GRID Computing

Foster, Kesselman and Tuecke introduced the definition of the term grid in the end of the 90s. The term is defined as the coordinated sharing of computational resources between different administrative domains in dynamic distributed environments. In 2002 Foster proposed to restrict the definition. A system constitutes a grid if it is decentralized, is based on open protocols and implements advanced quality of service mechanisms.

A grid is therefore a distributed system that aggregates under a single infrastructure computational resources from multiple organizations that are autonomous, heterogeneous and geographically distributed. The grid hides the computational resources specificities, exhibiting a single interface that allows transparent access to the infrastructure.

The motivation behind the grid is the resolution of complex computational problems with demanding computing and data requirements. The dimension and cost of the required resources make the resolution of this class of problems extremely difficult or even impossible in the context of the conventional architectures. The development of the data communication technologies has created the opportunity for the development of solutions based on the integration of geographically distributed resources making possible the deployment of the grid concept.

Today the grid technologies are an essential condition for the technological and scientific development and receive special support and attention from the governments. In this context the European Union has strongly supported the development and deployment of grid computing in the 5th and 6th frameworks. It is expected that Portugal will join this challenge, namely through the creation of a national grid programme as is the case in the other European countries.

In the most developed countries the interests and deployment of grid technologies in the commercial and industrial areas has been growing. In this context private companies have successfully used grid technologies to integrate internal resources and provide computing capacity for demanding processing tasks.

In Portugal, LIP a High Energy Physics research laboratory is the organization with the biggest participation in grid development and deployment projects such as DataGrid, CrossGrid, "Enabling Grids for E-SciencE" (EGEE) and the CERN project "LHC Computing Grid" (LCG). This effort aims to prepare the national Particle Physics community for the challenge raised by the new experiments being prepared at CERN. Each of these four experiments will produce 2 PetaBytes of data per year. Currently the LCG grid has already 78 sites and more than 7000 processors. LIP is one of these sites.

Last update: Tue Nov 9 18:17:00 UTC 2004

Webmaster: João Martins