Preface

This text aims to provide an intelligent near-beginner (as far as OSI is concerned) with an understanding of Open Systems Interconnection (OSI). Some previous acquaintance with data communications as presented in the many text books on that broad subject would be useful. The book is aimed at the reader who is curious enough to ask: "Why is it that way? What advantages does that approach give? Might there be other or better ways?"

This text is not an exposition of the technical detail of the OSI Standards. Rather it aims to explain why OSI is the shape it is, and to guide the reader in a critical examination of the OSI approach to specifying rules for computer communication (computer protocols). The text should be particularly valuable for those who are newly moving into positions where they are a part of a team developing applications using OSI, either in the International Standards' work or for their own firm. The text would also be useful for those sections of undergraduate and taught masters' courses that are dealing with OSI, either as the main text or as follow-on reading.

Much of the material of necessity represents personal perceptions and reasoning, as the real reasons for approaches and choices are rarely presented in International Standards or CCITT/ITU-T Recommendations (the primary definitive documents on OSI). The main purpose of ISO Standards and CCITT/ITU-T Recommendations is to present clearly the protocol to be implemented, not to explain the reasons for the choices. Frequently such reasons are buried in old working documents, maybe even only in private or national papers rather than being recorded in official international documents. In some cases reasons are merely in the heads of early workers, and are perhaps not even well articulated. It can also happen that earlier non-OSI protocols provided the basis for the OSI work, and reasons and rationale at the OSI level are simply "because that is the way it was done in xyz", and the search for real reasons has to go back a level. Nonetheless, I have been active in both progressing the OSI work and in presenting it at conferences and seminars for close on two decades, and the presentation in this text is believed to be a fair one.

The text commences with discussion of the organizations involved in OSI, and with the OSI architecture - the seven-layer model and related concepts, then goes on to consider each of the layers in turn (some in more detail than others), and ends with a single Chapter discussing the architecturally interesting features of a (necessarily) limited sample of application layer protocols.

The text does not attempt to give a complete treatment of OSI application layer standards: the choice of subjects to discuss has been made on the basis of whether the application concerned raises interesting or difficult new concepts, or helps to illustrate features discussed earlier in the text. In other words, the selection of applications treated is based on whether a presentation of a particular application is relevant to a general understanding of OSI and how to produce other OSI applications, rather than on whether the application is considered one of the more important in the market-place.

Most people reading this text will probably have at least heard of X.25, X.400, X.500, and the 7-layer model. This is not assumed in the main exposition, but examples and illustrations occur from time to time that will be less meaningful without some background knowledge of these areas. Such examples and illustrations can be skipped without much loss.

John Larmouth

1994