Wikipedia tells me that NEPOMUK stores it's data as RDF. As far as I've been able to figure out it uses Soprano to save the information. But where does it store it? And how does it search it (redlane, Sesame2?) Also, what exactly is the role of Strigi?

Strigi is a indexer, used to extract metadata and other information from files existing on your computer. Nepomuk uses Strigi to gather information and stores it in it's database.

Nepomuk stores it's data under ~/.kde4/share/apps/nepomuk/repository/main/data/. The exact sub-folder used is dependent upon the backend in use. Virtuoso is currently recommended for this. Redland is much too slow for Nepomuk, and Sesame2, while an improvement is not fast enough either unfortunately.