Abstract. Finding the optimal selection of an OWL reasoner and service interface for a specific ontology-based application is challenging. Over time it has become more and more difficult to match application requirements with service offerings from available reasoning engines, in particular with recent optimizations for certain reasoning services and new reasoning algorithms for different fragments of OWL. This work is motivated by real-world experiences and reports about interesting findings in the course of developing an ontology-based application. Benchmarking outcomes of several reasoning engines are discussed – especially with respect to accompanying sound and completeness tests. We compare the performance of various service and communication protocols in different computing environments. Hereby, it becomes apparent that these largely underrated components may have an enormous impact on the overall performance.