GridCoord d 1

Research lines or approaches to the past or current initiatives

The emergence of Grid in the French landscape has been mainly driven by the various public partners under the general umbrella of the Minister of Research and Technology. Depending on the activities and times, the Directorate of Research or the Directorate of Technology took the lead in promoting this line of research, but over the years, one can witness a mostly constant direction of action. This relative stability in the strategy in spite of the many disturbing political events of the past years can be traced to the very active influence of a few number of key people. To cite all but a few: Michel Cosnard for the software aspects, Dany Vandromme for the networking aspects, etc.

From the beginning, Grids have been considered in France as a subject of research by the Academia, and the structure of public funding made it possible to maintain this line of action. The two main public research institutes in the IST field, CNRS and INRIA, had key roles in motivating their research groups in this direction and in setting up the necessary “human networking” between them through a number of co-operative research programmes. In France, it can be seen that public research has taken the lead in the development of Grid technology, and could promote a long-term vision of the scientific approach, which culminated in the national GRID 5000 large research instrument project. Of course, industry was not disregarded, and it could also participate more through the “National Networks” RNRT and RNTL. But in all cases, the partnership was balanced with industry partners, and the public research group were not dependent on the often too short-term vision of industry.

Looking at the past, it can be seen that the two platforms “VTHD” for advanced networking and “e-Toile” for advanced Grid technology where really instrumental in letting the French community aggregate and co-operate on this line of research. This “platform-based” approach should thus be highly recommended for other countries wishing to reach a suitable momentum in this area. Also, the French approach to platforms, with matching funding from both public institutions and industry, proved to be very efficient: actually, a positive synergy emerged from this approach, and academia and industry could collaborate without one being dependent on the other. At least in this very domain this arises as a necessity if one wants to focuses on long-term visions and challenges, instead of short-term, technical improvements.

Since the early beginnings of GRID 5000 Project, the notion of a “large instrument” for Grid research has appeared as a central reference. Actually, the vision is that GRID 5000 is a “large instrument” for research in computer science, much in the same sense as a telescope in Chile is a “large instrument” for astro-physicists. It should be directed by a public body, independently of any short-term, technical commitment; it should be widely opened to foreign collaboration; it should be designed so that new, and most probably unforeseen experiments can be run. This implies that the platform cannot offer any industry-like commitment for quality/continuity of service, in complete contrast with an industry product.

Funding structure and co-ordination

As far as France is concerned, the funding structure of the Grid technology activity is mainly dependent on the Ministry of Research and Technology. In the current governmental organisation, this is a sub-minister of the Ministry of National Education, Academic Education and Research, directed by François Fillon. The Delegate-Minister for Research (http://www.recherche.gouv.fr/) is François D'Aubert.

The Ministry of Research and Technology is divided into a number of Directorates:

• 
  The Research Directorate is led by Elisabeth Giacobino. Michel Bidoit is in charge of Mathematics, Information, Communication Sciences and Technologies (STIC, in French).
  
• 
  The Technology Directorate is led by Jean-Jacques Gagnepain. Joseph Mariani is in charge of Information and Communication Technologies.
  

In addition to the Ministry, France has a number of Scientific and Technological Public Institutes (EPST, in French), which have specific missions for scientific research and innovation. In the IST sector, one can find the CNRS (http://www2.cnrs.fr/en/8.htm) (circa 10,000 researchers), which is multi-disciplinary. Regarding IST, CNRS includes 750 full-time researchers and a budget of 140 M€ (excluding salaries). One can also find INRIA (http://www.inria.fr/inria/enbref.en.html) which is specialized in Computer Science and Automatics, which holds circa 500 full-time researchers with a total budget of 135 M€ (excluding salaries). These two institutes have their own President and Executive Directors, and a large autonomy regarding their scientific politics with respect to the Ministry.

Before describing the various funding bodies in detail, it should be stressed that the whole system is just now being deeply reshaped, with the creation in February 2005 of the National Agency for Research (http://www.gip-anr.fr/en/index.htm) (ANR, according its French denomination), which is essentially meant to be a French adaptation of the US NSF. This central agency is planned to become the unique source of support for all the public programmes described below: RRIT, ACI, etc. The French funding system is now just in the transition, so that many aspects are still unclear. We describe below the status at the end of 2004, that is, before ANR became active, stressing the aspects which are thought to continue in the new system.

National Research and Innovation Networks (RRIT)

A main channel for grid funding and co-ordination in France are the National Research and Innovation Networks programmes (http://www.recherche.gouv.fr/technologie/reseaux/) (RRIT in French). They associate the Ministry of Research and Technology together with the Ministry of Industry. These programmes are a major tool for French national R&D through a number of activities. They provide a formal framework for partnership between public research and industry, at the level of funding, legal aspects, scientific focusing, evaluation, etc. RRIT programmes support co-operative R&D projects, with the minimal rules that they should involve at least 3 partners, of which 1 from industry, and 1 from public research. Each programme defines its own joint objectives, priorities, call structure, proposals, evaluation, selection, project follow-up, etc. RRIT programmes also promote accompanying activities: annual conference, reports, etc. It is part of their mission to pay special attention for SMEs, start-ups and spin-offs. The general rule is that the public and industry funding should overall match for each single programme.

As many as 16 such RRIT programmes have been launched in all areas, and 4 in the Information and Communication Technologies:

• 
  Telecommunications (RNRT), 1998
  
• 
  Micro and nanotechnologies (RMNT), 1999
  
• 
  Software engineering (RNTL), 2000
  
• 
  Audiovisual and multimedia (RIAM), 2001, co-supported by the Ministry of Culture.
  

All these 16 programmes have been funded on the Research and Technology Fund (FRT) of the Ministry of Research. The overall funding amounted to 200 M€ in 2004. A number of smaller public agencies for innovation brought also specific support for SMEs (ANVAR, etc.)

Grid technology has been mainly promoted through two of these programmes: RNRT and RNTL, and it can probably be stated that the larger part of the industry involvement in developing Grid technology has been done within these programmes, on a 50-50 support basis.

The RNRT and RNTL programmes proposed a numbers of project categories.

• 
  Exploratory projects (market horizon: 5 years). They are basic research oriented, addressing technical challenges. They usually result in a preliminary research prototype which validates the feasibility of the project. They are funded by the Ministry of Research and Technology.
  
• 
  Precompetitive projects (market horizon: 3 years). They aim at technology transfer, resulting in pre-industrial product or services. They are funded by the Ministry of Industry.
  
• 
  Testbeds/Platforms. These are large projects offering technical facilities to external users, both from academia and industry. They are co-funded by the two Ministries.
  

The RNRT and RNTL programmes have been instrumental in the development of grid technology in France. The platforms have played a key role in this. A more detailed presentation of the RNRT and RNTL platform can be found below in this document. Each of the programmes has funded between 20 and 30 platforms, often in co-ordination with other European teams involved in similar projects. Among these many platforms, two have played a crucial role: VTHD and e-Toile, as described earlier in this document.

The RNTL programme has been adopted by the new French ANR Agency, whereas the RNRT has been stopped as of now.

Concerted Incentive Actions (ACI)

The ACI programmes (http://www.recherche.gouv.fr/recherche/fns/index1.htm) are directly funded by the Ministry of Research itself, using the National Science Fund (FNS, in French). Only Public Research laboratories can apply for funding, though they may work in association with industrial partners. In 2002, the FNS amounted to 150 M€, and the IST projects were funded by 4 M€. In 2003, the figures were 160 M€, of which 12 M€ for IST and 1.5 for specific Computing/Communication infrastructures.

Most of the funds for IST were devoted to the ACI GRID programme (http://www.recherche.gouv.fr/recherche/fns/grid.htm) launched in 2001 under the direction of Michel Cosnard until 2004, and then Thierry Priol. Brigitte Plateau is the Chairwoman of the Scientific Committee. Five categories of projects were defined: Multi-disciplinary projects, Software projects, Emergent teams, Community animation, International co-operation. In the period 2001-2002, as many as 80 3-year projects were submitted and 40 selected, for a total funding of around 6 M€.

The ACI GRID projects were instrumental in creating a real collaboration between the academic research groups throughout the country, and letting these groups develop novel (and often risky!) ideas about Grid technology. A selection of interesting ACI GRID projects can be found below in this document.

In 2003 two additional ACI programmes of interest were launched by the Ministry of Research: ACI “Data Masses” (ACI MD) (http://www.recherche.gouv.fr/recherche/fns/mdonnees.htm) under the direction of Bernard Péroche, and ACI “Security and Informatics” (ACI SI) (http://www.recherche.gouv.fr/recherche/fns/securinf.htm) directed by Claude Kirchner. In opposition to ACI GRID, these two programmes are co-funded by the Ministry, CNRS, INRIA and a number of additional partners (for instance, Ministry of Defence). The ACI MD programmes took over the finishing ACI GRID to fund many projects in Grid technology, with a special emphasis on the management of large amounts of data. Each of them funds 20-30 projects each year.

The ACI GRID programme was already planned to be stopped this year. The ACI MD and ACI SI programmes were still supposed to run for a couple of years. There have been unofficial announcement that they should eventually be adopted by the new French ANR Agency, but there is no evidence yet of any official decision.

The national GRID 5000 Project

As mentioned before, the VTHD high-performance networking RNRT platform and the e-Toile effective grid deployment RNTL platform have been instrumental in building the grid technology community in France, first at the academic level, and also by tightening the links between public research groups and industry: IBM, Sun Microsystems, HP, etc.

All this momentum resulted in the definition of the ambitious GRID 5000 collaborative project (http://www.grid5000.org/). This project aims at building an experimental Grid platform gathering 8 sites geographically distributed in France. The main purpose of this platform is to serve as an experimental testbed for research in Grid computing. This high collaborative research effort is funded by the French Ministry of Research, INRIA, CNRS, the local universities of all sites and some regional councils. Moreover, depending on the local relations of each site, industry partners have been involved through discount rates for hardware equipment and/or research contracts. This concerns mainly IBM, Sun, and Apple.

The originality of this project is that it is in some sense bottom-up, in contrast with the previous top-down actions. A number of groups decided to put in common their resources, whatever their origin, and to collaborate into Grid'5000 is a research effort developing a large-scale nation wide infrastructure for Grid research. 10 laboratories are involved, nation wide, in the objective of providing the community of Grid researchers a testbed allowing experiments in all the software layers between the network protocols up to the applications: applications, algorithms, runtime, middleware, operating systems, network protocols. Grid'5000 will allow Grid experiments France-wide in all these software layers.

The current plans are to assemble a physical platform featuring 8 clusters, each with a hundred to a thousand PCs, connected by the Renater Education and Research Network. All clusters will be connected to Renater at 1 Gb/s (10 Gb/s is expected in near future). This is the main Grid Infrastructure in France. More details about this project can be found on the webat the reference site (http://www.grid5000.org) and in the companion D.3.1.2 Deliverable: “Survey of universities and research labs”.

The grand total of funding which has been gathered by the various sites through all the possible channels, both national and local, can be estimated today at 7.5 M€ for hardware equipment.

As emphasized above, we stress that this does not include the salaries of the researchers and publicly funded engineers involved in the project, which are civil servants with a permanent position directly funded by the state. As many as 810 sites are involved, and at least 10 persons at each lab are actively contributing to the project (some full-time, some part-time), it can be safely estimated that the project induces an additional cost of 40-60 person-years.