Résumé
This document is not meant to reinvent the wheel, it mostly is a compilationof database tuning concepts, howevere with a persona! touch. It is based
upon considerations made by various writers, such as [Hainaut 1986],
[O'Neill 1994], [Date 1990], [Finkelstein 1988] and [Elmasri 1994]. It gives
an overview of possible parameters aimed to physically tune a database.
Moreover, it concerns with the problem of index selection. Chapter 1, introduces the process of physical database tuning within the process of data modeling. It reveals the pitfalls of selecting the appropriated indexes. Chapter 2, describes the data operations and deals with execution methods for queries and joining tables. It abstracts the data access operations into a small set of easy to understand and to analyze query types. Chapter 3, deals with physical data allocation parameters and access structures. It lists a set of parameters that might be helpful during database allocation. It describes and evaluates various data access structures, such as B-Trees, Clusters and Hash indexes. Chapter 4, tries to consolidate chapter 2 and chapter 3 into a small set of I/0 cost relations. It lists a set of relation that might be used to determine rapidly I/0 costs for a given query type and access structure. Chapter 5, is based upon a study made by [Finkelstein 1988] to implement a physical design tool for relational database. It describes a methodology for physical database tuning and lists some tuning guidelines and heuristics used to reduce query execution time. Chapter 6, gives a practical overview off various considerations that might arise during the quest for the optimal index solution. However, we will not pretend the case study to be exhaustive, as we start with a limited set of data entities, requirements and queries. This document is aimed to introduce, help and guide the database designer in its first attempts of database tuning.
| la date de réponse | 1996 |
|---|---|
| langue originale | Français |
| L'institution diplômante |
|
| Superviseur | Jean-Luc Hainaut (Jury) |