Research project

CubeStar is a multidimensional processing tier on top of a single relational database system or a set of distributed database systems by applying CORBA. The new layer is the basis for analytical computations as usual in OLAP, standard reporting or data mining. Besides creating a versatile and concise multidimensional data model CubeStar especially has concentrated on query optimization by applying preaggregates, i.e. precomputed query results. Within the project an intelligent strategy for buffering preaggregates based on the current query behaviour was established. Further complex algorithms for query rewriting were developed that can benefit even from an only minimal overlapping of query and preaggregate.