Service-Oriented Architecture (SOA) represents an emerging architectural approach that is able to tackle challenges in the contemporary service-based economy, in which the global market revenues are shifting from the manufacture of traditional off-the-shelf products to the provision of diversified services that suffice for customers’ needs. In such a service-based economy, one can envisage an entirely “service-oriented” world, where a massive number of distributed services with different natures and capabilities are provided by various professionals around the world. Problems arise when business applications demand desirable services through different sources and providers that are appropriate for their own benefits and preferences. Therefore, it can be very challenging to design an SOA infrastructure that enables users to exploit this great level of service heterogeneity and quantity. One of the key issues in service-oriented architecture is to achieve efficient service discovery and loosely-coupled service distribution while maintaining a satisfactory degree of scalability, usability, and Web consistency. This thesis deals with SOA infrastructure-level design and implementation issues. It approaches this SOA infrastructure within the scope of Web services, which capture an important, and perhaps the best, ‘realisation’ of SOA. It investigates and formulates how public Web services distributed across the World Wide Web can be augmented by a software platform that enables scalable, user-centred,semantic-enabled, and integration-oriented service retrieval, selection, and matching. The primary goal of this thesis is thus to propose a conceptual framework of an enhanced SOA infrastructure with regard to service distribution and discovery. / It also aims to design and implement a platform (PWSP), by means of which a large number of public Web services on the Web can be distributed based on service demands, retrieved based on service descriptions, selected based on service qualities, and matched based on service messages in a user-centred, scalable, and Web-consistent manner without augmenting existing Web services standards.

Web services composition is emerging as a promising technology for the effective automation of application-to-application collaborations. The application integration problems have been subject of much research in the past years. However, with growth in importance of business process automation and highly dynamic nature of the Internet, this research has taken on a new significance and importance. Adequate solutions to this problem will be very important to make enterprise systems more flexible, robust and usable in the future. In this dissertation, we present a novel approach for the declarative definition and scalable orchestration of composite Web services in large, autonomous, heterogeneous, and dynamic environments. We first propose a composition model for composing Web services in a personalized and adaptive manner. We model composite Web services based on statecharts. To cater for large amounts of dynamic Web services, we use the concept of service community that groups services together and is responsible for the runtime selection of services against user's preferences. We use the concept of process schema that specific users can adjust with their personal preferences. A set of exception handling policies can be specified to proactively react to runtime exceptions. We then propose a tuple space based service orchestration model for distributed, self-managed composite services execution. We introduce the concept of execution controller that is associated with a service and is responsible for monitoring and controlling service executions. The knowledge required by a controller is statically extracted from the specification of personalized composite services. We also present techniques for robust Web services provisioning. The techniques presented in this dissertation are implemented in Self-Serv, a prototype that provides a set of tools for Web service composition and execution. Finally, we conduct an extensive usability and performance study of the proposed techniques. The experimental results reveal that our system i) provides an efficient support for specifying, deploying, and accessing composite services; ii) is more scalable and outperforms the centralized approach when the exchanged messages become bigger; and iii) is more robust and adaptive in highly dynamic environments.

No description available.

Reconfiguration dynamique des architectures orientées services

<p>För att dagens system och deras allt mer komplexa applikationer ska kunna integreras med</p><p>varandra krävs det att de kommunicerar via tjänster. Denna tjänsteorienterade integration</p><p>uppnås genom att man använder sig av Service Oriented Architecture (SOA) som bygger på</p><p>löst kopplade tjänster som kommunicerar med varandra på ett standardiserat sätt. En viktig</p><p>del i en integrationslösning är Enterprise Service Bus (ESB). I denna rapport kommer vi</p><p>förklara grunderna i tjänsteorienterad integration och sedan fördjupa oss i ESB. Då ESB är ett</p><p>luddigt begrepp ska vi på ett enkelt och lättbegripligt sätt ge vår syn på begreppet, samt dess</p><p>fördelar och nackdelar. Vi kommer även att ge marknadens syn på ESB genom en</p><p>enkätundersökning som innefattar både leverantörer, konsulter och kunder.</p> / <p>If today’s software systems and their complex applications shall be able to integrate with each</p><p>other, they have to communicate through services. This service oriented integration can be</p><p>accomplished by using Service Oriented Architecture (SOA) where all components are</p><p>loosely coupled and communicate in a standardized way. An important part when building an</p><p>integrated solution is the Enterprise Service Bus (ESB). In this report we will explain the</p><p>basics of SOA and take a more detailed look at the world of ESB. The concept of ESB is not</p><p>well defined and hence means different things to different people. We are going to present an</p><p>interpretation of the ESB and its benefits and disadvantages. To find out what the market</p><p>thinks about ESB we have been talking to producers, consultants and customers.</p>

L'intérêt va croissant parmi les communautés scientifiques pour le partage de données et d'applications qui facilitent les recherches et l'établissement de collaborations fructueuses. Les domaines interdisciplinaires tels que les neurosciences nécessitent particulièrement de disposer d'une puissance de calcul suffisante pour l'expérimentation à grande échelle. Malgré les progrès réalisés dans la mise en œuvre de telles infrastructures distribuées, de nombreux défis sur l'interopérabilité et le passage à l'échelle ne sont pas complètement résolus. L'évolution permanente des technologies, la complexité intrinsèque des environnements de production et leur faible fiabilité à l'exécution sont autant de facteurs pénalisants. Ce travail porte sur la modélisation et l'implantation d'un environnement orienté services qui permet l'exécution d'applications scientifiques sur des infrastructures de calcul distribué, exploitant leur capacité de calcul haut débit. Le modèle comprend une spécification de description d'interfaces en ligne de commande; un pont entre les architectures orientées services et le calcul globalisé; ainsi que l'utilisation efficace de ressources locales et distantes pour le passage à l'échelle. Une implantation de référence est réalisée pour démontrer la faisabilité de cette approche. Sa pertinence et illustrée dans le contexte de deux projets de recherche dirigés par des campagnes expérimentales de grande ampleur réalisées sur des ressources distribuées. L'environnement développé se substitue aux systèmes existants dont les préoccupations se concentrent souvent sur la seule exécution. Il permet la gestion de codes patrimoniaux en tant que services, prenant en compte leur cycle de vie entier. De plus, l'approche orientée services aide à la conception de flux de calcul scientifique qui sont utilisés en tant que moyen flexible pour décrire des applications composées de services multiples. L'approche proposée est évaluée à la fois qualitativement et quantitativement en utilisant des applications réelles en analyse de neuroimages. Les expériences qualitatives sont basées sur l'optimisation de la spécificité et la sensibilité des outils de segmentation du cerveau utilisés pour traiter des Image par Raisonnance Magnétique de patients atteints de sclérose en plaques. Les expériences quantitative traitent de l'accélération et de la latence mesurées pendant l'exécution d'études longitudinales portant sur la mesure d'atrophie cérébrale chez des patients affectés de la maladie d'Alzheimer.

Researchers in software engineering proposed design method for distributed applications to construct a set of communicating system components from a global behavior. The joint behaviors of these components must precisely satisfy the specified global behavior. The next concern is to transform the constructed models of these components into executable business processes by ensuring the exchange of asynchronous messages among the generated business processes. The introduction of Service-Oriented Architecture (SOA) has helped to achieve this goal. SOA provides high flexibility in composing loosely-integrated services that can be used among business domains to carry out business transactions; this composition is known as service orchestration. Moreover, SOA supports Model Driven Architecture (MDA) such that services modeled as UML Activity Diagrams (AD) can be transformed into a set of Business Execution Language (BPEL) processes. Many researchers discussed the transformation of UML AD and the Business Process Modeling Notation (BPMN) into BPEL. However, they did not discuss the practical limitations that some of these transformations impose. This thesis addresses the imitations of the transformation from UML AD to BPEL processes using the IBM Rational Software Architect (RSA). We showed here that the tool is unable to create the correct BPEL artifacts from UML AD components in certain cases, for instance when the behavior includes the alternative for receiving single or concurrent messages, a weak loop, or certain choice activities. Furthermore, we provided novel solutions to the transformations in these cases in order to facilitate the transformation from UML AD to BPEL.

Researchers in software engineering proposed design method for distributed applications to construct a set of communicating system components from a global behavior. The joint behaviors of these components must precisely satisfy the specified global behavior. The next concern is to transform the constructed models of these components into executable business processes by ensuring the exchange of asynchronous messages among the generated business processes. The introduction of Service-Oriented Architecture (SOA) has helped to achieve this goal. SOA provides high flexibility in composing loosely-integrated services that can be used among business domains to carry out business transactions; this composition is known as service orchestration. Moreover, SOA supports Model Driven Architecture (MDA) such that services modeled as UML Activity Diagrams (AD) can be transformed into a set of Business Execution Language (BPEL) processes. Many researchers discussed the transformation of UML AD and the Business Process Modeling Notation (BPMN) into BPEL. However, they did not discuss the practical limitations that some of these transformations impose. This thesis addresses the imitations of the transformation from UML AD to BPEL processes using the IBM Rational Software Architect (RSA). We showed here that the tool is unable to create the correct BPEL artifacts from UML AD components in certain cases, for instance when the behavior includes the alternative for receiving single or concurrent messages, a weak loop, or certain choice activities. Furthermore, we provided novel solutions to the transformations in these cases in order to facilitate the transformation from UML AD to BPEL.

This thesis presents a semantic-based modeling approach for describing Service-Oriented Architectures (SOA). Ontologies are utilized as a major representation mechanism for describing various elements available in the architecture. The methodology proposes an architecture specification mechanism to constuct a unified ontology that enables transition from design concerns to the modeling elements. A multi-level modeling is also achieved by employing Model-Driven Engineering (MDE) techniques to describe various models at different stages of the software architecture. This aims to organize service-oriented models within a number of architecture viewpoints in order to provide an architectural perspective for SOA. The use of ontologies for model specification also allows us to make use of ontology mapping to specify the transformation between different models. Additionally, we present a case study to demonstrate the proposed methodology on a real-world healthcare scenario.

This research combines the methodology proposed in Axiomatic Design Theory (ADT) with a service oriented decomposition approach for systematic development of Service Oriented Architecture compliant systems. A previous study had applied ADT to component oriented development where simultaneous specification and decomposition of models related to requirements, design, product domain, and components were supported. Recently, Web services have gained popularity and they became a more desired alternative to components. This research sets the foundation for service-oriented modeling and development with ADT support through enhancing the component oriented work conducted before. The goal is to be able to consider customer needs viewed in the domain context, together with the requirements and design so that efficient development can take place based on existing Web services. The system under development is viewed as a hierarchy of process models where leaf-level processes correspond to Web services.