Self Organizing Semantic Topologies in P2P Data Integration Systems

A semantic topology is a peer overlay network connected via semantic links, constructed using schema mappings and used for peer querying. The large-scale and dynamic environments of P2P networks dictate the use of automatic schema matching, which was shown to carry with it a degree of uncertainty. Therefore, peers prefer network topologies that improve their ability to answer queries effectively, by reducing uncertainty. We introduce a model for a peer database management system that manages the inherent uncertainty of automatic schema matching, the amplification of this uncertainty over transitive mappings, and its impact on query processing. We then briefly present the research challenges involving a dynamic topology setting where peers can change their neighbor set selection.