Local and global query optimization mechanisms for relational databases
Kazuhiro Satoh, Masashi Tsuchida, Fumio Nakamura, 一彦 大町 · Very Large Data Bases · 1985
In a relational database management system (DBMS), query optimization is essential since the DBMS must produce candidate internal access strategies from a high level query and select the optimal one. The goals of inplementing a good optimization mechanism are: (a) to maximize the performance of selected access strategies and (b) to minimize the optimization overhead. We have developed an optimization mechanisin, built it in a relational DBMS product at Hitachi. The mechanism can estimate precise tuple selectivity of predicates in queries for optimal access strategy selection, and employs a new method, called the cascade method, for optimization overhead reduction. This paper also proposes a global optimization mechanisn which executes optimization over multiple database queries.