ANR
Presentation and Objectives
The FORUM project deals with the problem of information integration
in a large and highly dynamic information space.
As communication infrastructures advanced, in particular the evolution
of Internet technologies, needs for ubiquitous access to distributed
information sources increased. Today, data analysis and integration techniques
are becoming more and more prominent features of enterprise and government
systems. They offer tremendous opportunities for empowering users and organisations
in a variety of application domains including electronic commerce, scientific
databases, enterprise information integration, digital government, etc.
The aim of the FORUM is to leverage and seamlessly extend current
mainstream information integration technologies to cope with information
sources interoperability problem in Web like environment. More precisely,
the main objective of the project is to develop a scalable and flexible
information sharing infrastructure that enables effective use of a potentially
large and dynamic collection of information sources. We build our approach
on an existing P2P system, namely the Xpeer architecture. Xpeer organises
the information space in a network of peer communities.
In this context, we plan to investigate the following critical and open
research issues:
- In order to alleviate the integration task, and hence to enable
a flexible information sharing infrastructure that copes with the dynamic
nature of the information space, we aims at developing techniques for
automatic mapping discovery between heterogeneous information sources.
- To support content based querying in P2P systems, we propose to
investigate novel and flexible query rewriting techniques that copes with
the absence of a centralized mediated schema.
- To cope with the large size of the information space, we propose
to investigate design of new query rewriting algorithms that scale in
the number of available information sources.
- To deal with the dynamic aspect of P2P context at semantic level,
we propose to investigate the mapping graph restructuring and their impact
on query optimization.
A prototype will be experimented on the CEMAGREF application