Includes bibliographical references (p. 411-419) and index. + CD-ROM. Ch. 1. About Time! -- Ch. 2. Queueing Theory for Those Who Can't Wait -- Ch. 3. Systems of Queues -- Ch. 4. Distributed Performance Management -- Ch. 5. Commercial Parallelism -- Ch. 6. Parallel Systems -- Ch. 7. Multiprocessor Systems -- Ch. 8. Client-Server Applications -- Ch. 9. World Wide Web Servers -- Ch. 10. Small Numbers, Big Consequences -- Ch. 11. Paths, Potentials, and Probabilities -- Ch. 12. Large Transients in Packet-Switched Networks -- Ch. 13. Large Transients in Circuit-Switched Networks -- Ch. 14. The Dynamics of Scaling -- App. A. PDQ User Guide -- App. B. Professional Organizations -- App. C. Guidelines for Making Multiprocessor Applications Symmetric -- App. D. Glossary. Here is practical new perspective on performance analysisone that makes this difficult subject understandable and useful. It demystifies the complex underlying formulas, and shows how to apply performance analysis to all kinds of large-scale computer systems, including client/server and communications systems. At the heart of this book is a unique methodology called performance-by-design. Timeless and independent of any specific technology, it is a tested approach for ensuring that goals are set and met. The emphasis is on assessing overall performance in large-scale systems, not just tuning one particular hardware subsystem or software component. The chief tool supporting this methodology is the queueing analyzer, the C source code of which is provided on the accompanying diskette. This tool enables you to bypass such arcane concerns as the mathematics of queueing theory, and cut right to the real-world chase. This book will help you master performance analysis and apply it at either end of the development cycle: to support decisions in the early stages of system design, or to evaluate your existing system.