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

Control of Distributed Processes

IP.com Disclosure Number: IPCOM000131335D
Original Publication Date: 1978-Jul-01
Included in the Prior Art Database: 2005-Nov-10
Document File: 11 page(s) / 40K

Publishing Venue

Software Patent Institute

Related People

Harold S. Stone: AUTHOR [+4]

Abstract

Cost-effective control of distributed computing systems requires practical models for reassigning program modules among the processors in real time. The introduction of the microprocessor has made distributed processing an increasingly popular notion in the computer ndustry. Economies of fabrication have substantially reduced the cost of replicating processors in a system, making multipleprocessor systems economically attractive. Concurrent with the hardware developments in microprocessors has been the development of data base management systems with sophisticated techniques for sharing data. These two major trends in the computer industry -- microprocessors and large shared data bases -- have created an environment that is fostering the growth of distributed computation. The idea, of course, is that each user has the local processing power that he requires in the form of a private dedicated computer. He can access a central bank of data when necessary for either current information or peak computational power, when these needs cannot be met by his private computer. While the idea behind distributed computation is tantalizing, there are serious practical and theoretical problems that must be solved to realize the potential of the idea. In a practical sense, the communication links in a distributed network are inherent bottlenecks. Typical bandwidths available on high-speed communication links are still a factor of 10 to 100 lower than memory bandwidths. Protocol and error protection further reduce the bandwidth available to the user. In the face of high penalties for communication, the practical solution for using a distributed system is to do as little communication between processors as possible, relegating such communication to file transfers before and after major blocks of computation are run. Future data links, using fiber optics and broadband communications satellites, may remove the bottlenecks we see now. At that time the communication bottleneck will all but disappear, so that distributed processing can take on a new character, with interprocessor accesses becoming frequent and beneficial.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 10% of the total text.

Page 1 of 11

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.

Control of Distributed Processes

Harold S. Stone

Shahid II Bokhari

University of Massachusetts

Cost-effective control of distributed computing systems requires practical models for reassigning program modules among the processors in real time.

The introduction of the microprocessor has made distributed processing an increasingly popular notion in the computer ndustry. Economies of fabrication have substantially reduced the cost of replicating processors in a system, making multipleprocessor systems economically attractive. Concurrent with the hardware developments in microprocessors has been the development of data base management systems with sophisticated techniques for sharing data. These two major trends in the computer industry -- microprocessors and large shared data bases -- have created an environment that is fostering the growth of distributed computation. The idea, of course, is that each user has the local processing power that he requires in the form of a private dedicated computer. He can access a central bank of data when necessary for either current information or peak computational power, when these needs cannot be met by his private computer.

While the idea behind distributed computation is tantalizing, there are serious practical and theoretical problems that must be solved to realize the potential of the idea. In a practical sense, the communication links in a distributed network are inherent bottlenecks. Typical bandwidths available on high-speed communication links are still a factor of 10 to 100 lower than memory bandwidths. Protocol and error protection further reduce the bandwidth available to the user. In the face of high penalties for communication, the practical solution for using a distributed system is to do as little communication between processors as possible, relegating such communication to file transfers before and after major blocks of computation are run. Future data links, using fiber optics and broadband communications satellites, may remove the bottlenecks we see now. At that time the communication bottleneck will all but disappear, so that distributed processing can take on a new character, with interprocessor accesses becoming frequent and beneficial.

On the theoretical side of the issue is the fact that the behavior of programs in a distributed environment is not well understood. We are today in a situation analogous to the development of virtual memory systems in the middle 1960's. At that time the promise of virtual memory systems was quite clear, but the early implem...