Communications for cooperating systems Book

Preface:

PREFACE

The past two years, during which this book was written, seem to have had a more rapid evolution of communication technology than I have ever witnessed, in the thirty-five years of my participation in it. Happily, the theme has been the devlopment of very practical commonality in the midst of exploding diversity. My hope is that this work will further that commonality.

System Design Objectives

The system designer of communications networks is pressed to achieve three objectives:

1. Obtain a return on the investments already made in the networks
2. To evolve the network based on today's technical possibilities and economic realities
3. To have a sharp eye on the future trends and possibilities, so as to maximize the future life and value of today's decisions

The wise designer carefully balances among these three sometimes conflicting objectives.

Book Objectives

This book adds to the in-depth technical understanding needed for such network planning and design. The intended audience is experienced systems engineers and other practitioners who are already familiar with basic telecommunications principles. Another intended audience is the senior or graduate student in telecommunications, who has had earlier courses in basic communications.

The book describes the urrent netowrk technologies and the future trends in cooperating systems. Thsis term is used in a broad sense, encompassing what also has been called distributed processing in distributed systems or client-server computing. In general, we refer to all such systems as cooperating systems.

This bookbringsinto focus the rapid evolutions and consolidations that are occurring in a multivendor environment. It describesthe needs for and approaches toa macroarchitecture that incorporates multiple communication architectures. This particularly includes the strategic integration of OSI and SNA (with its extension, APPN), and interoperation with TCP/IP, NetBIOS, and other protocol systems. Major underlying questions are: How do SNA, OSI, TCP, NetBIOS, SAA, and AIX relate? How do they or can they all fit together? How do they relate to the industry-standard Distributing Computing Environment of the Open Software Foundation (OSF)? How do they use the coming broadband facilities? The resulting view is that of a macroarchitecture for heterogeneous systems with anevolutionary thrust toward multiprotocol interoperability and standards.

Any technical base for a network plan mut try to integrate both current and future requirements. This book is therefore based on a vision as to where the industry requirements and technologies will take networks in the coming years. This vision is largely a view of the multivendor world, as seen through the eyes of many IBM professionals, since I have crystallized this view from conversations with over two hundred technical professionals in the various IBM laboratories, as well as the general literature. My goal is to provide an objective, factual report by an observer of a large-scale, evolutionary, communications-development process. The interpretations are, of course, my own, and are not necessarily those of the IBM Corporation.

The aim of this book is to educate, not to train. Specific architectures and products are frequently referred to, to provide concrete examples. Details needed for actual use, however are often omitted to prevent overlead and to conserve space. Some detail, useful for understanding, is put in the end-of-chapter Technical Reference sections. Important concepts and terms are included in the Glossary.

Organization of the Book

Contrary to most prior treatments of communications architecture, the book is organized from the top down--that is, starting with the upper layers and working down, rather than starting at the link level and working up. This top-down approach is taken in the sequence of chapters and within each chapter.

The book is divided into six parts, beginning with a general overview and progressing to a discussion of the four macro-layers of the architecture and a closing chapter.

1. An overview of the evolving architecture, its reasons for existence, the trends that motivate it, and its major structures
2. Common application-services, such as network management and distributed data services
3. End-to-end data-exchange facilities, such as conversation services and remote procedure calls
4. End-to-end transport-service-providers, including end-to-end recovery, flow control, and inter-subnetwork services of OSI, SNA, and TCP/IP
5. Common link and subnetwork-access facilities, such as X.25, LANs, and high-speed subnetworks
6. Current multiple protocol systems illustrations, involving elements such as NetBIOS, SNA, Novell/IPX, X.25, and TCP/IP

At the end are an extensive glossary, general references, cited references, a list of the standards on which the book is based, adn a comprehensive index. There also are the answers to the end-of-chapter exercises. These exercises seek to highlight the key points covered in each chapter.

Some typographic standards have been used. Commands and calls appear in Courier. Parameters appear in lowercase italics. Values of parameters and return codes appear as small capitals.

Some readers will benefit from only the overview chapters and the beginnings of some chapters. Others, in need of more information will benefit from a more complete reading.

Relation To Products

The emphasis is on architecture, rather than products. An architecture incorporates the results of research and standards efforts, so as to provide a consistent, up-to-date structure for products to draw from. Products, in fact, usually implement only a subset of a broader architecture. Products are frequently mentioned in thi book, however, to add realism to the architecture and to illustrate the diversity of interconnections that are feasible. Where specific product information is needed, the reader is strongly urged to seek up-to-date and more detailed information from the appropriate product manuals.

International Standards

A major part of today's evolution of networks is guided by international standards. These are heavily referenced in the text. Again, however, the treatment of standards is educational and is not presumed to be complete or definitive. For actual implementations, the reader must refer to the current standards documentation.

Acknowledgements

The view of communications systems design that is reflected in this book is a composite of many complementary videws, which have been gathered over several years. Prominent among these sources have been the technical staffs at the IBM development laboratories at Raleigh, North Carolina. Significant input has also been obtained from other IBM locations, particularly from IBM laboratories at Austin, Texas; Dallas, Texas; Fujisawa, Japan; Heidelburg, Germany; Hursley; England; La Gaude, France; Palo Alto, California; Rochester, New York; Santa Clara, California; Toronto, Canada; Yorktown, New York; and Zurich, Switzerland. The open literature contributed the rest.

Reluctantly, it is not possible to give adequate credits to the many hundreds of IBM people who have made significant contributions here. It would fill a book in itself. Major insights for this book were, however, contributed by Eric Broockman, George Deaton, Jim Gray, V. Hoberecht, Phillipe Janson, Ellis Miller, and Jerry Mouton, with whom I consulted regularly. The following overall reviewers made valuable suggestions in many chapters: M. Pozefsky, N. Ellis, J. Pickens, and R. Ahlgren. Specific acknowledgements are given in the chapters to the sixty-four persons who privided primary input or reviews for certain chapters. My sincere thanks go to them and to the many other persons who graciously contributed to my understanding.

In general, this textwould not have been possible without the initial encoursagement of Ellen Hancock, vice-president and general manager of communications systems, and the steady support and cooperation of John Hunter, director of communications systems architecture and technology, and Rick McGee, manager of communications systems architecture at IBM.

Finaly, many thanks are due to Betty Cypser who produced the initial artwork for all of the diagrams in this book, using PageMaker. The extensive reviews would not have been possible without that.

Any errors remaining in the text, are, of course, my own responsibility.