Guy Vidal-Naquet

Petri nets and regular languages

It is shown that the regularity problem for firing sequence sets of Petri nets is decidable. For the proof, new techniques to characterize unbounded places are introduced. In the class L0 of terminal languages of labelled Petri nets the regularity problem in undecidable. In addition some lower bounds for the undecidability of the equality problems in L0 and L are given. L0λ is shown to be not closed under complementation without reference to the reachability problem.

A Generic Execution Framework for Models of Computation

The model driven engineering approach has had an important impact on the methods used for the conception of systems. However, some important difficult points remain in this domain. In this paper, we focus on problems related to the heterogeneity of the computation models (and therefore of the modeling techniques) used for the different aspects of a system and to the validation and the execution of a model. We present here a language for describing computation models, coupled with a generic execution platform where different computation models as well as their composition can be interpreted. Our goal is to be able to describe precisely the semantics of the computation models underlying domain specific languages, and to allow the interpretation of these models within our platform. This provides for a non ambiguous definition of the behavior of heterogeneous models of a system, which is essential for validation, simulation and code generation

Equivalence between Schedule Representations: Theory and Applications

Multiprocessor scheduling problems are hard because of the numerous constraints on valid schedules to take into account. This paper presents new schedule representations in order to overcome these difficulties, by allowing processors to be fractionally allocated. We prove that these representations are equivalent to the standard representations when preemptive scheduling is allowed. This allows the creation of scheduling algorithms and the study of feasibility in the simpler representations. We apply this method throughout the paper. Then, we use it to provide new simple solutions to the previously solved implicit-deadline periodic scheduling problem. We also tackle the more general problem of scheduling arbitrary time-triggered tasks, and thus in particular solve the open multiprocessor general periodic tasks scheduling problem. Contrary to previous solutions like the PFair class of algorithms, the proposed solution also works when processors have different speeds. We complete the method by providing an online schedule transformation algorithm, that allows the efficient handling of both time-triggered and event-triggered tasks, as well as the creation of online rate-based scheduling algorithms on multiprocessors.

A communication mechanism for resource isolation

Sharing resources between multiple untrusted clients requires a shared service that provides access to the resources upon client requests. But executing these requests needs other resources, like memory or CPU time, which must be carefully allocated. In this paper, we investigate a communication mechanism that allows access to shared services without changing existing allocation decisions. This is achieved by systematically using the new resource lending principle, that allows a service to use the resources of its clients to perform the request. We present an easily understandable design model for this communication mechanism named the thread lending model, that completely avoids any allocation by the service, and demonstrate its implementation in our prototype OS Anaxagoros. We finally investigate the consequences of using this model on the structure and implementation of the shared services.

An Introduction to Time-Constrained Automata

We present timing-constrained automata (TCA), a model for real-time computation in which agents behaviors are modeled by automata constrained by time intervals. In this model time does not change automata behavior: on the contrary, it is automata execution that changes the timing constraints. This allows the model to well lend itself to analysis as well as concrete execution. TCA actions model can have multiple start time and deadlines, can be aperiodic, and can change dynamically following a graph, the time-constrained automaton. This allows expressing much more precise timing constraints than classical periodic or sporadic model, while preserving the ease of scheduling and analysis. We provide some properties of this model as well as their scheduling semantics. We show that timing-constrained tasks can be automatically derived from source-code, and optimally scheduled on single processors using a variant of EDF. We explain how timing constraints can be used to guarantee communication determinism by construction, and used to study when possible agent interactions happens.

Deterministic Languages of Petri Nets

One way of using Petri nets for modelling systems is to represent states by markings, and actions that modify the states by transitions. In order to represent the fact that one event can cause different modifications of the system, one has to label the transitions. For example, in a school, the beginning and the end of classes are signalled, or labelled, by the ringing of a bell. One possible way of investigating the functioning of a system is to examine the sequences of labels corresponding to the firing sequences of transitions of the Petri net that models the system. Several classes of languages have been defined and studied: languages of sequences of firing of transitions when one event or signal causes one action, and languages of words labelling firing sequences of transitions when one event can cause several actions.

A Realistic Model of Real-Time Systems for Efficient Scheduling

In order to take semantical aspects into account for the scheduling problem and obtain scheduling results for a wide class of systems, we extend the Petri net scheduling approach for real-time systems. Our study focuses on tasks with conditional statements. Classical approaches consider only the worst case execution time that occurs in the different branches of conditional statements and don’t take the semantic of the tasks into account. We show that this pessimistic model can lead to wrong conclusions for scheduling. We extend the task model with conditional statements, and the notion of schedule is replaced by the notion of scheduling tree. We propose then a model approach using Petri nets in order to explicitly take conditional instructions and their semantics into account.

Tree scheduling versus sequential scheduling

We present a new approach of validation for critical real-time applications: the tree based approach. This approach explicitly takes the conditional statements and the semantics contained in the tests into account. We substitute sequential schedules by scheduling trees. We thus add new scheduling possibilities to those provided by the sequential approach. We then give conditions which make the two approaches equivalent (i.e. give same results for schedulability).

A Multi-Level Framework for Validation of Ontology-Driven and Community-Based Web Services Composition

This paper proposes an Ontology-driven and Community-based Web Services OCWS framework which aims at automating discovery, composition and execution of web services. The purpose is to validate and to execute a users request built from the composition of a set of OCWS descriptions and a set of user constraints. The defined framework separates clearly the OCWS external descriptions from internal realistic implementations of e-services. It identifies three levels: the knowledge level, the community level and e-services level and uses different participant agents deployed in a distributed architecture. First, the reasoner agent uses a description logic extended for actions in order to reason about: i consistency of the pre-conditions and post-conditions of OCWS descriptions and the user constraints with ontologies semantics, ii consistency of the workflow matching assertions and the execution dependency graph. Then the execution plan model is generated automatically to be run by the composer agents using the dynamic execution plan algorithm DEPA, according to the workflow matching and the established execution order. The community composer agents invoke the appropriate e-services and ensure that the non functional constraints are satisfied. DEPA algorithm works dynamically without a priori information about e-services states and has interesting properties such as taking into account the non-determinism of e-services and reducing the search space.

Proceedings Fourth Interaction and Concurrency Experience, ICE 2011, Reykjavik, Iceland, 9th June 2011.

This paper presents an investigation of the notion of reaction time in some synchronous systems. A state-based description of such systems is given, and the reaction time of such systems under some classic composition primitives is studied. Reaction time is shown to be non-compositional in general. Possible solutions are proposed, and applications to verification are discussed. This framework is illustrated by some examples issued from studies on real-time embedded systems.