Object-oriented query processing in a multidatabase environment: Integrating structural and operational approaches

This research has focused on the development of the Donají multidatabase query processing environment. The Donají environment provides a global object-oriented conceptual schema as the front-end interface for submitting global queries expressed using an object-oriented algebra. Data sources in Donají can be structured databases as well as non-database sources with object-oriented interfaces. Additionally, local data sources are not necessarily disjoint and information from data sources may be replicated or related from one source to another. Global queries may require accessing more than one data source and also may require adding information from any data source. Structured data sources such as relational databases are accessed through structural mappings of the local data organizations to the global conceptual schema. Non-database sources are accessed through operational mappings that result in an invocation of operations at the local level.; Results and contributions from this research include the establishment of the Donají semantic framework consisting of an object-oriented metadata model, a global object identity mechanism, and a multidatabase query processor. The metadata model captures descriptions of global and local schemas in a uniform notation. Virtual constructs are incorporated in the metadata model for supporting the integration of related data sources. The global object identity mechanism supports the global representation of local entities and the assembly of semantically related local objects under a unique global identifier. The multidatabase query processor uses an object algebra as the primary query language. Within the algebra, Donají is able to adopt proven transformation rules for distributed queries. Virtual constructs are also defined in the algebra notation and used directly in the simplification and solution of queries. The object algebra has being extended in Donají, by defining a set of meta-operators that permit the use of semantic information about the distribution of data sources during the decomposition of global queries into local subqueries, and during the consolidation of results from local evaluations into global objects. Communication between the semantic framework components is accomplished by a set of mediators that access the information in the metadata model and support the creation and local submission of subqueries.