Community-oriented Information Integration
Yannis Katsis (UCSD)

Many online communities, be they private, commercial, academic or 
governmental, have structured data that they would like to integrate and 
share between their members. However information integration of structured
data is currently the privilege of only big communities that have the 
resources required for such large-scale projects. Indeed the construction of 
scientific integration portals, such as BIRN (Biomedical Informatics Research 
Network) and GEON (The Geosciences Network) is still a very large-scale 
effort, which has considerable financial cost and takes many years to 
initiate and accomplish. The main reason is that they rely on traditional 
integration technology, where all integration tasks are carried out by a 
central integration authority, which quickly becomes the bottleneck.

This leaves out all other communities with structured information integration 
needs whose limited time and financial budget rules out the traditional 
integration solution. To address this need, we introduce the community-based 
integration paradigm which enables systems that integrate and query 
structured data into a global database, at no cost to any central authority. 
This is achieved by decentralizing the setup and maintenance tasks, pushing 
them to the independent community members.

In this talk we indentify the major setup and maintenance tasks that have to 
be carried out in an integration system and show how they can be distributed 
to the individual community members. Being a summary of both our past 
research and our future plans, this talk contains both full-blown solutions 
and proposals on future work, which you can help shape through your comments 
and/or get involved if you are interested.