Section: Contracts and Grants with Industry
Sardes is a member of the CNRS research network GDR 725 ASR “Architecture, Systme et RŽseau”. See http://asr.cnrs.fr/ .
Project SelfWare (ANR)
The goal of project Selfware is to provide a framework for autonomous management of large scale distributed applications including legacy software systems. The expected results are reducing operator mistakes (a major cause of failure for large Internet services), saving hardware ressources, and minimizing administrator efforts. The main challenge is to be able to manage any arbitrary complex legacy system with minimum effort. This framework will be evaluated for two actual middleware infrastructures: (i) a clustered J2EE application Server ( JOnAS ) and (ii) a Message Oriented Middleware ( Joram ). Experimental scenarios will illustrate self-optimizing and self-healing.
The project partners are INRIA Grenoble-Rhône-Alpes; France Telecom Orange Labs; Ecole des Mines de Nantes; Bull; IRIT/ENSEEIHT, Toulouse.
The project runs from June 2006 to December 2008.
Project Modyfiable (ANR)
The goal of Modyfiable is to study the formal basis of “dynamic modularity”, i.e. aspects related to dynamic creation, deployment, linking, and update of modules and components, considered as first-class programming abstractions. In contrast to classical modularity, which only deals with modular program construction, dynamic modularity is concerned with the manipulation and control of modules and components, understood as units of modularity and deployment, during execution.
The project partners are: INRIA Grenoble-Rhône-Alpes ( Sardes project team); École Normale Supérieure de Lyon (Plume team).
The project runs from January 2006 to December 2008.
Project ScorWare (ANR)
The SCOrWare project aims at providing an open source implementation of the recent Service Component Architecture (SCA) specifications defined by the Open SOA collaboration, an industrial consortium in the domain of software engineering, including BEA Systems, IBM Corporation, IONA Technologies, Oracle, Red Hat, Rogue Wave Software, Siemens, Sun Microsystems, and Sybase. Briefly, SCA defines a new architecture and programming model for Service Oriented Applications (SOA) based on the component paradigm and supporting several service description languages like WSDL and Java interfaces, several programming languages such as Java, C++, and BPEL, several communication protocols between applications such as SOAP, CORBA, Java RMI, and JMS.
The project aims to build a configurable and manageable SCA platform built upon the Fractal component model, with advanced functionalities such as dynamic reconfiguration of SCA component assemblies, a binding factory supporting different communication protocols between SCA components, a transaction service, a semantic trading service of SCA components, a deployment engine of autonomous SCA architecture and graphical administration consoles. The project partners are INRIA; Artenum; EBM WebSourcing; Edifixio; INT, Evry; IRIT, Toulouse; Obeo; OpenWide.
The project runs from January 2007 to December 2008.
Project Flex-e-Ware (ANR)
The Flex-e-Ware project aims at defining software development tools and mechanisms dedicated to flexible, reconfigurable embedded systems. It aims to develop a common platform for the development of component-based embedded systems, based on the integration of the Fractal and Lightweight CCM component models.
The project partners are INRIA (Adam and Sardes project teams); CEA List; ENST, Paris; LIP6, Paris; France Telecom; Schneider Electric; STMicroelectronics; Teamlog; Trialog; Thales.
The project runs from January 2007 to December 2009.
Project Aravis (ANR-Minalogic)
The ARAVIS Project aims at addressing the challenges raised, both at the hardware and software levels, by the production of highly integrated multiprocessor systems on chip (MPSoCs) designed for demanding applications such as video encoding/decoding and software-defined radio communications. Due to the complexity of the manufacturing process, the latest generations of chips exhibit peculiar features that must be taken into account : (i) massively parrallel processing units, (ii) irregular behavior and aging of the processing units due to unavoidable defects of the manufacturing process. The ARAVIS project strives to provide a hardware and software platform suited to the adaptation requirements raised by the needs of such emerging hardware technologies and applications. The proposed approach encompasses three contributions: (i) a symmetric hardware architecture based on an asynchronous interconnect with integrated voltage/frequency scaling, (ii) a set of regulation algorithms based on control theory to optimize quality of service and energy consumption, (iii) a component-based runtime environment and related software tools to ease the dynamic management of applications and execution resources.
The project partners are STMicrolectronics, CEA-LETI, TIMA and INRIA (Necs and Sardes project teams). The project runs from October 2007 to September 2010.