomial time algorithms are known for computing exact worst-case response times for this task model. Nevertheless, the problem is known NP-hard and there cannot exist a constant approximation algorithm for response time computation, unless P=NP. We propose a fully polynomial time approximation scheme (FPTAS) for computing response time upper bounds under resource augmentation. The resource augmentation is defined as the processor speedup factor bounded by (1 + 1/k), where kdef [1 = ε]-1 for any constant ε ∈ (0;1), the FPTAS accuracy parameter. This algorithm is best possible in the sense that resource augmentation is indeed necessary for an efficient response time calculation.

Model Driven Timing Analysis for Real-Time Systems

This paper stresses the difficulty for a system designer to use an appropriate real-time task model for his system, and to choose the associated scheduling analysis tests/dimensioning techniques. We propose a model-based approach tackling this difficulty. We focus on the schedulability analysis tree used by our method in order to help the designer to dimension, and then to validate his system.

An efficient response-time analysis for real-time transactions with fixed priority assignment

In the general context of tasks with offsets (general transactions), only exponential methods are known to calculate the exact worst-case response time (WCRT) of a task. The known pseudo-polynomial techniques give an upper bound of the WCRT. In this paper, we present a new worst-case response-time analysis technique (mixed method) for transactions scheduled by fixed priorities, and executing on a uniprocessor system. This new analysis technique gives a better (i.e. lower) pseudo-polynomial upper bound of the WCRT. The main idea is to combine the principle of exact calculation and the principle of approximation calculation, in order to decrease the pessimism of WCRT analysis, thus allowing improving the upper bound of the response time provided while preserving a pseudo-polynomial complexity. Then we define the Accumulative Monotonic property on which a necessary condition of feasibility is discussed. We also propose, to speed up the exact and mixed analysis, the dominated candidate task concept that allows reducing significantly the number of critical instants to study in an analysis.

Feasibility analysis of real-time transactions

The objective of this paper is two-fold: give a survey of response time analysis (RTA), and contribute to schedulability analysis for the real-time transaction model. The RTA is studied under fixed priority policies (FPP), while schedulability analysis assumes an optimal scheduling algorithm (like the deadline driven scheduling algorithm EDF) in a preemptive context on uniprocessor systems. We compare the transaction model to the family of multiframe models, then present the exact, and approximated methods, as well as a tunable method to compute the RTA. Finally we present a new schedulability analysis method and an efficient algorithm to speed up this test.