Browse Prior Art Database

A HEURISTIC APPROACH TO COMPUTER SYSTEMS PERFORMANCE IMPROVEMENT, I: A FAST PERFORMANCE PREDICTION TOOL

IP.com Disclosure Number: IPCOM000128629D
Original Publication Date: 1974-Dec-31
Included in the Prior Art Database: 2005-Sep-16
Document File: 15 page(s) / 54K

Publishing Venue

Software Patent Institute

Related People

Stephen R. Kimbleton: AUTHOR [+3]

Abstract

The complexity of current computer systems strongly inhibits management usage of analytic modeling techniques in seeking 'good' computer system performance. As a result, heuristically-based approaches are in order. Implementation of such an approach requires: (1) a means for determining the performance of a given computer system processing a given collection of jobs in accord with a specified schedule, (2) a means for achieving an improved system from a given system, and (3) a technique for determining when to stop this iterative process. The objective of this paper is to describe an analytically driven approach to computer system performance prediction which can achieve execution speeds of approximately two orders of magnitude faster than real time for production batch installations while providing detailed information on device utilizations and delays. Thus, it provides a basic required capability for the development of heuristic approaches to computer system performance improvement.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 7% of the total text.

Page 1 of 15

THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.

A HEURISTIC APPROACH TO COMPUTER SYSTEMS PERFORMANCE IMPROVEMENT, I: A FAST PERFORMANCE PREDICTION TOOL

ARPA ORDER 2110. 2223

ISIIRR-74-20 March 1975 Stephen R. Kimbleton A Heuristic Approach to Computer Systems Performance Improvement, I: A Fast Performance Prediction Tool INFORMATION SCIENCES INSTITUTE

46 76 Admiralty WaylMarinadel ReylCalifornia 90291 UNIVERSITY OF SOUTHERN CALIFORNIA (213) 822-1 S 11

Preparation of this paper was supported by the Advanced Research Projects Agency under Contract No. DAHC15 72 C 0308, ARPA Order No. 2223, Pro-gram Code No. 3D30 and 3P10, and the Office of Naval Research, Information Systems Program under Contract N00014-67-A- 0181-0035 (NR 049-311). The views and conclusions contained in this document are those of the author and should not be interpreted as necessarily representing the official policies, either expressed or implied, of the Advanced Research Projects Agency, the Office of Naval Research, or the U.S. Government. This document approved for public release and sale; distribution is unlimited.

ABSTRACT

The complexity of current computer systems strongly inhibits management usage of analytic modeling techniques in seeking 'good' computer system performance. As a result, heuristically- based approaches are in order. Implementation of such an approach requires: (1) a means for determining the performance of a given computer system processing a given collection of jobs in accord with a specified schedule, (2) a means for achieving an improved system from a given system, and (3) a technique for determining when to stop this iterative process. The objective of this paper is to describe an analytically driven approach to computer system performance prediction which can achieve execution speeds of approximately two orders of magnitude faster than real time for production batch installations while providing detailed information on device utilizations and delays. Thus, it provides a basic required capability for the development of heuristic approaches to computer system performance improvement.

KEY WORDS: Performance prediction, simulation, analytic models, system design, performance analysis

CR CATEGORIES: 2.44 3.50 4.305.5

ACKNOWLEDGMENTS

The author would like to express his appreciation to Dr. Thomas E. Bell who critically reviewed an earlier draft of this paper, to Mr. James Cochran who transferred ASIM to the Institute TENEX computer system and executed all experimental runs, to Mr. Nelson Lucas who

University of Southern California Page 1 Dec 31, 1974

Page 2 of 15

A HEURISTIC APPROACH TO COMPUTER SYSTEMS PERFORMANCE IMPROVEMENT, I: A FAST PERFORMANCE

PREDICTION TOOL

prepared the figures, and to Mrs. Linda Tisnado who prepared the report. A condensed version of this report will appear in the Proceedings of the 1975 National Computer Conference. >

INTRODUCTION

Satisfaction of the organizational information processing requireme...