Self-adaptive automatic data base design

  • Authors:
  • Michael Hammer

  • Affiliations:
  • MIT Laboratory for Computer Science, Cambridge, Massachusetts

  • Venue:
  • AFIPS '77 Proceedings of the June 13-16, 1977, national computer conference
  • Year:
  • 1977

Quantified Score

Hi-index 0.00

Visualization

Abstract

Physical data base design, the selection of organizational structures and access mechanisms for a data base, is one of the most important responsibilities of a Data Base Administrator (DBA). A DBA often has difficulty in performing this task; he lacks the information needed to choose a design that is well matched to the data base's mode of use. This paper presents the design principles of an automatic system that has the ability to choose the physical design for a data base and to adapt this design to changing requirements. The components of such a system include: an information gathering module that collects global statistics on the overall usage pattern of the data base; a predictor that projects observed usage statistics into the future; a design evaluator that computes a figure of merit for any proposed design; and a heuristic proposer that synthesizes a small set of candidate designs for detailed consideration. These principles have been applied to the design of a system that selects secondary indices for an inverted file system.