Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

System Architecture for Distributed Data Management

IP.com Disclosure Number: IPCOM000131222D
Original Publication Date: 1978-Jan-01
Included in the Prior Art Database: 2005-Nov-10
Document File: 12 page(s) / 46K

Publishing Venue

Software Patent Institute

Related People

Richard Peebles: AUTHOR [+4]

Abstract

Successful implementation of most distributed processing systems hinges on solutions to the problems of data mangement, some of which arise directly from the nature of distributed architecture, while others carry over from centralized systems, acquiring new importance in their broadened environment. Numerous solutions have been proposed for the most important of these problems. In a distributed computer system, multiple computers are logically and physically interconnected over ";thin- wire"; /low bandwidth) channels and cooperate under decentralized system-wide control to execute application programs. Examples of thinwire systems are Arpanet, the packet-switched network of the U.S. Defense Communications Agency, and Mininet, a transaction-oriented research network being developed at the University of Waterloo. These may be contrasted with high-bandwidth or ";thick-wire"; multiprocessor architectures, such as the Honeywell 6080 or the Pluribus IMP. A practical consequence of thin-wire design is that processing control is in multiple centers. No one processor can coordinate the others; all must cooperate in harmony as a community of equals. The key issue is that interprocess communication is at least an order of magnitude slower when the communicating tasks are in separate computers than it is when they are executing in the same machine. Therefore, no single process can learn the global state of the entire system nor issue control commands quickly enough for efficient operation, so that multiple centers of control are implied.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 8% of the total text.

Page 1 of 12

THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.

This record contains textual material that is copyright ©; 1978 by the Institute of Electrical and Electronics Engineers, Inc. All rights reserved. Contact the IEEE Computer Society http://www.computer.org/ (714-821-8380) for copies of the complete work that was the source of this textual material and for all use beyond that as a record from the SPI Database.

System Architecture for Distributed Data Management

Richard Peebles * and Eric Manning

University of Waterloo, Ontario, Canada

1

Richard Peebles * and Eric Manning University of Waterloo, Ontario, Canada

2

Introduction Successful implementation of most distributed processing systems hinges on solutions to the problems of data mangement, some of which arise directly from the nature of distributed architecture, while others carry over from centralized systems, acquiring new importance in their broadened environment. Numerous solutions have been proposed for the most important of these problems.

In a distributed computer system, multiple computers are logically and physically interconnected over "thin- wire" /low bandwidth) channels and cooperate under decentralized system-wide control to execute application programs. Examples of thinwire systems are Arpanet, the packet- switched network of the U.S. Defense Communications Agency, and Mininet, a transaction- oriented research network being developed at the University of Waterloo. These may be contrasted with high-bandwidth or "thick-wire" multiprocessor architectures, such as the Honeywell 6080 or the Pluribus IMP. A practical consequence of thin-wire design is that processing control is in multiple centers. No one processor can coordinate the others; all must cooperate in harmony as a community of equals.

The key issue is that interprocess communication is at least an order of magnitude slower when the communicating tasks are in separate computers than it is when they are executing in the same machine. Therefore, no single process can learn the global state of the entire system nor issue control commands quickly enough for efficient operation, so that multiple centers of control are implied.

This definition does not imply that the computers are geographically distributed. On the contrary, the machines might all be located in one room; they could be considered a loosely coupled multiprocessor rather than a network. The choice of machine location is a result of economic and political considerations; that is, the distributed system architecture is amenable to both single-site and widespread machine location.

The debate between centralization and decentralization has thus taken something of a new twist. The old arguments for centralized systems were based on economies of scale in computer architecture, simplification of computer center operation, and the need for an integrated data base. The arguments for economies of scale are highly questionable. Arguments for simpli...