Exploratory analysis of access path length data for a data base management system

  • Authors:
  • D. P. Gaver;S. S. Lavenberg;T. G. Price

  • Affiliations:
  • Naval Postgraduate School, Monterey, California;-;-

  • Venue:
  • IBM Journal of Research and Development
  • Year:
  • 1976

Quantified Score

Hi-index 0.02

Visualization

Abstract

An exploratory approach is taken to analyze a vast quantity of data recorded during the running of the data base management system IMS (Information Management System). The collection of data analyzed is a sequence of access path lengths for a daylong period. The number of segments accessed by IMS when searching a data base in order to retrievea specified segment for a user is called an access path length. Part of the motivation for the analysis is to suggest reasonable stochastic models for the access path length sequence that can be conveniently utilized as input models for a stimulation model of an IMS installation. The exploratory approach taken to the data involves the use of graphical displays and simple numerical summaries to reveal characteristics of, and patterns in, the data. Some simple ways are presented in which the structure of the data revealed by the analysis can be incorporated into an input model for a system simulation.