MESCAL Research team

Middleware efficiently scalable

Team presentation

The goal of the MESCAL project-team is to design and validate middleware and services in order to efficiently exploit large distributed infrastructures built on aggregation of commodity components and/or commodity clusters at metropolitan, national or international scale. Our applications are intensive scientific computations. The methodology used is based on the design of bricks that scale efficiently, using modeling and performance evaluation. of target architectures, software layers and applications.

Research themes

The research efforts of MESCAL are aimed at designing and implementing middleware for clusters and grids that scale efficiently.This includes deployment tools, batch schedulers and distributed file systems. The team's research directions also include modelling, simulation, evaluation and optimization of computation grids and more generally, of large discrete event systems, using deterministic as well as stochastic techniques.

International and industrial relations

MESCAL is a partner of the network of excellence CoreGrid as well as of the national initiative Grid 5000. Academic collaborations include the University of Leiden (NL), North Carolina State University (USA) and the universities of Rio Grande Do Sul (Brazil). Our main industrial partners are Bull, France Telecom and MandrakeSoft.

Keywords: Grid computing Middleware Performance