Push down optimization in a distributed, multi-database system

摘要

High performance query optimization in a heterogeneous distributed multi-database system. An efficient technique is disclosed to enable an interface module, located between a host computer and a back- end database system, to perform maximal query or subquery push down. That is, the interface module is configured to select either the entire query or the largest subquery(s) that can be forwarded to a single database instance within the back-end database system without extensive decomposition or extraneous commands. The interface module has stored therein a data structure having information concerning the data stored in, and the capabilities of, each of the back-end databases in the heterogeneous environment. Based on this information, the interface module determines whether a query or subquery satisfies two criteria. First, the interface module determines whether a single database instance within the heterogeneous environment contains all of the data referenced in the query or subquery. This is referred to as data coverage. Next, the interface module determines whether the same single database instance provides all the functions or capabilities needed to satisfy the query or subquery. This is referred to as function or capability coverage. If both of these criteria are met, the query or subquery can be pushed down to the single database instance.