ROWLEX is created primarily to browse and manipulate RDF graphs in memory. The RDF graph itself is represented by an instance of the RdfDocument class (e.g., you can load an RDF file from disk by calling new RdfDocument(@"C:\myRdfGraph.rdf");
. The instances of the ROWLEX generated classes (aka ROWLEX proxy classes) represent a node in the graph. These proxy instances hold direct reference to their RdfDocument. Therefore, you may execute ROWLEX both on the server and on the client side, as long as keep the RdfDocument instance the proxy class instances in the same AppDomain. Or in other words, keep the RdfDocument and its proxy classes on the same end of the wire.
If you wish to process RDF on the client side with Silverlight, it should not be a problem. Send the RDF graph over the wire as an RDF/XML document, load the graph on the client side into the RdfDocument and you are free to browse/manipulate it with the instances of your proxy classes.
You might want to read my answer on this question, too, for deeper understanding what goes on under the hood.