Information grids : managing and mining semantic data in a grid infrastructure :
open issues and application to geno-medical data
LIRIS, INSA Lyon
Abstract:
Over these last eight years, computing grids have received an increasing attention both from
the scientific community and the major industrial companies. Indeed, computing grids propose
a very effective, economic and scalable way to build high performance infrastructures.
Basically a grid a composed of a set of computing resources (CPUs, memories, storage areas,
softwares...) distributed over a network and a middleware designed to allow sharing and using
these resources (resource manager, resource broker, authentication system, replica manager, etc.).
Grids have initially been used for high performance scientific computing (e.g. nuclear physics,
climate studies...). So, until now, most of middleware tools have been designed in this application
framework. More recently, attempts have been done to open grids to other application fields. Indeed
grids appear as a promising candidate infrastructure paradigm for managing applications distributed
over a MAN/WAN and designed to share information and services among users. Thus, aside computing
grids, information grids and community grids are now considered with a very high interest.
In this perspective, by their very demanding requirements and their socioeconomic impact, biomedical
applications are certainly one of the most promising domain for using grids.
However, grids yet lack most of data management tools mandatory to pretend implement semantically
rich applications like biomedical applications :
- data warehouse systems and tools have not yet been implemented on grids ;
- data mining algorithms and knowledge discovery tools need to be adapted to grids ;
- collaborative data brokers able to share and exchange data in an economic, safe and efficient
way while preserving privacy have to be developed.
In this position talk, we will first draw a global picture of the latest achievements and open issues
in grid computing. Then, we will focus on specific issues regarding information grids, i.e.
the management and mining of semantic data in distributed and heterogeneous grid infrastructures.
Finally, we will illustrate the links that information grids show with related research fields,
e.g. pervasive information systems and Internet service integration.
This talk is based on an ongoing project leaded by the LIFL lab (Lille, France), the IRIT lab
(Toulouse, France) and the LIRIS lab (Lyon, France) which aims to build a toolbox for managing
and mining genome and medical data in collaborative grid networks.
hameur@irit.fr - Workshop
Home Page