DartGrid
DartGrid is a data integration framework using semantic web technologies. It features in visuliazed semantic mapping tools (Currently support relational-RDF/OWL mapping, XML-RDF/OWL mapping is still under development), and a SPARQL-SQL query component based on rewriting-query-using-view approach.
More information could be found at DartGrid.
A roadmap for future development
2007/06/13 This is a list of future direction we have in mind right now that might be explored into in the furture. It might be changed a lot for real implementation.
Full OWL mapping support: Currently, DartGrid only supports mapping from relational data to RDFS, and the simplest OWL syntax at the schema level. The query engine does not have OWL support yet. Besides, with more experiences in applications such as TCM and neuroscience, much more complex mappings exist. This new mapping tool will cover more complex mapping cases and will be tested against by the life science demo use cases and the THALIA Testbed. The new mapping tool will take a browser-based framework using AJAX-based toolkits such as openlink ajax framework.
Reasoning support for query rewriting One of the advantage of semantic web technologies is the reasoning capability. However, performance while dealing with large amount of semantic data has becoming an unavoidable bottleneck for most of the currently available RDF/OWL stores. Relational databases have been full-fleged in many performance issues, especially for query processing. DartGrid takes an mapping-and-query-rewriting-based approach to laying semantic web upon relational databases to take full advantage of the performance of commertial DBMSs. However, DartGrid has limited support for reasoning, particularly insufficient of supporting OWL reasoning. A reasoner will be coupled with the query rewriting component to enable full OWL reasoning capability of query answering.
Semantic query endpoint planning This was inspired by a note from openlinsw, which has given detailed description on this problem. Currently, DartGrid only support selecting candidate data sources to fullfill the query task, but more complex case might exist when several data sources need to be combined to form a workflow to fullfill a query task. In this case, more complex query scheduling and planning approach might be necessary.