Main / Documentation
Rechercher:

OSGi™ Users Group - France

Documentation

Cette rubrique liste des ressources documentaires en français sur la technologie OSGi™.

Livres Blanc/Whitepapers

  • David Menga et Michel D'Hooge, "OSGi™ et UPnP™ : Des technologies complémentaires", Livre Blanc, Juin 2005, PDF

Livres/Books

  • Didier Donsez, "Chapitre 7 : La plate-forme dynamique de services OSGi", 36 pages, dans le livre en ligne "Intergiciels et Construction d'Applications Réparties", Coordinateurs: Sacha Krakowiak, Jacques Mossières, HTML ,PDF

Articles/Papers

  • Atelier de travail OSGi @ UbiMob 2006, 5 septembre 2006, Actes
  • Yvan Royon and Stéphane Frénot, Multiservice Home Gateways: Business Model, Execution Environment, Management Infrastructure, IEEE Communications Magazine, vol. 45, no. 10, October 2007 (doc)
The home gateway market is undergoing deep changes. On one side, home networks are evolving, getting dynamic and federating more and more devices. On the WAN side, new actors appear, such as multimedia content providers. From both sides emerge new features and new management needs. In this article we highlight challenges and benefits of moving more intelligence to the home gateway, making it more than a simple interconnection device. We argue that we need a full-fledged execution environment on gateways to address the evolution of business models. We present this evolution and summarize existing solutions for execution environments and management for home gateways. Then we propose two improvements that reflect the new requirements. The first improvement is a high-level virtualization of service gateways, and the second one recommends an end-to-end dynamic multiprovider management system

Thèses de Doctorat/PhD Dissertations

Les nouveaux espaces de vie, n´es de la prolifération d’objets communicants, rendent la problématique de l’interaction de plus en plus complexe à la fois pour l’utilisateur et le concepteur. Le champ de recherche de l’informatique ambiante tend à apporter des solutions pour ces espaces mais la problématique de l’interaction reste entière. Dans cette thèse, nous abordons le problème de l’interaction homme-environnement (IHE) par une approche transverse autour de deux communautés de recherche, l’IHM et les approches logicielles à composant et service. Nous présentons un modèle original d’interaction homme-environnement, sous un aspect conceptuel et d’implémentation. Ce modèle a donné lieu à une plate-forme d’´etude IHE qui permet de construire des applications dédiées à l’interaction homme- environnement, en prenant en compte des contextes d’utilisation très divers.
Les serveurs d'applications sont largement utilisés et éprouvés comme environnement hébergement et d'exécution d'applications Internet. Ils permettent d'assurer la fiabilité, les performances et la sécurité de ces applications. Dans le contexte de ce travail, les serveurs d'applications Java EE, la construction des applications et la gestion de leurs cycles de vie est spécifié et offre de nombreuses possibilités. Les services techniques offerts par le serveur et utilisé par les applications ne propose pas les même facilités d'assemblage et de gestion et souffrent d'un manque de flexibilité dans les implantations open source actuelle. Ils sont de ce fait peut adaptés aux nouvelles applications Internet. Celles-ci demandent de nouveaux types de services techniques (capteurs, multimÈdia) et étendent également les supports matériels des d'exécution des serveurs à des environnements plus contraints. Cette thèse propose une architecture permettant d'obtenir un serveur d'applications personnalisable (démarrage, arrêt et configuration des services techniques au besoin et à chaud), extensible (facilite l'ajout de nouveaux services utilisables par les applications) et taillé sur mesure pour les applications (déploiement et retrait des services techniques nécessaires et de leurs dépendances). L'architecture proposée s'appuie sur une architecture orientée services, qui permet d'obtenir le dynamisme désiré sans perdre les propriétés inhérentes des serveurs d'applications. Cette approche est validée par une expérimentation sur un des serveurs open source actuelle, JOnAS. Elle permet de mettre en avant les nouvelles propriétés obtenues par la mise en oeuvre de l'architecture proposée.
This work aims to propose an adaptable, extensible and dynamic architecture for applications servers' core by relying on service-oriented architectures. Applications servers are widely used and tested as hosting and execution environments for Internet applications. They guarantee the reliability, the performances and the security of those applications. In our context, the Java EE applications servers, building applications and managing their lifecycle is specified and offers many capabilities. Technical services offered by the server do not have the same assembly and management properties and suffer from a lack of flexibility in current open source implementations. For those reasons, they do not fit current Internet applications' needs that require new kind of technical services (sensors, multimedia) and also extend execution supports to more constraints environments. This thesis proposes an architecture that permits to obtain a customizable (on-demand and hot start, stop and configuration of technical services), extensible (simplify the integration of new services useable by the applications) and scalable (deployment and retreat of required services and their dependencies) applications server. The proposed architecture relies on a service-oriented architecture that brings the required dynamism without loosing applications servers' inherent properties. This approach is validated by experimentation on an existing open source server, JOnAS. It permits to highlight the new properties obtained by the implementation of the proposed architecture.
  • Yvan Royon, "Environnements d’exécution pour passerelles domestiques", Thèse de Doctorat d'Informatique de l'Institut National des Sciences Appliquées de Lyon, Décembre 2007.
The home gateway market evolves towards new business models. Today’s model, called triple play, brings Internet connectivity to our homes, along with two additional services : telephony and video over IP. These three services are managed by a single business entity : the access provider. Constructors and operators propose a replacement model, called multi-play, which opens the market to multiple voice and video providers. This allows end-users to access more content and to choose between competing service plans. A third business model then becomes possible. Since the market is open to multiple business actors, and since home gateways become more and more powerful, why should we limit services to voice and video ? An increasing number of other services exist, such as health care, support for the disabled, home automation and telesecurity. The notion of service must be broadened, and the home gateway must be enhanced to be able to host any kind of service. We call this model multi-service. Technical novelties with multi-service home gateways are that they support deploying, executing and managing several software modules that come from different providers. This translates into specific needs in terms of execution isolation, remote management, deployment infrastructure, and programming model. Execution isolation is a compromise between the level of isolation and the impact on perfor- mance. Since home gateways have limited hardware resources, performance and complexity are key factors. They need a level of isolation that allows to separate software modules based on their providers. When choosing how this separation is enforced, we create variants in the business model ; for instance, some modules, such as a video codec or a web server, may or may not be shared among service providers. Each provider manages his own services. Similarly, variants are possible : one is that each provider chooses his own management technology, another is that the access provider dictates one. The deployment infrastructure must scale, both service-wise and user-wise, and propose update mechanisms. Variants lie in life cycle management granularity, e.g. pausing services may or may not be possible. Lastly, the programming model impacts decoupling between modules, their reusability, and their time-to-develop. To implement these four families of needs on resource-constrained targets, we focused on two execution environments, which allow different aforementioned variants : Java/OSGi and C/Linux. OSGi boasts a clean, service-oriented programming model, and already manages deploy- ment and life cycle of modules. However, it lacks multi-user features and the related isolation mechanisms. GNU/Linux offers many tools for isolation and module deployment. However, it needs a tight integration and a unified management of such tools. Therefore, we propose to fill these lacking features on these two environments, so they can conform to the multi-service model.

Cours/Courses

Blogs

Divers/Misc

  • Rien pour l'instant