Large control centers, as needed in road traffic, typically manage highly dynamic environments. They process vast amounts of information from heterogeneous data sources about a large number of real-world objects, which are anchored in time and space. In such systems, human operators are vulnerable to information overload and, thus, may fail to be aware of the overall meaning of available information and its implications. With BeAware, we propose a software framework that supports the development of situation awareness applications for control centers. The contribution of this paper is twofold: First, we integrate existing ontologies with spatio-temporal reasoning concepts, focusing on extensibility. We introduce meta-modeling concepts that allow us to assess and project situations and actions using semantic web technology. Second, we compare the runtime performance of the situation comprehension capabilities of a generic, ontology-driven implementation and a domain-specific relational-database-backed implementation, and discuss the strengths and shortcomings of each approach.