The Aqua approximate query answering system

TLDR

Aqua is a system that delivers fast, approximate answers to common OLAP aggregate queries. It operates atop any commercial DBMS by precomputing statistical synopses, rewriting queries to run on these synopses with quality guarantees, and maintaining the synopses incrementally as the database changes.

Abstract

Aqua is a system for providing fast, approximate answers to aggregate queries, which are very common in OLAP applications. It has been designed to run on top of any commercial relational DBMS. Aqua precomputes synopses (special statistical summaries) of the original data and stores them in the DBMS. It provides approximate answers along with quality guarantees by rewriting the queries to run on these synopses. Finally, Aqua keeps the synopses up-to-date as the database changes, using fast incremental maintenance techniques.