The Model-Assisted Global Query System for Multiple Databases in Distributed Enterprises

An integrated enterprise information management system must be able to (1) support end usersÕ needs in retrieving information at a global level and (2) facilitate application programs to access, aggregate, and update information maintained in multiple local systems.
Towards query formulation, on-line intelligence empowers the user interface to provide assistance in the articulationas well as allow for high-level, intuitive representation of queries. Specifically, for the articulation of queries, the system would utilize its knowledge on the enterprise information resources (or, metadata; including information models, implementation models and business and operating rules) to alleviate semantic ambiguity, facilitate logical analysis, and enhance adaptive construction during the formulation of queries.

Similarly, the representation itself could accommodate heterogeneity in local models across the enterprise (through, e.g., the knowledge on data equivalence), without having to impose a single view (e.g., an Òintegrated schemaÓ of all databases) for all users to comply by. Making use of the metadata, model-based global query system shown here accomplishes the following:

• query optimization, by decomposing the global query into optimal local queries, taking into account both semantics and performance;
• query translation, i.e., the local queries are then coded in their respective data manipulation languages; and finally
• result integration: results from local systems are assembled to answer the global query.