With inference you will be able to apply virtual RDF and RDB2RDF /OBDA technology in use cases reserved for traditional OWL reasoners (slow with lots of data) or RDF triples stores based on forward or backward chaining (inefficient reasoning). Interesting use cases for inference in RDB2RDF/OBDA include:
On-the-fly Ontology-based Data Access
SPARQL over RDFS/OWL mappings
Data Integration with DB federation
Extremely high performance, even in the presence of very large data sets, complex queries and large ontologies.
ontop implements the latest generation query rewriting techniques that translate your SPARQL into SQL in the best possible way. The SQL queries are so efficient that they often allow provide 10x times better performance than other SPARQL engines (open source or even commercial), and in complicated queries/data sets, up to 500x times.