Preface

Local area network (LAN) performance at many organizations is a topic as frequently discussed as the weather. However, unlike the weather, for which our options of control are extremely limited, we have a significant ability to control the performance of LANs.

As I began to revise this book, I returned to my original outline several times to think about the issues I face as a network manager. Some of those issues include responsibility for establishing new LANs as well as expanding existing networks and providing the communications capability for users on different networks to communicate with one another or access a number of mainframe applications residing on different computers. As a network manager with a fixed budget, there are numerous economic- related problems I need to consider. For example, when interconnecting two LANs via a wide area network (WAN), I need to determine an optimum transmission line rate. If the line rate selected does not provide sufficient bandwidth, then user productivity will suffer. If the bandwidth exceeds what users require, my organization will be paying for wasted resources. Thus, selecting an appropriate line rate to interconnect geographically separated LANs represents both a technical and financial issue. Another common issue facing network managers and LAN administrators occurs when considering whether or not to use multiport devices or dual devices for reliability. In addition, network managers and LAN administrators commonly are tasked with estimating network traffic for networks they plan to install but for which no prior traffic estimate exists. Other frequently performed tasks include estimating the effect of network configuration changes on the performance of Ethernet and Token Ring networks and understanding how to design switch-based networks to prevent data traffic bottlenecks. These and other issues can be considered to represent a core set of operational issues you will want to answer to ensure that your networks not only meet the requirements of your user community but, in addition, are not overconfigured. Concerning the latter, as sure as the sun rises in the East, you will never hear an end user complain that network performance is too good. Thus, you will more than likely never be told that you overconfigured your network or transmission facilities used to interconnect networks, although doing so can result in a considerable expenditure of funds beyond those necessary. Once you realize this, you must also recognize that in an era of budgetary constraints, it is just as important not to oversize your network components and transmission facilities as it is not to undersize those components and transmission facilities.

In presenting various LAN performance issues, I have also attempted to provide you with the tools to answer those issues based upon your specific networking requirements. To accomplish this goal, mathematical models of different LAN performance issues are developed throughout this book models that you can use to determine answers to your specific networking requirements. And because the execution of some models is tedious and time consuming, I also developed several QBASIC programs and Excel spreadsheet models that you can use to facilitate your computations .

Because a book must be finite in length, there are probably some issues I have overlooked or, due to space constraints, have not covered. If there is a particular issue you would like me to work on, please feel free to contact me and I will do my best to answer your questions. Perhaps your questions, as well as the answers to those questions, will result in their inclusion in a fifth edition of this book, which will allow future readers to gain from your experiences. As a professional author, I value your comments and encourage you to write me through my publisher or you can contact me directly via e-mail at [e-mail].