Schema mediation for large-scale semantic data sharing

摘要: Intuitively, data management and integration tools should be well suited for exchanging information in a semantically meaningful way. Unfortunately, they suffer from two significant problems: typically require common comprehensive schema design before can used to store or share information, are difficult extend because evolution is heavyweight may break backward compatibility. As result, many large-scale sharing tasks more easily facilitated by non-database-oriented that have little support semantics.The goal of the peer system (PDMS) address this need: we propose use decentralized, extensible architecture which any user contribute new data, even mappings between other peers’ schemas. PDMSs represent natural step beyond systems, replacing their single logical with an interlinked collection semantic individual schemas.This paper considers problem mediation PDMS. Our first contribution flexible language mediating schemas extends known formalisms our complex architecture. We precisely characterize complexity query answering language. Next, describe reformulation algorithm generalizes both global-as-view local-as-view algorithms. Then several methods optimizing initial set experiments studying its performance. Finally, define consider global problems managing