GridCoord d 1

Annex: current projects

Project name	Description	Funding	Authority	Principal Participants
Austrian Grid	Boosting e-Science by the investigation of methodologies and the creation of an e-Science environment	2,7M€	BM:BWK	Universities of Graz, Innsbruck, Linz, Salzburg, Vienna and several hospitals

Annex B - France

This document presents a survey of funding bodies and industry in France. It is written from the viewpoint of the FP6-IST Programme, Unit F2 “Grid-based systems for Complex Problem Solving”, in opposition to “Grid as research infrastructures” as addressed in Unit F3.

We strongly emphasize that all costs below are provided in the French way, which is exclusive of the salaries of researchers or engineers. In France, most of public researchers are on permanent civil servant positions, and their salaries are provisioned directly by the state, independently of the specific projects they are working on. Only very few non-permanent public researcher positions are opened, so that this cost is negligible in practice with respect to the million-euro amounts considered below. As a matter of fact, the charged cost for a freshman researcher at INRIA is 45,000 Euros/year, and twice as much for a confirmed, senior researcher.

Short history of the Grid initiatives in the past

The work on grids in France can be traced back to 1995 with the first attempt to interconnect clusters together. As another key event, Ian Foster was an invited speaker at the Euro-Par '96 Conference organized by LIP in Lyon (at that time, directed by Michel Cosnard) to present the first results of the I-Way experiments of SuperComputing '95, considered as the first “real” experiment in Grid computing. (It interconnected 17 networked sites and ran as many as 60 applications.) The French community was thus very fast to catch up with the first glimpses of Grid. The seminal Grid book of Foster and Kesselman was widely disseminated in France.

The interest for Grid computing spread very fast in France, thanks to several factors. Firstly, the CNRS maintains a number of Research Collaborative Networks (GDR, according the French denomination) which connect French researchers working in similar areas, somewhat similar to a national Network of Excellence. One of them (http://www.arp.cnrs.fr/) is devoted to the area in Architecture, System, Networking and Parallelism (GDR ARP). The various groups interested in Grid computing got quickly connected through this structure, and started to collaborate throughout the country on this subject. Secondly, the INRIA started designing at this time its first so-called “Strategic road map” (http://www.inria.fr/inria/strategie/index.en.html) which was eventually adopted in 1999. INRIA caught up very quickly in the Grid area, and included this item in its plan, under the #1 “Strategic priority” for the Institute coined as “Designing and mastering the future network infrastructures and communication services platforms”:

“The globalization of computer resources and data -GRID- is a very promising organizational principle that consists in trivializing access to distant computing and storage resources to the extreme, to place a very large number of these resources, no matter how heterogeneous and geographically distributed, at the service of a given application. This principle shifts from the concept of server to that of service. It also makes it possible to access remote data and software without having to transport them. The standardization of the needed middleware is far from being obtained and multiple performance, security and access rights management problems must be solved on full-scale applications.”(1999)

A number of French groups dwelt into this new area. To cite a few of them, sorted by geographical order: Paris Sud (Orsay), Lille, Nancy, Lyon, Grenoble, Nice Sophia-Antipolis, Toulouse, Bordeaux, Rennes, etc.

At the same time, thanks to the very active lobbying of a number of French researchers, the French Ministry of Research opened a number of "National Networks" specialized in various fields of fundamental and applied research considered as critical for the French industry. The idea was to establish a partnership between public research and industry, and to support co-operative Research and Development projects. Each project had to be supported by at least 3 partners, of which one from industry and one from public research. A special care was devoted to SMEs, start-ups and industrial spin-offs of public labs. All projects had to show evidence of matching funding from industry.

One of the first was the National network on Networking and Telecommunications (http://www.telecom.gouv.fr/rnrt/index_anglais.htm) (RNRT in French), co-supported with the Minister of Industry and Telecommunications. This programme, launched in 1998, was steered by a committee made of scientific researchers, industry managers, telecommunication operators, and various publicly funded organisations supporting innovation. The RNRT Programme was very active until recently, supporting a large number of projects, organizing an annual conference and various thematic workshops, and publicizing project reviews. The projects were selected according to their ability to strengthen cooperation between public and industrial research, and to encourage the transfer of innovations from academic laboratories to the industry and match results with market needs. As many as 200 or so projects were funded for the 1999-2002 period (4 years), with a total public funding of 185 M€, matched by a private funding of 211 M€.

In a very similar spirit, the same partners launched in 2000 the National Network for Software Engineering (http://www.recherche.gouv.fr/technologie/reseaux/rntl.htm) (RNTL, in French), also co-funded by the Minister of Research and the Minister of Industry. These projects are of various natures, but all must include both academic and industry researchers. “Exploratory Projects” are meant to support mostly academic research, which could eventually lead to some industrial application. In such projects, the Project leader is from academia, and the funding comes from the Ministry of Research. “Pre-competitive Projects”, in contrast, take already mature academic developments to transfer them into an industrial product. In this case, the Project leader is from industry, and the funding is from the Ministry of Industry. This programme is still running. It is now steered by the newly-created National Research Agency (http://www.gip-anr.fr/en/index.htm) (ANR, in French). For the period 2000-2002 (3 years), as many as 120 projects were funded, with a total public funding of 102 M€, and a matching private funding of 89 M€.

Both the RNRT and the RNTL Programmes played a key role in the emergence of Grid in the French landscape, both in academia and industry, through their Platform programmes. RNRT funded the VTHD/VTHD++ Platform (http://www.vthd.org/?wpid=7111) which set up a "very high speed networking" infrastructure between a number of academic and industrial (France Telecom, etc.) key players in France. The RNTL launched the e-Toile Platform (http://www.cnrs.fr/cw/en/pres/compress/eToile.htm) the first large-scale deployment of a grid testbed in France, based on the Globus toolkit. These two projects were run in close synergy, and were crucial in paving the way towards today's achievements.

A crucial step in the history of Grids in France was the launching of an Incentive Concertive Action (ACI, in French) in the domain of Grids, exclusively funded by the National Science Fund (FNS) of the Ministry of Research. The so-called ACI GRID Programme (http://www.recherche.gouv.fr/recherche/aci/grid.htm) (GRID stands for: Computing and Data Resources Globalization!) started in 2001, and was one of the first ACI programmes to be set up. It was directed by Prof. Michel Cosnard, and Prof. Brigitte Plateau was the Chairwoman of the Scientific Board. Since 2004, Dr. Hab. Thierry Priol directs it. This programme played a key role in supporting the development of fundamental research in Grid technologies by academic groups. It funded as many as 15 projects per year since its debut, with an annual budget of 3MEuros, mainly for equipment and operation (this does not include the salaries of researchers). The momentum achieved through the ACI GRID Programme in Academia eventually converged with the academic/industry RNRT VTHD and RNTL e-Toile platforms to give birth to the GRID 5000 Project, launched in 2003. This national project is aimed at building a “large instrument” for Grid research, made of 5000 processors spread across 8 sites all over the country, and interconnected by a very-high speed network. The ACI GRID Programme has been now stopped.

The GRID 5000 Project has originally been funded by the ACI GRID. An additional funding came from another ACI Programme launched in 2003, devoted to "Data Masses" (ACI MD), which is still running, in the new context of the newly-founded National Research Agency (ANR). More details about the GRID 5000 Programme, its funding coordination structure can be found below.