“If ever a profession was to be driven by modesty, constant learning and progress, it is [the profession of the performance engineer].” With these words, Fiche and Hebuterne conclude their ambitious work on performance evaluation, a field of the vast discipline known as teletraffic. The authors are well aware that “it is obviously impossible to translate into a single work all the facets of an activity which encompasses many systems and a great variety of domains such as standardization, modeling, measurement, field trials, observations, [and so on].” Nevertheless, they succeed in covering the most relevant topics of this field in some detail. Of course, one can question whether 30 pages are sufficient to teach the necessary statistics, or what can be learned about simulation in only 16 pages.
The book is divided into ten chapters. The first chapter presents the basic concepts of telecommunications networks and traffic. An overview of different network architectures (star, meshed, ring, and tree) and technologies (circuit switched and packet switched) is presented, and their main representatives (fixed telephony, mobile telephony, and data) are briefly described. The differences between offered traffic and handled traffic, between load A and load B, and between user plane and control plane are explained. The concepts of stationarity, busy hour call attempts, and traffic intensity are introduced.
The fundamental notion of quality of service (QoS), and how it is reflected in the standards of the International Telecommunication Union (ITU), is explained in chapter 2. The performance parameters attached to QoS are also described. The authors offer a very useful overview of the relevant standards that were developed in the past ten to 15 years.
After this introduction, the basic theoretical tools are studied: probability theory and its laws in chapters 3 and 4, and the theory of statistics in chapter 5. These theoretical tools are then applied to the different domains of performance evaluation. Chapter 6 presents the major elements of reliability theory, a topic that is often overlooked in other textbooks on performance. Chapter 7 sets out the main issues of queueing theory. Its importance in the field of teletraffic cannot be overestimated.
If the analytical approach to performance evaluation reaches its limit, then simulation often provides the means to overcome difficulties. Chapter 8 presents a short overview of discrete-event simulation. The most difficult task for a performance engineer is to build a model of the real system under investigation. In chapter 10, a set of models is defined for the control and transport plane, which can be applied to various real-life problems. This is very useful for anybody working in an industrial environment. In chapter 10, a performance methodology is described that, when applied to the various phases of equipment and network development, should ensure that performance and QoS objectives are met. This is, again, very useful for the practitioner. Appendices with important mathematical theorems and statistical tables, a short list of references, and an index conclude this fine book.
Most of the content of the book is timeless, and where it is of topical interest, like the engineering of Internet traffic, the book, first published in French in 2003, is up to date. The text gives the reader a good idea of what performance evaluation is, and can serve as a handbook for telecommunications engineers. The style is very terse, though. Many formulas are not derived, and some proofs of important theorems are missing. Therefore, I would not recommend this book for the teaching of performance evaluation in an academic environment.
