History: Difference between revisions
No edit summary |
|||
(4 intermediate revisions by 2 users not shown) | |||
Line 1: | Line 1: | ||
{{Portal|User}} | |||
== Going up to 2003 == | == Going up to 2003 == | ||
Most key actions leading to the creation of Grid'5000 are linked to the [http://www-sop.inria.fr/aci/grid/public/acigrid.htm creation], on February 28th, 2001, by the French ministry of research of the [http://www-sop.inria.fr/aci/grid/public/ ACI GRID program], initially directed by Michel Cosnard. | Most key actions leading to the creation of Grid'5000 are linked to the [http://www-sop.inria.fr/aci/grid/public/acigrid.htm creation], on February 28th, 2001, by the French ministry of research of the [http://www-sop.inria.fr/aci/grid/public/ ACI GRID program], for ''Globalisation des Ressources Informatiques et des Données'', initially directed by Michel Cosnard. | ||
In the context of this program, Franck Cappello and Olivier Richard collaborate in the [http://www-sop.inria.fr/aci/grid/public/ ACI GRID CGP2P] project, and talk around | In the context of this program, Franck Cappello and Olivier Richard collaborate in the [http://www-sop.inria.fr/aci/grid/public/ ACI GRID CGP2P] project, and talk around 2001-2002 about the need of a large scale experimental platform in France to experimentally validate research in the domain of Grids. They present their case to Michel Cosnard, who answers that it would be easier to finance 5 sites rather than a unique large cluster. Mr Cosnard commissions a [https://www.grid5000.fr/documents/CT/rapport-grid5000-V3.pdf study], coordinated by Yves Robert, on the opportunity and usefulness of such a configuration. Franck Cappello and Olivier Richard manage to convince Yves Robert of the usefulness of such an instrument during a conference entitled ''Ecole thématique sur la globalisation des ressources informatiques et des données'' that takes place in Aussois in December 2002. In 2003, this report is presented by Michel Cosnard and ???? to the ministry of research. | ||
In the same timeframe, CNRS | In the same timeframe, CNRS funded 2 projects of interest for Grid'5000, in the context of RTP 8 Grid (pluri-disciplinary research network), directed by Brigitte Plateau. | ||
* Franck Cappello : ''Etude préparatoire pour une plate-forme de grille expérimentale d'échelle nationale'' from 2003 to 2004. This study leads to the construction of a prototype with old hardware and 4 to 8 | * Franck Cappello : ''Etude préparatoire pour une plate-forme de grille expérimentale d'échelle nationale'' from 2003 to 2004. This study leads to the construction of a prototype with old hardware and 4 to 8 nodes per clusters. | ||
* Raymond Namyst : Méthodologie de programmation des grilles | * Raymond Namyst : Méthodologie de programmation des grilles | ||
Line 20: | Line 22: | ||
* usage of a resource manager (OAR) for experiment scheduling and resource reservation | * usage of a resource manager (OAR) for experiment scheduling and resource reservation | ||
The prototype is replaced by larger clusters and the platform is opened to users in 2005. The [[Grid5000:School2006|first spring school]] ( | The prototype is replaced by larger clusters and the '''platform is opened to users in 2005'''. The [[Grid5000:School2006|first spring school]] (February 2006) is a tipping point for usability and usage of Grid'5000, as a lot of effort by the technical team was put into writing the different tutorials given during that school and putting everything together so these tutorials would work. | ||
in March 2007, a scientific report entitled ''Report on the Scientific Developments and Results of Grid’5000'' was coordinated by Franck Cappello and successfully presented in July 2007. At that date Inria offered to step in to ensure funding for the technical team and continued operations of the successful platform. | in March 2007, a scientific report entitled ''Report on the Scientific Developments and Results of Grid’5000'' was coordinated by Franck Cappello and successfully presented in July 2007. At that date Inria offered to step in to ensure funding for the technical team and continued operations of the successful platform. | ||
== ADT phase (2008-2012) == | == ADT phase (2008-2012) == |
Latest revision as of 12:41, 21 April 2015
Going up to 2003
Most key actions leading to the creation of Grid'5000 are linked to the creation, on February 28th, 2001, by the French ministry of research of the ACI GRID program, for Globalisation des Ressources Informatiques et des Données, initially directed by Michel Cosnard.
In the context of this program, Franck Cappello and Olivier Richard collaborate in the ACI GRID CGP2P project, and talk around 2001-2002 about the need of a large scale experimental platform in France to experimentally validate research in the domain of Grids. They present their case to Michel Cosnard, who answers that it would be easier to finance 5 sites rather than a unique large cluster. Mr Cosnard commissions a study, coordinated by Yves Robert, on the opportunity and usefulness of such a configuration. Franck Cappello and Olivier Richard manage to convince Yves Robert of the usefulness of such an instrument during a conference entitled Ecole thématique sur la globalisation des ressources informatiques et des données that takes place in Aussois in December 2002. In 2003, this report is presented by Michel Cosnard and ???? to the ministry of research.
In the same timeframe, CNRS funded 2 projects of interest for Grid'5000, in the context of RTP 8 Grid (pluri-disciplinary research network), directed by Brigitte Plateau.
- Franck Cappello : Etude préparatoire pour une plate-forme de grille expérimentale d'échelle nationale from 2003 to 2004. This study leads to the construction of a prototype with old hardware and 4 to 8 nodes per clusters.
- Raymond Namyst : Méthodologie de programmation des grilles
In 2003, Thierry Priol replaces Michel Cosnard as the head of ACI GRID.
ACI phase (2003 - 2007)
In 2003, the Grid'5000 project is formally started, as 7 projects of ACI GRID. Franck Cappello is nominated as head of the steering committee. In October 2005, Pierre Neyron is hired to assist Franck Cappello in the coordination of the technical team.
During that period, many strategic choices are made
- isolation from Internet
- reconfiguration through reboot and Kadeploy
- dedicated network links between the sites
- usage of a resource manager (OAR) for experiment scheduling and resource reservation
The prototype is replaced by larger clusters and the platform is opened to users in 2005. The first spring school (February 2006) is a tipping point for usability and usage of Grid'5000, as a lot of effort by the technical team was put into writing the different tutorials given during that school and putting everything together so these tutorials would work.
in March 2007, a scientific report entitled Report on the Scientific Developments and Results of Grid’5000 was coordinated by Franck Cappello and successfully presented in July 2007. At that date Inria offered to step in to ensure funding for the technical team and continued operations of the successful platform.
ADT phase (2008-2012)
Inria's support would take the form of an ADT (technological development action) named ALADDIN-G5K. A proposal, coordinated by Thierry Priol, Franck Cappello and David Margery was written in late 2007, with the objectives of running Grid'5000 with production level system administration and developing better tools for its usage. The proposal was accepted and the project formally started on June 1st, 2008. The key meeting to prepare this proposal was held on October 24th, 2007, in Paris.
Between fall 2007 and June 2008, a new organisation is put in place. Thierry Priol takes the role of director, Franck Cappello as scientific director and David Margery as technical director (On December 1st, 2007).
In April 2009, Frédéric Desprez becomes scientific director, Franck Cappello deputy scientific director as Thierry Priol leaves the project for bigger responsibilities.
GIS phase (2012 -> ?)
In May 2012, Grid'5000 becomes supported by the Scientific interest group (GIS) Grille 5k (to be renamed). The Scientific interest group has representatives form CNRS, CEA, Inria, CPU (representing the universities), RENATER, Institut Telecom and CDEFI (representing engineering schools) as founding members. Michel Cosnard heads the steering committee, who has approved of
- Frédéric Desprez as director,
- David Margery as technical director,
- Laurent Gydé, Lucas Nussbaum, Chirstian Perez, Emmanuel Jeannot and the director as members of the executive office
- Lucas Nussbaum, Olivier Richard, Adrien Lebre and the technical director as members of the architecture committee.