The distributed data architecture framework applies to more than just
querying and receiving data from remote repositories in response to a
local query. It also applies to the basic concept behind distributed
file systems and distributed computing.
An example of distributed data management from 1982 comes from IBM:
http://www.research.ibm.com/journal/sj/313/ibmsj3103E.pdf
In this paper, the authors discuss how IBM's Distributed Data
Management architecture, a component of their System Application
Architecture, was designed to allow computing devices to share files
and database contents.
In a 1994 paper, several researchers from the University of
California, Berkeley, discuss their development of Mariposa, an
experimental distributed data management system. In this paper, the
researchers make reference to several database management systems,
such as SIRIUS-DELTA and Distributed INGRES (citations available in
the text) that also used the same architectural approach:
http://cs.chungnam.ac.kr/~ykim/courses/grad-ddb99/papers/icde94-mariposa.pdf
This paper provides a good review of the multiple approaches possible
for distributed data architecture and management.
Both of these works contributed to the thinking that culminated in the
development of the Distributed Relational Database Architecture
Specification published by the Open Group in 1999. The direct link to
the "DRDA, Version 2, Volume 3: Distributed Data Management (DDM)
Architecture" document is:
http://www.opengroup.org/onlinepubs/009608699/toc.pdf
An online example of a distributed data architecture that has
continued to evolve, is the DataFerrett project by the US Census
Bureau and the Bureau of Labor Statistics. Started in 1997, the
project was intended to provide a "generalized search system for
extracting and tabulating data across heterogeneous statistical data
sources":
http://www.thedataweb.org/support/user/chapter01.html
Another example of distributed data management is Vivisimo (
http://www.vivisimo.com ), a meta-search engine. This system takes a
single query, and retrieves results from multiple search engines
simultaneously, clustering the results by concept in real-time to
provide context-sensitive results. This system was launched in 2000,
but is based on research and prototype systems that meet your pre-2000
criteria.
I hope that this provides you with the information you are seeking. If
you require additional information or clarification, please feel free
to post your request and I will do my best to expand on it.
Regards,
aht-ga
---------------------------------------------
Google Search Strategy: "distributed data" management architecture |
Clarification of Answer by
aht-ga
on
11 Aug 2003 22:33 PDT
An additional real-world example would be Wal-Mart. As the
acknowledged masters of supply chain management, Wal-Mart has been
using several variations on distributed data management since the
80's. Their RetailLink interface is both a distributed data
architecture as well as a distributed query system. First launched in
1991 with their larger suppliers, then expanded in 1997 to include any
supplier with an Internet connection, RetailLink allows Wal-Mart's
suppliers to request in real-time inventory levels, sales patterns,
and current demand for the goods that they sell to the retail chain.
The same data, which originates in each store and is also cached and
archived at Wal-Mart's corporate headquarters, feeds Wal-Mart's
business intelligence tools so that the retailer can continually tweak
its product mix and account for regional differences. Their business
intelligence systems also pull in live data from external data sources
such as weather, economic indicators, and local events to drive
business decisions.
You can read more information in the following paper:
http://www.napmhou.org/Supply%20Chain%20Management%20Research%20Paper.doc
|