Querying data sources in a SuperPeer Data Management System

Numerous data integration systems have been proposing powerful query mechanisms of data sources. However these systems do not facilitate the evolution of existing schemas or the integration of new schemas. Many applications induce the continual arrival of data sources and thus of new schemas. These applications require more flexible systems. The PDMS (Peer Data Management Systems) supply an interesting alternative. In this paper we propose a SuperPeer Management System allowing local interoperability in a SuperPeer cluster and permitting global cooperation between the SuperPeers. We define first the architecture components and the query processing in terms of description logics. The descriptions of queries and of data source capabilities can be annotated with constraints on the attribute values domain in order to improve the data sources discovery process. We then propose a solution for query processing in such systems by adapting an existing algorithm.