Areski Flissi

Deploying on the Grid with DeployWare

In this paper, we present DeployWare to address the deployment of distributed and heterogeneous software systems on large scale infrastructures such as grids. Deployment of software systems on grids raises many challenges like: 1) the complexity to take into account orchestration of all the deployment tasks and management of software dependencies; 2) the heterogeneity of both physical infrastructures and software composing the system to deploy; 3) the validation to early detect errors before concrete deployments; and 4) scalability to tackle thousands of nodes. To address these challenges, DeployWare provides a metamodel that abstracts concepts of the deployment, a virtual machine that executes deployment processes on grids from DeployWare descriptions, and a graphical console that allows to manage deployed systems, at runtime. To validate our approach, we have experimented DeployWare with a lot of software technologies, such as CORBA and SOA-based systems, on one thousand of nodes of Grid5000, the french experimental grid infrastructure.

Norine, the knowledgebase dedicated to non-ribosomal peptides, is now open to crowdsourcing

Since its creation in 2006, Norine remains the unique knowledgebase dedicated to non-ribosomal peptides (NRPs). These secondary metabolites, produced by bacteria and fungi, harbor diverse interesting biological activities (such as antibiotic, antitumor, siderophore or surfactant) directly related to the diversity of their structures. The Norine team goal is to collect the NRPs and provide tools to analyze them efficiently. We have developed a user-friendly interface and dedicated tools to provide a complete bioinformatics platform. The knowledgebase gathers abundant and valuable annotations on more than 1100 NRPs. To increase the quantity of described NRPs and improve the quality of associated annotations, we are now opening Norine to crowdsourcing. We believe that contributors from the scientific community are the best experts to annotate the NRPs they work on. We have developed MyNorine to facilitate the submission of new NRPs or modifications of stored ones. This article presents MyNorine and other novelties of Norine interface released since the first publication. Norine is freely accessible from the following URL: http://bioinfo.lifl.fr/NRP.

Une infrastructure à composants pour des applications ubiquitaires

Multiplication of mobile devices and generalized use of wireless networks imply changes on software applications design and execution. This new challenge is now called ubiquitous computing. In this paper, we present a software infrastructure dedicated to ubiquitous contextual services. They are designed as assemblies of distributed software components and are dynamically discovered according to end-users physical location and device capabilities and automatically deployed on it. We have implemented this software infrastructure and an example of service on top of the OMG CORBA Components Model and the OpenCCM open source platform.

Une démarche dirigée par les modèles pour construire les machines de déploiement des intergiciels à composants

Les intergiciels a composants permettent dautomatiser le deploiement des applications. Cette fonction, appelee machine de deploiement, instancie les applications a partir de leurs descriptions architecturales. Malheureusement, a lheure actuelle, chaque intergiciel met en œuvre sa propre machine de deploiement. Ainsi, aucune capitalisation nest proposee aussi bien sur le plan conceptuel quoperationnel. Afin de favoriser cette capitalisation, cet article propose une demarche dirigee par les modeles (a la OMG MDA) pour la construction des machines de deploiement des intergiciels a composants. Cette demarche introduit un profil UML de workflow permettant de definir des modeles de deploiement independants des intergiciels a composants cibles. Un tel modele est ensuite raffine pour chaque intergiciel cible puis le modele obtenu est projete, via des transformations, vers diverses plates-formes dexecution. Les differents modeles et transformations de cette demarche sont illustres sur la construction dune machine de deploiement de composants CORBA mise en œuvre avec le modele de composants Fractal.

Bioinformatics tools for the discovery of new lipopeptides with biocontrol applications

As conventional or chemical pesticides have negative impact on environment and health of both farmer and consumers, it becomes relevant to develop alternative solutions to limit their use. In this context, innovative strategies to accelerate the development of biocontrol agents are welcome. For a decade of years, it has been demonstrated that lipopeptides are very efficient weapons against fungi responsible for crop diseases. Lipopeptides are secondary metabolites, produced by many microorganisms including beneficial rhizobacteria. The lipopeptide biosynthetic pathways include nonribosomal peptide synthetases. These modular enzymatic complexes work as assembly lines to build the peptides step by step, leading to the production of original peptide compounds with specific features as the presence of non proteinogenic monomers and cyclic and branched structures. In this paper, Florine and Norine bioinformatics tools, especially dedicated to non-ribosomal synthetases and their products are presented. Their use is mainly focused on the discovery of lipopeptides produced by Bacillus or Pseudomonas because they seem to represent a versatile reservoir of active secondary metabolites with promising activities for applications in phytosanitary area.

Une démarche orientée modèle pour déployer des systèmes logiciels répartis

Le deploiement de systemes distribues met en jeu de nombreuses technologies heterogenes. Ladministrateur systeme doit 1) maitriser le deploiement de chaque logiciel, 2) ladapter aux proprietes des machines et 3) lexecuter en respectant lordre de ses dependances. Ces tâches sont fortement propices aux erreurs. Dans cet article, nous presentons DeployWare, une approche a base de modeles pour le deploiement de systemes distribues complexes. Cette approche repose sur un meta-modele en deux parties. La premiere permet de decrire les proprietes, les dependances et actions a effectuer pour deployer des logiciels. La seconde permet dassembler des instances de logiciels. Ces deux parties sont realisees de maniere a rendre possible la verification comportementale des procedures de deploiement et des systemes. Les modeles DeployWare sont projetes vers une plate-forme dexecution a base de composants qui gere automatiquement lheterogeneite des machines et lorchestration des dependances.