MIL primitives for querying a fragmented world

In query-intensive database application areas, like decision support and da ta mining, systems that use vertical fragmentation have a significant perfo rmance advantage. In order to support relational or object oriented applica tions on top of such a fragmented data model, a flexible yet powerful inter mediate language is needed. This problem has been successfully tackled in M onet, a modern extensible database kernel developed by our group. We focus on the design choices made in the Monet interpreter language (MIL), its alg ebraic query language, and outline how its concept of tactical optimization enhances and simplifies the optimization of complex queries. Finally, we s ummarize the experience gained in Monet by creating a highly efficient impl ementation of MIL.