Showing papers by "Philip A. Bernstein published in 2000"

Data warehouse scenarios for model management

Abstract: Model management is a framework for supporting meta-data related applications where models and mappings are manipulated as first class objects using operations such as Match, Merge, ApplyFunction, and Compose. To demonstrate the approach, we show how to use model management in two scenarios related to loading data warehouses. The case study illustrates the value of model management as a methodology for approaching meta-data related problems. It also helps clarify the required semantics of key operations. These detailed scenarios provide evidence that generic model management is useful and, very likely, implementable.

Context-based prefetch – an optimization for implementing objects on relations

Abstract: When implementing persistent objects on a relational database, a major performance issue is prefetching data to minimize the number of round-trips to the database. This is especially hard with navigational applications, since future accesses are unpredictable. We propose the use of the context in which an object is loaded as a predictor of future accesses, where a context can be a stored collection of relationships, a query result, or a complex object. When an object O's state is loaded, similar state for other objects in O's context is prefetched. We present a design for maintaining context and for using it to guide prefetch. We give performance measurements of its implementation in Microsoft Repository, showing up to a 70% reduction in running time. We describe several variations of the optimization: selectively applying the technique based on application and database characteristics, using application-supplied performance hints, using concurrent database queries to support asynchronous prefetch, prefetching across relationship paths, and delayed prefetch to save database round-trips.

Panel: Is Generic Metadata Management Feasible?

Abstract: The database field has worked on metadata-related prob-lems for 30 years. Examples include data translation and migration, schema evolution, database design, schema / ontology integration, XML wrapper generation, data scrubbing and transformation for data warehouses, mes-sage mapping for e-business, and schema-driven web site design. Tools that address these problems are strikingly similar in their design. Arguably, we are making very little progress, since we keep reapplying the same old 1970’s techniques of data translation [9] and views to one new problem after another, without getting much leverage from each succeeding step. Despite all the research on the above tools, we have so far been unable to offer general-purpose database technology that factors out the similar aspects of these tools into generic database infrastructure. This panel addresses the following questions: Is it feasible to develop a generic infrastructure for managing models? If so, what would it need to do, beyond what’s offered in the best object-oriented databases and repositories? Can we devise a useful generic notion of model that treats all popular information structures as speciali-zations (SQL schemas, ER diagrams, XML DTD’s, object-oriented (OO) schemas, web site maps, make scripts, etc.)? Can we produce a generic model manipulation alge-bra that generalizes transformation operations devel-oped for data integration and translation, such as union, match, difference, and merge? What about generic operations on mappings between models, such as invert and compose? What is the role of an expression language that cap-tures the semantics of models and mappings, not only for design but also for run-time execution? Does a generic approach offer any advantages for model manipulation areas of current interest, such as data integration and XML? If the skeptics are right that a generic approach to model management is unachievable pie-in-the-sky, are writers of metadata-driven applications doomed forever to writing special-purpose object-at-a-time code for navigating their information structures? If so, what is the leverage that the database field can offer for these problems?

Data warehouse scenarios for model management

Abstract: Model management is a framework for supporting meta-data related applications where models and mappings are manipulated as first class objects using operations such as Match, Merge, ApplyFunction, and Compose. To demonstrate the approach, we show how to use model management in two scenarios related to loading data warehouses. The case study illustrates the value of model management as a methodology for approaching meta-data related problems. It also helps clarify the required semantics of key operations. These detailed scenarios provide evidence that generic model management is useful and, very likely, implementable.