On the feasibility of implementing a large relational data base with optimal performance on a minicomputer

  • Authors:
  • Glenn K. Manacher

  • Affiliations:
  • University of Illinois at Chicago, Chicago, Illinois

  • Venue:
  • VLDB '75 Proceedings of the 1st International Conference on Very Large Data Bases
  • Year:
  • 1975

Quantified Score

Hi-index 0.00

Visualization

Abstract

A less general model of Codd's relational model is proposed, for which queries and updates are maximally efficient. The differences in access method from Codd's are minor enough so that almost all the compactness of expression and quality of data independence remains. The first new mechanism is an "algebra" of dependencies superimposed on the set of relations. These dependencies, called "subsumptions", permit the data base designer to impose a structure on the set of relations, thereby joining them together into a shape reflecting the overall organization of the data. The second new mechanism is a new format in which fields are ordered, and lexical ordering of key values is enforced. It is shown that this format, used in conjunction with certain specific kinds of subsumptions, gives rise to a query sequence that is natural, linear, and intuitively related to access efficiency. Finally, an alternative notion of "Normal form" is proposed for which the overhead for maintaining the form is truly minimal, and a set of efficient basic operations is proposed.