Ontology is increasingly seen as a key factor for automatic processing of machines, and interoperability between heterogeneous information systems. Ontology mapping is becoming a crucial aspect in providing the background knowledge required for solving heterogeneity problems between semantically described data sources, and accessing distributed information repositories. Developing such ontology mapping has been a core issue of recent ontology research. In this paper we present a three-layer framework to (semi-)automatically discovering ontology mapping. We show how such resulting mapping is used for resolving semantic interrogation tasks, and enabling runtime semantic interoperability across heterogeneous information systems using semantic web technologies.