Section: Partnerships and Cooperations

National Initiatives

Inria ADT GinFlow (2014-2016)

Participants : Christine Morin, Matthieu Simonin, Cédric Tedeschi.

The GinFlow technological development action funded by Inria targets the development of a fully-operational workflow management system based on the HOCL-TS software prototype developed during the PhD thesis of Hector Fernandez between 2009 and 2012. Also, it allows the integration of this software with the TIGRES workflow engine developed at the Lawrence Berkeley National Lab so as to make the workflows submitted using the TIGRES programming model run in a decentralized fashion. These developments led to the release of the GinFlow software and its deposit at the APP (Agence de Protection des Programmes).

Inria ADT SaaP (2016-2018)

Participant : Martin Quinson.

The SaaP technological development action (SimGrid As A Platform) funded by Inria targets the refactoring of SimGrid to make it ready to use in production and teaching contexts. Our ultimate goal is to sustain the development of the framework by involving 5 to 10 companies that are using it internally. Our target of the teaching context is thus an intermediate goal, as we think that the best solution to ensure the adoption of our tool by the industrial engineers is that they discovered the tool during their studies.

The technical actions envisioned for this ADT are the complete rearchitecturation of the software (to make it easier to script a new model within the tool kernel) and a reorganization of the interfaces (for a better integration in the Java and python language). This work is lead in collaboration with the whole SimGrid community, which provide valuable feedback.

Inria IPL Discovery (2015-2019)

Participants : Anne-Cécile Orgerie, Matthieu Simonin, Cédric Tedeschi.

The Inria IPL Discovery officially started in September 2015. It targets the design, development and deployment of a distributed Cloud infrastructure within the network's backbone. It will be based upon a set of building blocks whose design will take locality as a primary constraint, so as to minimize distant communications and consequently achieve better network traffic, partition management and improved availability.

Its developments are planned to get integrated within the OpenStack framework. An energy/cost benefit analysis of the fully distributed Discovery architecture will also be performed to show the energy efficiency of the chosen approach.

Inria IPL CityLab (2015-2018)

Participant : Christine Morin.

The Inria Project Lab (IPL) CityLab@Inria (https://citylab.inria.fr) studies ICT solutions toward smart cities that promote both social and environmental sustainability. A strong emphasis of the Lab is on the undertaking of a multi-disciplinary research program through the integration of relevant scientific and technology studies, from sensing up to analytics and advanced applications, so as to actually enact the foreseen smart city Systems of Systems. City-scale experiments of the proposed platforms and services are planned in cities in California and France, thereby learning lessons from diverse setups.

Myriads investigates advanced cloud solutions for the Future Internet, which are critical for the processing of urban data. It leverages its experience in cloud computing and Internet of services while expanding its research activities to the design and implementation of cloud services to support crowd-Xing applications and mobile social applications.

In 2016, Christine Morin was involved in the MOOC entitled Villes Intelligentes : défis technologiques et sociétaux (Smart cities : technological and social challenges) run on the FUN platform from January to March 2016. She prepared eight sequences on urban data management in clouds. In 2016, we also conducted a comparative experimental evaluation of data stream processing environments executed on clusters and in a cloud. We compared the performance and energy consumption of Heron, Storm and SparkStreaming frameworks.

Inria IPL Hac Specis (2016-2020)

Participants : Anne-Cécile Orgerie, Martin Quinson.

The goal of the HAC SPECIS (High-performance Application and Computers: Studying PErformance and Correctness In Simulation) project (http://hacspecis.gforge.inria.fr/) is to answer methodological needs of HPC application and runtime developers and to allow to study real HPC systems both from the correctness and performance point of view. To this end, we gather experts from the HPC, formal verification and performance evaluation community.

The Anh Pham started a PhD thesis in November 2016, co-advised by Thierry Jéron (team SUMO, formal methods) and Martin Quinson. The envisionned work will pursue the previous efforts to formally assess distributed applications within the SimGrid framework.

COSMIC PRE (2016 - 2018)

Participants : Benjamin Camus, Anne-Cécile Orgerie.

The distributed nature of Cloud infrastructures involves that their components are spread across wide areas, interconnected through different networks, and powered by diverse energy sources and providers, making overall energy monitoring and optimization challenging. The COSMIC project aims at taking advantage of the opportunity brought by the Smart Grids to exploit renewable energy availability and to optimize energy management in distributed Clouds. This PRE, led by Anne-Cécile Orgerie also involves Fanny Dufossé from Dolphin team (Inria Lille) and Benjamin Camus, who has started a 18 months post-doc in October 2016 in the context of this project.

MIHMES ANR Investissements d'Avenir (2012 - 2018)

Participants : Yvon Jégou, Christine Morin, Manh Linh Pham, Nikos Parlavantzas.

The MIMHES project (http://www.inra.fr/mihmes) led by INRA/BioEpAR aims at producing scientific knowledge and methods for the management of endemic infectious animal diseases and veterinary public health risks. Myriads team provides software tools to efficiently manage and ease the use of a distributed computing infrastructure for the execution of different simulation applications.

In 2016, we further developed a distributed framework which allows to exploit multiple compute servers in parallel. Parallelism is exploited both at server level using OpenMP and at data-center level using this framework. To facilitate the deployment of the workloads on heterogeneous environments, this framework limits the requirements concerning the server configurations. They need not share any file system, the workloads can be programmed in differing programming language. These servers need only the capability to communicate through the network. The system allows to dynamically add and stop servers. To some extend, it is tolerant to server failures. The framework had being repackaged to facilitate its reuse for new workloads. We also worked on the automated deployment of the framework on top of one or multiple IaaS clouds.

PIA ELCI (2015-2018)

Participant : Anne-Cécile Orgerie.

The PIA ELCI project deals with software environment for computation-intensive applications. It is leaded by BULL. In the context of this project, we collaborate with ROMA and Avalon teams from Lyon: we co-supervise a PhD student (Issam Rais) funded by this project with these teams on multicriteria scheduling for large-scale HPC environments.

CNRS PEPS EcoSmart (2016)

Participant : Anne-Cécile Orgerie.

Smart Grids are connected to telecommunication networks and can thus optimize the production, distribution and consumption of electricity. Virtualized distributed systems (Clouds) are the major players in providing services over the Internet. The success of these on-demand services makes the energy consumption of these systems worrying. This project aims to optimize the energy consumption of these large consumers, namely virtualized distributed Clouds consisting of computing, storage and communication resources. The objective is to exploit the capabilities offered by smart grids to control the consumption of these systems and be able to influence it according to the availability or the nature of the electricity used.

CNRS GDS EcoInfo

Participant : Anne-Cécile Orgerie.

The EcoInfo group deals with reducing environmental and societal

impacts of Information and Communications Technologies from hardware to software aspects. This group aims at providing critical studies, lifecycle analyses and best practices in order to improve the energy efficiency of printers, servers, data centers, and any ICT equipment in use in public research organizations.