Data Cubes in Dynamic Environments

The data cube, also known in the OLAP community as the multidimensional database, is designed to provide aggregate information that can be used to analyze the contents of databases and data warehouses. Previous research mainly focussed on strategies for supporting queries, assuming that updates do not play an important role and can be propagated to the data cube in batches. While this might be sufficient for most of today’s applications, there is growing evidence that modern interactive data analysis applications will have to balance update and query costs. Two techniques for maintaining data cubes in dynamic environments are described here. The first, Relative Prefix Sums (RPS), supports a constant response time for ad-hoc range sum queries on the data cube, while at the same time greatly reducing the update costs compared to prior approaches. The second, the Dynamic Data Cube (DDC), guarantees a sub-linear cost for both range sum queries and updates.