Section: Partnerships and Cooperations

International Initiatives

INRIA Associate Teams

MOCAA

• Title: Models Composition, Aspects and Analysis

• INRIA principal investigator: Benoît Baudry

• International Partner:

  • Institution: Colorado State University (United States)

  • Laboratory: Colorado State University, Software Assurance Lab

• Duration: 2006 - 2011

• See also: http://www.irisa.fr/triskell/matt/

• Computer-based systems have been growing in complexity at an exponential rate (roughly 10 fold increase every ten years) for more than 40 years. Like in other sciences, people have been relying more and more on modeling to try to master this complexity. Modeling, in the broadest sense, is indeed the cost-effective use of a simplified representation of an aspect of the world for a specific purpose. Because in software a model has the same nature as the thing it models, this opens the possibility to automatically derive software (and other artifacts such as test cases, performance profiles, or documentation) from its model. This property is well known from any compiler writer (and others), but it was recently be made quite popular with initiatives such as Model Integrated Computing (MIC) or OMG's Unified Modeling Language (UML) and Model Driven Architecture (MDA), globally known as Model Driven Development (MDD). In this context, models are formally described and can be automatically manipulated for refinement, composition, test case generation, documentation; All those operations are model transformations. This collaboration aims at better understanding how classical software engineering practices (design patterns, validation, methods, IDEs) can be adapted to develop model transformations. Clément Guy worked in collaboration with Prof. Robert B. France (from the software engineering domain), as well as with Prof. Sanjay Rajopadhye (from the optimizing compiler domain) to cross-fertilize both domains. In particular, he was studying the possibility to extend existing model typing to fit the needs of reusing model transformations.

INRIA-CONFAP

• Title: Software Testing for Cloud Computing (TAAS)

• International Partner:

  • Universidade Federal do Paraná.

• Principal investigator: Gerson Sunyé

• Duration: 2011 - 2012

• Cloud computing is consolidating as an important paradigm for information technology to provide resources and Internet-based services. In clouds, a large amount of resources (e.g., memory, CPU, disk) is shared between several storage and processing machines or nodes, providing scalable environments.However, building reliable applications for clouds is a difficult task, because developers must face several non-trivial issues, such as: large-scale distribution, fault tolerance, massive data processing, hardware and software heterogeneity. In general, a cloud involves clusters and grids of nodes distributed over the Internet, where each new node shares its resources with the rest of the system, ensuring the scalability of clouds.

  Since cloud applications are becoming ubiquitous in society's critical activities (health, economics, governments, etc.), they must ensure that the eventual failures of nodes do not affect the applications running on it. Large-scale distribution increases risks related to the loss of data because of nodes that fail, delay in computation times because of unreliable distribution strategy, etc. and several algorithms are proposed to increase their tolerance to faults. Thus, quality factors such as: reliability, robustness, availability and performance are essential. The main practice to ensure these factors, as well as the correctness, is the systematic use of testing during the different stages of development.

  In this project, we propose to adapt and improve the testing architectures previously developed. More precisely, we propose to adapt the existing architecture for cloud environments, to define a testing language that supports the specification of large-scale tests as a whole and to provide both, a generator of test data and a fault injector, to reproduce real cloud environments.

INRIA International Partners

Following the Diva STREP project, we keep an active collaboration with the SINTEF institute. François Fouquet visited SINTEF for 8 weeks. During this visit, we combined the results of Kevoree and the result of the Moderate from SINTEF project to provide a dynamic component model for a micro-controllers based Internet of Things. Indeed, as the Internet of Things promises new ways for humans to interact with computing systems by seamlessly integrating resource constrained devices and traditional computing environment. These new computing environments are highly volatile and force applications to embed self-adaptive behaviors. The contribution of this collaboration is μ-Kevoree: a plain C implementation of the Kevoree runtime which can be deployed on poor in resources micro-controllers. Evaluation of memory usage, reliability and performance shows that μ-Kevoree is a viable solution with strong benefits over adaptation through dynamic firmware upgrades.

Visits of International Scientists

Internships

• Hanen HAOUAS (from Mar 2011 until Aug 2011)

  • Subject: Autonomously Optimizing Service-Based Application Dependability in Smart Building

  • Institution: Ecole Nationale des Sciences de l'Informatique (Tunisia)

• Wuliang Sun

  • Subject: Discovering the boundaries of a Modelling Space

  • Institution: Colorado State University (United States)

Participation In International Programs

Thanks to the MoCAA Equipe associée, Clément Guy realized a three-month stay in 2011 at Colorado State University (USA). He worked in collaboration with Prof. Robert B. France (from the software engineering domain), as well as with Prof. Sanjay Rajopadhye (from the optimizing compiler domain) to cross-fertilize both domains. In particular, he was studying the possibility to extend existing model typing to fit the needs of reusing model transformations.