Adaptive optimization in a database programming language:

Abstract: "Contemporary database research is concerned with the design of new programming languages to support advanced database applications such as software engineering information systems, office systems, and computer-aided design and manufacturing. Glue-Nail is a deductive database system t...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
1. Verfasser: Derr, Marcia A. (VerfasserIn)
Format: Buch
Sprache:English
Veröffentlicht: Stanford, Calif. 1992
Schriftenreihe:Stanford University / Computer Science Department: Report STAN CS 1460
Schlagworte:
Zusammenfassung:Abstract: "Contemporary database research is concerned with the design of new programming languages to support advanced database applications such as software engineering information systems, office systems, and computer-aided design and manufacturing. Glue-Nail is a deductive database system that offers two complementary languages, one declarative and the other procedural, for writing such applications. This dissertation describes the design and implementation of IGlue, an interpreter for the Glue language. It focuses on the problem of query optimization: converting a declarative specification of a query into an efficient plan for executing the query
Glue programs exhibit characteristics, such as a predominance of temporary relations and changing relations, that present problems for conventional query optimizers. This dissertation explores new approaches to optimization that accommodate these characteristics by adapting query plans at run-time. Several techniques are proposed and investigated, including criteria for automatically deciding when to create and drop indexes, strategies for adaptively reoptimizing queries, and a method for dynamically creating pointers between joining tuples. A series of experiments were performed to compare adaptive approaches to conventional ones, as well as to evaluate alternative adaptive strategies
The empirical results of these experiments provide insight into which strategies are the most effective in the IGlue interpreter. Although the adaptive optimization approaches described here were developed in the context of a deductive database system, many of the ideas generalize to other programming languages that manage and manipulate collections of temporary and persistent objects with dynamic properties.
Beschreibung:Stanford, Calif., Univ., Diss.
Beschreibung:XI, 143 S.