4 - Query Optimization Concepts and Methodologies in Multidatabase Systems

Abstract

In a multidatabase system (MDBS), the participating databases are autonomous. The schemas of these databases may be different in various manners even though the same information is represented. The execution of a global query in this environment involves levels of translations. Query optimization can be partially achieved at each of these levels. This chapter discusses past research at these optimization levels. A problem with translation is that it lacks a convenient representation of the integrated schema at the system level and a sound mathematical basis for data manipulation in an MDBS. To resolve this problem, the concept of hyper-relation, and its use as a powerful and succinct model for the global-level representation of heterogeneous database schemas are also presented. A hyper-relation has the structure of a relation, but its contents are the schemas of semantically equivalent local relations in the databases. The most critical level of optimization is at the execution strategy level. Difficulties of optimization at this level is that each participating database system does not own the information and the mechanism required for converting the data of one database to another to resolve the data-type conflict problems. A major difference between strategies and the past methods is that it does not require the regeneration of the cost models of the participating databases. Hence, it also minimizes the indeterminacy existing in MDBS optimizations.