Over time, businesses became aware that the collective information contained in
their various systems had great potential for analyzing market trends and improving
their business processes. However, their OLTP systems were unsuitable for executives
to perform this task, given the poor performance and complex interface for ad hoc
analysis queries (e.g., aggregated multitable joins and nested correlated subqueries).
Moreover, insufficient integration or history made some queries simply impossible.
Partly to address the problem of integrating data from prerelational systems for analysis purposes, IBM proposed the notion of an “information warehouse”. Although performance
problems delayed acceptance of this idea for some years, a later proposal for
a “data warehouse” by Bill Inmon (1993) was enthusiastically embraced by the business
community, and nowadays most large companies have a data warehouse.