Current technology for locating and validating composite services are not sufficient due to the following reasons. • Current frameworks do not have the capacity to create complete service descriptions since they do not model all the functional aspects together (i.e. the purpose of a service, state transitions, data transformations). Those that deal with behavioural descriptions are unable to model the ordering constraints between concurrent interactions completely since they do not consider the time taken by interactions. Furthermore, there is no mechanism to assess the correctness of a functional description. • Existing semantic-based matching techniques cannot locate services that conform to global constraints. Semantic-based techniques use ontological relationships to perform mappings between the terms in service descriptions and user requests. Therefore, unlike techniques that perform either direct string matching or schema matching, semantic-based approaches can match descriptions created with different terminologies and achieve a higher recall. Global constraints relate to restrictions on values of two or more attributes of multiple constituent services. • Current techniques that generate and validate global communication models of composite services yield inaccurate results (i.e. detect phantom deadlocks or ignore actual deadlocks) since they either (i) do not support all types of interactions (i.e. only send and receive, not service and invoke) or (ii) do not consider the time taken by interactions. This thesis presents novel ideas to deal with the stated limitations. First, we propose two formalisms (WS-ALUE and WS-π-calculus) for creating functional and behavioural descriptions respectively. WS-ALUE extends the Description Logic language ALUE with some new predicates and models all the functional aspects together. WS-π-calculus extends π-calculus with Interval Time Logic (ITL) axioms. ITL axioms accurately model temporal relationships between concurrent interactions. A technique comparing a WS-π-calculus description of a service against its WS-ALUE description is introduced to detect any errors that are not equally reflected in both descriptions. We propose novel semantic-based matching techniques to locate composite services that conform to global constraints. These constraints are of two types: strictly dependent or independent. A constraint is of the former type if the values that should be assigned to all the remaining restricted attributes can be uniquely determined once a value is assigned to one. Any global constraint that is not strictly dependent is independent. A complete and correct technique that locates services that conform to strictly dependent constraints in polynomial time, is defined using a three-dimensional data cube. The proposed approach that deals with independent constraints is correct, but not complete, and is a heuristic approach. It incorporates user defined objective functions, greedy algorithms and domain rules to locate conforming services. We propose a new approach to generate global communication models (of composite services) that are free of deadlocks and synchronisation conflicts. This approach is an extension of a transitive temporal reasoning mechanism.

Après un état de l'art complet nous avons détaillé notre approche de services sémantiques utilisant une méthode innovante pour la composition de services, les patrons de chaînes de traitements. Notre approche est constituée d'un annuaire sémantique hybride proposant différents niveaux de correspondances entre services, de règles de compositions automatiques dans le cas où une demande de services n'est pas fructueuse et enfin d'un moteur d'exécution supportant la résolution et la composition dynamique de services. Par la suite nous avons défini une méthodologie innovante, basée sur l'utilisation de taxonomies de services permettant de peupler rapidement un annuaire sémantique de services. Pour ce faire, nous avons mis au point une ontologie de haut niveau qui permet de lier la classe d'une taxonomie avec un service abstrait annoté sémantiquement. Enfin, nous avons évalué notre prototype à partir des chaînes de traitement mises en place dans les plate-formes déployées par Cassidian. / After a complete state of the art we detailed our semantic services approach which uses an innovative method for services composition: processing chains patterns. Our approach is composed on an hybrid semantic servicers registry which propose different levels of matching between services, some composition rules when the matching phase failde and an execution engine which is able to dynamically resolve and com^pose services. In order to solve the service regitry population issue, we have designed an upper ontology, which enables links between a service taxonomy class with a semantically annotated abstract service. Finally, we have evaluated our prototype using real processing chains used by Cassidian platforms.

According to the service orientation design paradigm there are business (BS) and electronic services (ES). BS encapsulate business concerns. ES encapsulate computing systems, information systems and software applications. In environments with a high number of BS and ES the decision on which ES provides the most suitable support for a certain BS is not a trivial task. The objective of the thesis is to provide models, methods, and techniques for preselection of ES for a given BS. Preselection is about reducing the large amount of ES to a significant smaller amount under the consideration of a particular BS.

According to the service orientation design paradigm there are business (BS) and electronic services (ES). BS encapsulate business concerns. ES encapsulate computing systems, information systems and software applications. In environments with a high number of BS and ES the decision on which ES provides the most suitable support for a certain BS is not a trivial task. The objective of the thesis is to provide models, methods, and techniques for preselection of ES for a given BS. Preselection is about reducing the large amount of ES to a significant smaller amount under the consideration of a particular BS.

Les Architectures Orientées Services (SOA) se sont progressivement imposées comme outil incontournable dans les échanges inter-entreprises grâce à leurs potentiels stratégiques et technologiques. Leurs mise en oeuvre est concrétisée à travers les services Web dont l'un des principaux atouts est leur composabilité. Avec l'émergence du Web sémantique la découverte et la composition de services Web sémantiques constituent un réel défi. Le processus de découverte s'appui généralement sur les registres traditionnels offrant des descriptions syntaxiques regroupés statiquement, ce qui pose un problème lié à l'hétérogénéité des descriptions syntaxiques et à la rigidité de la classification. Le processus de composition dépend à son tour de la qualité de l'appariement des services. Nous proposons dans cette thèse une architecture d'un framework qui couvre toutes les phases du processus de composition. Ensuite, nous proposons une mesure de similarité sémantique pour un appariement entre les descriptions des services Web. Le processus de découverte de services Web s'appuie sur la similarité entre les services, le formalisme d'Analyse de Concepts Formels et l'organisation des services en treillis. La composition ensuite repose sur l'établissement de services composites cohérents et pertinaents pour la fonctionnalité espérée. Les points forts de cette architecture sont l'adaptation et l'intégration des technologies sémantiques, le calcul de similarité sémantique et l'utilisation de cette similarité sémantique et du formalisme FCA afin d'optimiser le processus de composition. / Service Oriented Architectures (SOA) have been progressively confirmed as an essential tool in inter-companies exchanges thanks to their strategic and technological potential. Their implementation is realised through Web services. One of the main assets of services is their compostability. With the emergence of the semantic Web, the discovery and composition of semantic Web services become a real challenge. The discovery process is generally based on traditional registries with syntactic descriptions where services are statically grouped. This poses a problem related to the heterogeneity of syntactic descriptions and the rigidity of the classification. The composition process depends on the Web service matching quality processed in the discovery phase. We propose in this dissertation an architecture of a framework that covers all the phases of the composition process. Then, we propose a semantic similarity measure Web services. The Web services discovery process relies on the proposed similarity measure, the formal concept analysis (FCA) formalism, and the organisation of lattice services. The composition is then based on the establishment of coherent and relevant composite services for the expected functionality. The main strengths of this architecture are the adaptation and integration of semantic technologies, the calculation of semantic similarity and the use of this semantic similarity and the FCA formalism in order to optimise the composition process.