Yves Raynaud

A KQML-CORBA based architecture for intelligent agents communication in cooperative service and network management

Networks and Services management solutions are now moving from centralised platforms to distributed approaches. In order to reach a higher level of scalability and flexibility, an attractive solution can be based on cooperative intelligent agents. This raises new requirements for communication between management entities, as it implies not only exchange of pure management information but also exchange of cooperation messages or code.

Methodology of real-time system design using multiprocessors

Abstract We present a methodology devoted to real-time system design using multiprocessors. The methodology is proposed according to five aspects: environment, function, behaviour, performance and hardware. The proposed methodology is based on the decomposition of the system development life cycle into six stages: definition of requirements, analysis of the system environment, refinement of the system, analysis of the system behaviour, evaluation of performance and implementation. This methodology is validated by an application which aims to generate transfer frame telemetry that complies with CCSDS recommendations.

A multi-agent self-adaptative management framework

The self-management vision, which has raised much interest in recent years, is only an operational aim that does not provide any solution in its definition on how to attain this management autonomy. The general contribution of this article suggests concepts, methods and mechanisms to conceive a self-adaptative management framework, thus participating in this self-management vision. This framework is composed of organizational aspects based on multi-agent concepts, informational aspects based on different management ontologies and cooperative aspects based on specific agent behaviours which automate the accomplishment of management actions on the overall distributed environment. We have also implemented a prototype of this work to validate the feasibility of our approach and confirm the self-adaptation capacities of the system.

Distributed Systems Design: An Integrated Software Tool for Performance Analysis

In an operational context, designing sophisticated data processing systems involves a lot of difficulties such as choosing the good hardware components, dispatching the tasks on the various computers... But, above all, one of the most critical problem is to predict system’s performances; that means to evaluate, before construction, the architecture behavior under processes loading.

A generic conceptual framework for self-managed environments

The high complexity of existing managed environments has led to the need to increase the autonomy of network and service management solutions. This short article presents a generic multi-agent organizational framework dedicated to the management domain. This conceptual framework constitutes a first step towards the design of self-managed environments.

An active CIM dependency pattern for a consistent service state monitoring

Before significant effects may be felt in service utilization by the user, detection of QoS degradation is a desirable requirement of service management. Management approaches define management information for each managed resource in a distributed environment, but a problem occurring on a resource denies access to its management information. A service has dependency relationships with all the components in the system, software or hardware. The paper focuses on the service level that is influenced by the subjacent levels, network and system, and elements, and investigates how to express these influences known by management experts in order to automate their processing. Our solution is to focus on the informational model independently from the other ones. The network and service management (NSM) domain is now on the way towards a unified and technology independent information model thanks to the CIM/DMTF specification of the WBEM initiative. This is to ensure both interoperability and the federation of management platforms from an informational point of view in the sense that the communication point of view is abstracted. This tendency opens new management modeling capabilities that the paper develops for consistent service monitoring.

Flow control in ATM based local area networks

ATM networks are expected to provide the information transport for a rich mixture of services and applications, associated with which will be a broad spectrum of traffic types and transport performance needs [DePrycker, 1993] [Handel, 1991] .

EXTENSION DU FORMALISME ESTELLE À ĽÉVALUATION DE PERFORMANCES: ÉTUDE DES PERFORMANCES DU PROTOCOLE SSCOP POUR LES RÉSEAUX À HAUT DÉBIT ATM

Cet article presente une approche permettant ďobtenir des mesures de performance a partir ďune technique de description formelle, Estelle, normalisee par ľiso. he but de cette approche est de ne pas modifier la syntaxe ďEstelle afin de preserver les avantages fournis par la normalisation. Pour cela, des annotations relatives a ľaspect quantitatif sont ajoutaes a la spacification. Cette spacification annotae est ensuite traduite dans un langage de simulation, Modsim, afin de raaliser une simulation fournissant les rasultats de performance. Cette approche est illustrae en fin ďarticle par ľavaluation du protocolesscop de la coucheaal, les rasultats obtenus atant conformes a ceux obtenus a partir ďautres mathodes ďavaluations.This article presents an approach used to obtain performance measures from anISO-standardized formal description technique, Estelle. The underlying aim of this approach is not to modify Estelle syntax so as to maintain the advantages provided by conformance to theISO standard. To achieve this, annotations relating to the quantitative aspect have been added to the specification. This annotated specification is then tranlated into a simulation language, Modsim, to achieve a simulation capable of supplying performance results. At the end of the paper, this approach is illustrated by evaluating thesscop protocol of theaal layer, the results obtained being in compliance with those obtained using other evaluation methods.RésuméCet article présente une approche permettant ďobtenir des mesures de performance à partir ďune technique de description formelle, Estelle, normalisée par ľiso. he but de cette approche est de ne pas modifier la syntaxe ďEstelle afin de préserver les avantages fournis par la normalisation. Pour cela, des annotations relatives à ľaspect quantitatif sont ajoutäes à la späcification. Cette späcification annotäe est ensuite traduite dans un langage de simulation, Modsim, afin de räaliser une simulation fournissant les räsultats de performance. Cette approche est illusträe en fin ďarticle par ľävaluation du protocolesscop de la coucheaal, les räsultats obtenus ätant conformes à ceux obtenus à partir ďautres mäthodes ďävaluations.