Computer Capacity Planning Theory and Practice Shui F Lam Department of Mathematics and Computer Science California State University Long Beach, California K. Hung Chan Graduate School of Management University of California, Riverside Riverside, California ACADEMIC PRESS, INC. Harcourt Brace Jovanovich, Publishers Boston Orlando San Diego New York Austin London Sydney Tokyo Toronto Copyright © 1987 by Academic Press, Inc. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Academic Press, Inc. Orlando, Florida 32887 United Kingdom Edition published by ACADEMIC PRESS INC. (LONDON) LTD. 24-28 Oval Road, London NW1 7DX Library of Congress Cataloging-in-Publication Data Lam, Shui F. Computer capacity planning. Bibliography: p. Includes index 1. Computer capacity—Planning. I. Chan, K. Hung. II. Title. QA76.9.C63L36 1987 004.2'l 86-32220 ISBN 0-12-434430-5 (alk. paper) 87 88 89 90 9 8 7 6 5 4 3 2 1 Printed in the United States of America To our Parents, and Jo May and Jo Kay ACKNOWLEDGMENTS The authors wish to express appreciation to the many colleagues and professional acquaintances who have made helpful comments on this monograph. We are especially indebted to James Pick of the University of California, Riverside, Art Gittleman of the California State University, Long Beach, Ronald Reed and William Denton of Home Savings of America for their detailed reviews and insightful suggestions for improvements. We have also received invaluable assistance and ideas from Tsung-Jen Chiu, Apichart Karoonkornsakul, HeeKyung Sung, and Chris­ tina K.Y. Yau. An important element in the execution and completion of research project of this kind is the environment in which the effort took place. We thank Karl Anatol and Charles Austin of the California State University, Long Beach and Stepan Karamardian of the University of California, Riverside for their roles in providing this environment. Finally, to those hundreds of individuals who responded to our question­ naire survey, we are particularly grateful. We hope that the results of this research will be helpful to the survey respondents and others in the field of computer capacity planning. Any shortcomings of this manuscript are the responsibility of the authors. January 1987 Shui F. Lam K. Hung Chan vii COMPUTER CAPACITY PLANNING - THEORY AND PRACTICE PREFACE Increased computerization in business data processing has led to greater reliance on computers in business operations and growing data processing budgets. Many computer-related matters have become major management concerns. Satisfactory computer services depend greatly on the choice of configurations and the amount of capacity available on the computer system. To monitor and project computer workload and to plan for changes or expansions of computer configurations to meet future demand in a cost-effective manner is generally referred to as computer capacity planning. Because upgrading computer facilities often involves a large sum of money, improvements in capacity planning can mean sub­ stantial cost savings to the company. This study examines the theory and practices of computer capacity planning in a more rigorous manner than previous studies. The first part of this study summarizes and discusses the theories and techniques presented in recent literature on computer capacity planning. The second part focuses on examining the degree of applicability of the theories and techniques and how different organizations may adopt dif- xiii xiv Preface ferent capacity planning methods. It also identifies ways to improve the applicability of the theory and the quality of the practice on com­ puter capacity planning. A questionnaire was used to collect the necessary empirical evidences for this study. The approach taken in this monograph should make it a useful reference for a wide range of readers. Instead of using a "cookbook" style showing simply how to do capacity planning, we emphasize on as­ sessing the usefulness of the various approaches and techniques in capacity planning. The information provided in this monograph should be of interest to analysts and managers who are faced with the practical reality of computer capacity planning. It should broaden their under­ standing of this function by knowing how other companies carry out this function and what techniques are available in the literature that can be used to improve results. As management advisers, internal and external auditors as well as other quality assurance personnel should also obtain insight on how to measure, evaluate, and improve performance of their own or their clients' computer facilities. Equipment manufacturers and software developers should find the results of this study useful as they provide direct consumer information that will help guide the development of their new machines or new software packages. An essential way to assess consumer behaviors is to contact the consumers directly. As far as we can ascertain, this is the first study on computer capacity plan­ ning that directly solicits and analyzes a wide range of consumer infor­ mation in a systematic and scientific manner. Researchers may use the survey results to assess the future direc­ tion of research in computer capacity planning as these results provide Preface xv insight into the degree of applicability of the various techniques under different business environments. Finally, for students in computer science, information management and accounting/auditing, we offer a con­ cise description of the principles of computer capacity planning. The data and analysis of the practices enable students to better appreciate the usefulness of the capacity planning techniques in different circumstances. LIST OF FIGURES Figure 2.1 A Queueing Network Model with a Terminal 61 -Driven Workload (in Section 2.8.3). Figure 3.1 Sample Workload Graph (in Section 3.2.1). 74 Figure 3.2 Workload Versus Capacity (in Section 81 3.2.3). Figure 3.3 Workload Forecasts Assuming Steady Growth 113 (in Section 3.3.11). Figure 3.4 A Queueing Network Model for Three 115 Workload Classes (in Section 3.3.11). xvii LIST OF TABLES Table 1.1 Increase in Sales of the Ten Largest Computer 4 Companies (in Section 1.1). Table 3.1 Combining Interactive and Batch Components into 116 One Class (in Section 3.3.11). Table 3.2 Measurement for the Database and Data Entry 117 Workload (in Section 3.3.11). Table 3.3 Notations and Formulae for a Mixed Queuing 119 Network Model (in Section 3.3.11) Table 3.4 Performance Measures Calculated from the 124 Queueing Network Model (in Section 3.3.11). Table 4.1 Usable Responses to the Questionnaire Survey 141 on Computer Capacity Planning Practices (in Section 4.3). Table 4.2 Overall Ratings for Questions 2 to 8 in the 143 Questionnaire Survey (in Section 4.4). Table 4.3 Analysis of Variances for Overall Mean 144 Ratings (in Section 4.4). Table 4.4 Multiple Comparisons Using the Ryan-Einot- 145 Gabriel-Welsch F Test (in Section 4.4). Table 5.1 Breakdown Among Respondents Based on 168 Different Environmental Characteristics (in Section 5.2). Table 5.2 Mean Ratings by Group Based on the Four Most 174 Significant Environmental Factors (in Section 5.4). Table 5.3 F Test (One-Way Analysis of Variance)for the 176 Difference in Mean Ratings (in Section 5.4). Table 5.4 Specific Effects of the Four Most Significant 178 Environmental Factors on the Five Critical Capacity Planning Issues (in Section 5.4). Table 5.5 Mean Ratings by Industry (in Section 5.5). 190 XIX CHAPTER 1 ISSUES AND OVERVIEW 1.1 The Significance of Computer Capacity Planning Electronic computing is a relatively new industry. In a history of thirty some years, electronic data processing has been ingrained into many aspects of people's daily life and certainly has established a strong foothold in business and non-business organizations. With more and more data stored electronically, more operations under computer guidance or control, and more users relying on computers to carry out their functions, it is inconceivable for many companies to be able to continue normal operations without the computer. The availability, reliability, serviceability, ease of use, rate of response, and other computer related matters have become major management concerns. As more operations are being computerized and more users are gaining access to the computer through their terminals, computer functions are very visible and computer performance is under close scrutiny. Data process­ ing departments are often expected to provide satisfactory services in the most cost-effective manner. 1