Book Review - The Internet Protocol Journal - Volume 1, No. 1

GroupWare: Collaborative Strategies for Corporate LANs and Intranets , by David Coleman, ISBN 0-13-727728-8, Prentice-Hall PTR, 1997,

Some areas of science provide very poor training for dealing with primarily human processes. One might think that packet switching would be an exception because it lives on the stochastic nature of bursty communications. Because our knowledge of human and group activity is, at best, characterized by statistical assessments, those working in networking should do well in understanding and dealing with the unpredictable and human nature of communication, especially when it involves using networks.

So much for theory. In general, the world of lower-level networking has done little for the upper strata of computer-mediated human communication, except to provide a platform for the work of others. An apparent exception in the world of Internet technology is e-mail, yet it actually serves more as proof of the problem than as an exception. The basic facilities in Internet e-mail are the same today as they were 25 years ago. As nice as they are, the word "basic" is essential when characterizing them. Almost none of the Internet's standardized e-mail facilities are really targeted at providing automated or structural support for the work of a group.

GroupWare Defined
The collection of products and services designed to help people collaborate via computer, by direct interaction, or by information dissemination is called "GroupWare." Coleman's book is a revision of GroupWare: Technology and Applications. Written only 15 months earlier, the world changed more than enough in that time to require the revision. The first book had relatively little to say about the Internet, whereas this new book tries mightily to factor it into the equation. The result is a bit erratic, but the digressions serve to highlight how rapidly things are changing, rather than to suggest looking elsewhere for a better source on the topic.

The new book has an entirely different subtitle, giving a reasonable sense that the content targets more an understanding of system organization and function than detailed technical explanation. That's just fine, because the book really is not particularly technical. It covers the requirements and functions for supporting activity by groups.

Downsizing and working remotely are two very strong driving forces for increased use of GroupWare. This book is essentially an introduction to concepts, functionality, and use of systems that attempt to help staff members work together. Oddly, that does not only mean working together when physically separated, because there is discussion of meeting room assistance, such as with automated sense-of- the-group tallying devices.

The first two chapters introduce the topic, emphasizing that human and group process concerns dominate the field and are intimately tied to the aggressive efforts that organizations are making to run more productively and, frequently, with fewer people. The third chapter discusses functionality in terms of the World Wide Web. The book reflects the current enthusiasm for the Web, sometimes to the detriment of the appropriate use of messaging technology, although messaging is more prevalent among GroupWare than other kinds of commercial Internet systems.

The realm of GroupWare does not have a firm taxonomy. My own synthesis includes: Message (text and document) Exchange, Forms Exchange, Calendaring & Scheduling, Workflow, Presentations and Interactive Meetings, and Document Development and Sharing. The next six chapters cover the functional pieces of this GroupWare realm.

The next five chapters cover the major vendors of integrated GroupWare products: Lotus Notes, Novel GroupWise, TeamWARE, Hewlett-Packard, and Oracle Interoffice. HP's chapter discusses "strategy," suggesting the lack of a well-integrated product suite, but one more survey of the terrain is nonetheless useful. And that, perhaps, is the major reason for reading this book: It constantly emphasizes the human and process-oriented aspect of organizational behavior and the need to attend carefully both to the needs of the humans and the nature of the processes. It is easy to understand that an improper travel authorization, will bring an organization to its knees. It is easy to forget that the system is used by humans who well might not want the added complexity or rigidity of the system and who, therefore, must be part of the design and adoption effort. In my opinion, the book takes a rather more negative view about GroupWare acceptability than is necessary, but then I like such technology, and the average worker in the average organization does not.

The last six chapters of this book intermix case studies and Hahn, of Collabra and Netscape, points the reader to Chapter 17, "GroupWare & Reengineering: The Human Side of Change." Although one of the better considerations of these issues in the book, it is far from the only one.

A Useful Survey
If you have little familiarity with these "upper level application" areas of networking, the functionality, products, or use, then this book is a good one to read. You will not learn much about the underlying technology, nor will you be able to qualify as a "certified GroupWare support engineer," but you will obtain an extremely useful survey of the field, and you will obtain it from the perspective of human and organization use. As the Internet moves into the mass market, that perspective is a good one.

Dave Crocker
Brandenburg Consulting

High-Speed Networks
High-Speed Networks: TCP/IP and ATM Design Principles, by William Stallings, ISBN 0-13-525965-7 Prentice-Hall, 1997.

High-speed networks now dominate both the WAN and LAN markets. In the WAN market, data networks have evolved from packet-switching networks to ATM networks operating at 155 Mbps or more. In the LAN market, the staple 10-Mbps Ethernet is being replaced with 100-Mbps Fast Ethernet, Gigabit Ethernet, and even Asynchronous Transfer Mode (ATM) LANs. This book provides a survey of high-speed networks and the design issues related to them. Much of the book is devoted to the study of various techniques aimed at reducing network congestion.

The book is divided into seven sections. The first section deals with the fundamentals: TCP/IP principles; packet switching and Frame Relay networks; and internetworking principles. The second section provides an overview of ATM and Fast and Gigabit Ethernet. These two sections can easily be torn out of the book and serve as an excellent primer on today's modern networks. I am going to recommend to my employer that they be made mandatory reading.

In the third section of the book, Stallings focuses on one treatment of queuing theory, namely, how it is applied to modeling network behavior. Stallings has an undeniable gift for taking large complicated subjects and teaching the fundamentals, and then some, without belittling the subject at hand or the reader. This book is witness to this gift, and this chapter but one fine example. But once the reader has an understanding of queuing theory, Stallings throws a wrench in the gears. The chapter on self-similarity explains why traditional queuing models are inadequate when trying to predict the performance of Ethernet traffic and other self-similar streams. While this section is by far the most theoretical, it is at the same time necessary for the reader's understanding of network performance, and while many readers may not care to devote the time necessary to gain a complete understanding of self-similarity, astute students are urged to invest in more than a simple gloss-over of this section.

Having understood the basics of self-similarity, I hoped the fifth section of the book, on network traffic management, would be addressed with greater emphasis on delivering quality of service and the problems related to self-similarity. Instead, the material is based on traditional queuing models.

The fourth section, flow control, is divided into two categories. The first, link control mechanisms, focuses on some of the performance issues related to the use of Automatic Repeat Request (ARQ) link control protocols. The second category, transport control mechanisms, concentrates on the TCP flow control mechanism. I expected to find references to bugs in some TCP implementations exposed by high-volume WWW servers, but didn't. Stallings goes on to present an overview of some of the performance issues of TCP over ATM. As institutions begin upgrading their networks, this issue is sure to receive a great deal of interest. The section concludes with a look at the Real-Time Transport Protocol , another area sure to spark attention as the need to move large multimedia data across WANs, in real time, becomes more relevant.

The sixth section of the book covers Internet routing protocols and opens with a primer on graph theory. Four routing protocols (RIP, OSPF, BGP, and IDRP) are covered. The section concludes with a discussion of multicasting as an introduction to RSVP. This section sparked my curiosity enough to call for a visit to the WWW site for RSVP development.

Stallings shies away from directly addressing application-driven improvements aimed at increasing network performance. In today's Web/ CGI-driven world, I would expect this to be a topic of interest to many. Perhaps this is a subject for another book. But the topic is not entirely avoided. The last section of the book focuses on various lossless and lossy compression techniques. The quirkiness of material covered makes this section a darling.

This book rates an A+. Unlike most books about computers being published today, this book is neither superficial nor is it insulting to the reader. It is intended for both professional and academic audiences. Stallings' desire to truly educate is apparent. This is not a book about promoting the hype, this is a book about serious learning.

Neophytos Iacovou,
University of Minnesota
Academic & Distributed Computing Services