Global Query Processing and Optimization in the CORDS Multidatabase System

A multidatabase system (MDBS) integrates information from autonomous pre-existing local databases managed by heterogeneous local database management systems in a distributed environment. To achieve good overall system performance, efficient global query processing and optimization techniques are required in the MDBS. In this paper, several new techniques such as query sampling, query probing, and piggy-backing are introduced. These techniques were designed for an MDBS called the CORDS multidatabase system. The paper gives an overview of the global query optimizer in the system, including its architecture and some critical implementation techniques.