Some thoughts on the future of associative memories/processors in the solution of data base management problems

  • Authors:
  • P. Bruce Berra

  • Affiliations:
  • -

  • Venue:
  • SIGFIDET '74 Proceedings of the 1974 ACM SIGFIDET (now SIGMOD) workshop on Data description, access and control
  • Year:
  • 1974

Quantified Score

Hi-index 0.01

Visualization

Abstract

Associative memories and processors have been discussed in the literature for the past 15 years and a small number of hardware devices have actually been built (21). A number of applications have been considered and include air traffic control (22), computer graphics (24, 25), information retrieval (9), numerical analysis (15), networks (18, 19) and among others, data base management (2, 4-8, 10, 11, 13, 17). A significant percentage of present computer resources are expended in the processing of large data bases. Of this, a great deal is wasted by processing largely non-sequential data by sequential means. With hardware costs going down; and software and personnel costs going up it is important to look at the impact of non-sequential computer hardware on data base management (DBM). Based upon previous research in the application of associative processing devices to DBM, it appears that the opportunity exists for increased efficiencies by utilizing these devices in conjunction with sequential computers. In this paper, some of the advantages and limitations of these devices are presented and some brief thoughts are given concerning this author's view of the future. It is assumed here that the reader has a working knowledge of associative processors and DBM.