On efficient storage space distribution among materialized views and indices in data warehousing environments

In data warehousing environments (DWEs), two major techniques are used to reduce the OLAP query processing cost: materialized views and indices. Therefore, the DW administrator (DWA) should select views to be materialized and indices according to the available storage capacity. The DWA has to run an algorithm to select views to be materialized according to the storage space reserved for views. After that he/she runs another algorithm to select indices over base relations and materialized views according to storage space reserved for indices. In this paper, we consider the problem of distributing storage space among materialized views and indexes to minimize total query processing cost. We develop a cost model for executing a set of frequently asked queries using both materialized views and indices. Furthermore, we develop an iterative approach to automatically distribute the storage space among views and indices for a DW. We extend this approach to handle dynamic DWEs. Finally, we evaluate our algorithm to show its applicability with an example taken from the Informix white paper.