The metamorphosis of information management
On the high plateau of modern computer hardware, software systems are arising that are more than fast data processors or glorified adding machines-they are information refineries that can transform mere facts into knowledge on a vast scale. For instance, information-refining programs might transform the low-level data that described a hospital patient, transportation network or factory into a high-level synopsis. They might convert electronic file cabinets full of data into authoritative discussions of the objects or events (patient histories, wildflowers or automobile accidents) contained in those files. The development of information refineries is intimately bound up with the growing role of parallelism in computer science. Parallel hardware consisting of multiple subcomputers in a single box has increased the computing power available to users. Parallel software makes it possible to tap this power, and frequently it offers elegant solutions to complex information-management problems. The author and investigators elsewhere have been exploring a number of approaches to building such information refineries. Two kinds of machine hold promise: the information filter, which transforms an incoming stream of data into higher-level knowledge, and the smart data base, which sorts out interesting patterns from records of many similar objects or event.
- OSTI ID:
- 6913153
- Journal Information:
- Scientific American; (USA), Vol. 261:2; ISSN 0036-8733
- Country of Publication:
- United States
- Language:
- English
Similar Records
Data Locality Enhancement of Dynamic Simulations for Exascale Computing (Final Report)
Data Processing for the SNS EQ-SANS Diffractometer
Related Subjects
DATA BASE MANAGEMENT
TECHNOLOGY ASSESSMENT
PARALLEL PROCESSING
ARRAY PROCESSORS
COMPUTER ARCHITECTURE
DATA PROCESSING
INFORMATION SYSTEMS
INFORMATION THEORY
PROGRAMMING
PROGRAMMING LANGUAGES
MANAGEMENT
PROCESSING
990300* - Information Handling
990200 - Mathematics & Computers