A Relational Database Machine for Very Large Information Retrieval Systems
A relation is divided into several parts and stored in general purpose magnetic disks. A parallel access method and a clustering technique based on the extended K-d tree are employed to shorten data transfer time from disks to the main memory.
Selection operations are performed at the time of the data transfer by specialized hardware attached to a disk controller. A table based on the finite state machine is used to perform text search of selection operations.
The three-phase (filtering, sorting and comparing) join method is used to accelerate join operations. Each phase is performed in parallel by specialized hardware.
The performance of ADAM is two orders higher than that of conventional general purpose computers when applied to very large information retrieval systems. The cost performance is also one order of magnitude better.
KeywordsMain Memory Finite State Machine Random Access Memory Selection Operation Text Search
Unable to display preview. Download preview PDF.
- G.G. Langdon, Jr. et al., “Special Issue on Database Machines”, IEEE Trans. Comput., Vol.C-28 No.6, June 1979Google Scholar
- S. Uemura, T. Yuba, A. Kokubu, R. Ohomote and Y. Sugawara, “The Design and Implementation of a Magnetic-Bubble Database Machine”, Proc. IFIP 80, Oct. 1980, pp.433-438Google Scholar
- R. Epstein and P. Hawthorn, “Design Decisions for the Intelligent Database Machine”, Proc. NCC 80, May 1980, pp.237-241Google Scholar
- Y. Tanaka, “Bit-Sliced VLSI Algorithms for Search and Sort”, Proc. 10th VLDB, Aug. 1984, pp.225-234Google Scholar
- K. Kannan, “The Design of a Mass Memory for a Database Computer”, Proc. 5th Comp. Arch., Apr. 1978, pp.44-51Google Scholar
- T. Lang, E. Nahouraii, K. Kasuga and E.B. Fernandez, “An Architectural Extension for a Large Database System Incorporating a Processor for Disk Search”, Proc. 3rd VLDB, Oct. 1977, pp.204-210Google Scholar
- H.O. Leilich, G. Stiege and H. Ch. Zeidler, “A Search Processor for Data Base Management Systems”, Proc. 4th VLDB, Sept. 1978, pp.280-287Google Scholar
- S. Fushimi, M. Kitsuregawa, H. Tanaka and T. Moto-oka, “Multidimensional Clustering Technique for Large Relational Database Machines”, Proc. Foundations of Data Organization, May 1985, pp.226-235Google Scholar
- A. Mukhopadhyay, “Hardware Algorithms for Nonnumeric Computation”, Proc. 5th Comp. Arch., Apr. 1978, pp.8-16Google Scholar
- L.A. Hollaar and D.C. Roberts, “Current Research into Specialized Processors for Text Information Retrieval”, Proc. 4th VLDB, Sept. 1978, pp.270-279Google Scholar
- D.R. McGregor, R.G. Thomson and W.N. Dawson, “High Performance Hardware for Database Systems”, in “Systems for Large Data Bases” P.C. Lockemann and E.J. Neuhold(Eds.), North-Holland, 1976, pp.103-116Google Scholar