Abstract. Ontology-based modelling is becoming increasingly important in the design of complex knowledge management applications. However, many problems related to large-scale ontology development, deployment and collaborative maintenance of related metadata still remain to be solved. Making online modifications to an ontology whose concepts are simultaneously being used for metadata generation may potentially disrupt metadata semantics and even introduce inconsistencies. In this paper we analyzed and classified operations on ontology according to their impact on metadata. The approach is aimed at environments (like our own Knowledge Hub) relying on a relational database for storing ontologies and metadata assertions and it is based on the use of database triggers for automating metadata maintenance.