SDQNET: Semantic Distributed Querying in Loosely Coupled Data Sources

Web communities involve networks of loosely coupled data sources. Members in those communities should be able to pose queries and gather results from all data sources in the network, where available. At the same time, data sources should have limited restrictions on how to organize their data. If a global schema is not available for such a network, query processing is strongly based on the existence of (hard to maintain) mapping rules between pairs of data sources. If a global schema is available, local schemas of data sources have to follow strict modelling restrictions posed by that schema. In this paper, we suggest an architecture to provide better support for distributed data management in loosely coupled data sources. In our approach, data sources can maintain diverse schemas. No explicit mapping rules between data sources are needed to facilitate query processing. Data sources can join and leave the network any time, at no cost for the community. We demonstrate our approach, describing SDQNET, a prototype platform to support semantic query processing in loosely coupled data sources.