Query optimization in distributed database systems
In a distributed database system, data are managed on a network of computers. To access data distributed in different sites, the transmission of data over communication links is necessary. Since communication delay is substantial, query processing is considered strongly related to the performance efficiency of a distributed database system. Optimization of distributed query processing is studied here. With the objective of minimizing total data transmission cost, the semi-join tactics for query processing is applied. First, the properties of optimal semi-join programs for a special type of queries, called tree queries, are investigated. An execution graph is introduced for representing semi-join programs. From the structure of the excecution graph, two properties are derived that an optimal semi-join program has to satisfy. A new approach is suggested for processing join clauses in a query, and two more optimality properties are obtained. Based on these properties, the traditional semi-join processing method is revised such that the cost for processing a query can be reduced. Four improvement algorithms were developed; these algorithms can be used to improve the semi-join program generated by any heuristic query processing algorithm.
- OSTI ID:
- 5544539
- Country of Publication:
- United States
- Language:
- English
Similar Records
Distributed query processing in fast local networks
An optimization of queries in distributed database systems