Logo What's New

Let our core competencies complement yours...

 Company  About Us  Services  Technology  News  Contact Us  Search

Practical Computer Network Analysis and Design

James McCabe, author
Morgan Kaufman Publishers, ISBN 1-55860-498-7

Foreword by
David M. Piscitello
Core Competence, Inc.

Networks are nearly as prolific today as PCs were 10 to 15 years ago. Businesses of every size, service, and marketplace have come to embrace networking as a means of sharing resources, disseminating  information, advertising, conducting commerce, and extending their communities of interest through collaboration and joint venture. Elementary and secondary education systems are as networked today as many colleges and universities were as recently as five years ago. 

Nearly every transaction engaged in over the course of ones daily routine involves networking or can. You can bank, work, shop, gather news, invest, plan a vacation, and be entertained using networks. Networks, it seems, are everywhere.

Several factors contribute to the apparent ubiquity of networks. Personal computers are now business appliances and are fast becoming home appliances. The hardware costs of local area network interfaces and "hubbed" network connections is a small fraction of what it was only five years ago.  And the skill and software required to install and operate a small- to modest-size PC LAN is readily acquired. The design of such networks is, in most cases, straightforward.

Local area networks, it seems, are organic.  As an organization grows, so does its network. Through processes that are characteristically similar to cell meiosis and mitosis, a network segments or combines with other networks, changing its shape, composition, and complexity. The process, however, is often more complex and seemingly chaotic than cell union or division. The resulting networks are not always exact regenerations of the original. For in addition to changes in physical characteristics (e.g., CPU and storage), network components frequently undergo changes in ways not as readily analogized to cell biology. The use, application,  or "role" of a computer in a network can change: in many cases, computers become specialized and provide application as well as communications services. The relationship of the user of a computer to other users and server computers in a network can change. The geographical location of a computer in a network can change. The characteristics of the connections between computers may change. Depending on the number of hosts and LAN segments, considerable planning may required so that multisegmented LANs deliver services to the users of the concatenated set of LANs. 

LANs, or more precisely, the users of LAN computers, never seem quite happy existing in isolation from other LANs. For the users on any given LAN in an enterprise, there's always more and different and indispensable information on another LAN somewhere, whether that somewhere is down the hall, on the next floor, across campus, at a regional or main office, or at an international office, on a different continent. Users need to access this information after all, the information is indispensable or mission critical and so wide-area connections must be established between LANs. 

Accommodating growth and change in enterprise and multiorganizational networks adds a new dimension to what may already be a considerable technological challenge for an organization. Many organizations are not prepared to deal with change at all these levels. They often do not have and cannot produce clear definitions of user (application) requirements and information flows; without these, their ability to define appropriate service upon which to base physical and logical network design criteria is greatly limited.  So standard operating procedure for network design is often more seat-of-the-pants than anything else.

Why?  The answer is amazingly simple.  Network design is a science with too few references and guidelines for practitioners.  Chemists, biologists, nurses, and physicians have a wealth of references upon which they can rely for formulae, actions and reactions, prescriptive indications and contraindications. But there are few places to go where you will find a statement of the nature "if you observe symptom X and you apply this, then you should observe Y or Z" when you attempt to design or redesign a network.  Network design is hardly as exact a science as medicine and my wife reminds me that medicine is not as exact as we would hope.

A common cocktail-party analogy I make regarding network design is that networking is a lot like plumbing. But there are guidelines and municipal codes that plumbers apply to size pipes for water mains, water supply distribution, and drainage. A plumber can better estimate the number of sinks, showers, washers, and tubs a main supply pipe of a given diameter can support at a given pressure than a network designer can estimate computers on a network because the plumber has more exact information to work with.

A plumber needs to understand the number and diversity of water delivery systems (e.g., faucets and showerheads), where the delivery systems are located, and the amount of water pressure desired before determining the appropriate diameter of a main supply and internal delivery pipes. Similarly, network designers need to understand the number of users, the network resources consumed by the applications they use, and the individual and collective flows of information (which, unlike plumbing, is bidirectional!). 

Just as municipal codes provide plumbers with guidelines for calculating pipe diameter to achieve a given rate of water flow, this book provides guidelines for calculating information flows. From them, the network designer and team will come to understand how to characterize service, identify service requirements, and develop an information flow specification. Author Jim McCabe also explains how to develop criteria for evaluating network technology, and how to determine appropriate methods and strategies for physical and logical LAN interconnection. 

Practical Computer  Network Design and Analysis fills a much-needed gap in data-network reference literature. As a consulting editor for a professional series on computer networking, I have read dozens of books on network theory, and dozens more that describe protocol bits and their operation. The shelves of every bookstore I visit are bursting with how-to's for personal computer operating systems and software applications. These are important literature for computer scientists, but to design a network that meets user expectations, one needs more.

This book is something special.  Jim describes a systems approach to designing networks that is disciplined and exacting.  I refused to relinquish my preliminary manuscript because it's quite frankly not merely the only practicum I've encountered for network design, but it's also highly useful, and it has quickly become indispensable.

[Company] [About Us] [Services] [Technology] [News] [Contact Us] [Search]

Contact our webmaster with questions or comments regarding this site.
Copyright 2008 Core Competence, Inc. All rights reserved.