Browse Prior Art Database

A TestbedApproach to the Design of Computer Communication Network

IP.com Disclosure Number: IPCOM000131536D
Original Publication Date: 1982-Oct-01
Included in the Prior Art Database: 2005-Nov-11
Document File: 13 page(s) / 46K

Publishing Venue

Software Patent Institute

Related People

Kenneth Brayer: AUTHOR [+4]

Abstract

Even ideal simulations can 't predict real performance, but with this method Mitre didn't have to guess. They measured the performance of their routing system on the actual nodes. The traditional approach to discrete digital simulation was to develop a mathematical or statistical model to represent a process, program the model on a large-scale computer, and then execute the model to obtain performance results. The products of a design or simulation effort were a listing of computer code, a document describing the computer code, and a set of results that exemplified the solution of the problem being considered. This approach usually involved considerable programmer effort and cost from $400 to $2000 per hour in computer time on large-scale machines, depending on the time of day. Moreover, the total cost of the computer time in a major simulation effort could even equal or exceed the total cost of human labor in developing the model, exercising it, performing the analysis, and writing the final report. For the last few years, the Mitre Corporation has faced the problem of developing candidate communication and distributed processing systems for possible use by military customers. Historically, we have taken the simulation approach, and on occasion have even simulated computers on computers. However, if at the completion of the effort we were asked what would happen if operating parameters or the computer system design was changed, we have had to make major software changes and then buy more computer time to get performance results.

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

Page 1 of 13

THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.

This record contains textual material that is copyright ©; 1982 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.

A TestbedApproach to the Design of Computer Communication Network

Kenneth Brayer and Valerie Lafleur

The Mitre Corporation

Even ideal simulations can 't predict real performance, but with this method Mitre didn't have to guess. They measured the performance of their routing system on the actual nodes.

The traditional approach to discrete digital simulation was to develop a mathematical or statistical model to represent a process, program the model on a large-scale computer, and then execute the model to obtain performance results. The products of a design or simulation effort were a listing of computer code, a document describing the computer code, and a set of results that exemplified the solution of the problem being considered. This approach usually involved considerable programmer effort and cost from $400 to $2000 per hour in computer time on large-scale machines, depending on the time of day. Moreover, the total cost of the computer time in a major simulation effort could even equal or exceed the total cost of human labor in developing the model, exercising it, performing the analysis, and writing the final report.

For the last few years, the Mitre Corporation has faced the problem of developing candidate communication and distributed processing systems for possible use by military customers. Historically, we have taken the simulation approach, and on occasion have even simulated computers on computers. However, if at the completion of the effort we were asked what would happen if operating parameters or the computer system design was changed, we have had to make major software changes and then buy more computer time to get performance results.

On one computer networking project in 1974, a simulation program was developed for use on an IBM 370/155 computer in simulating a network for 47 processors. Ninety minutes into the first production run, after Mitre had expended $350,000 in staff and computer time, the computer pronounced that the 1.62M bytes then available was insufficient. Management canceled the project. Clearly, another way, the testbed method, had to be found.

The Survivable Communications Group is now working on a different approach that involves simulations on microcomputers. Having carefully examined the relevant cost, the group has concluded that with present microcomputer technology and the availability of languages and language development tools, Mitre can invest the same total funds in staff labor and can purchase computers at a lower cost than that for mainframe time. Because of the natu...