High Availability Network Fundamentals (With CD-ROM)

Rating:
Summary: Important addition to any network library
Review: Like any book on availability this one addresses a number of interrelated topics - service level management, capacity and performance management, and specific techniques for achieving high availability.

I like this book because it introduces proven engineering techniques in the first five chapters, beginning with a discussion of availability and building upon this foundation with necessary math, a discussion of network topologies, factors that affect availability, and end-to-end availability management. This part of the book will not only provide valuable information to network professionals at all levels of experience and knowledge, but is also applicable to IT operations managers and infrastructure support professionals who are involved in service level management.

The remaining chapters are realistic case studies that are based on Cisco products that give excellent insights into attaining and managing availability in various environments - a small ISP, a large enterprise and an enterprise VoIP network.

In addition to a wealth of information, this book comes with a CD ROM that contains a generic hardware availability and reliability calculator (in Microsoft Excel format - I was able to use it with an older version of Excel 97, which is a nice touch), and a terse PDF of instructions. It also contains worked examples of this spreadsheet for various Cisco products, the most valuable of which in my opinion are the Cisco 12000 processing engine and system level calculation spreadsheets that accompany Chapter 9 "A Large VoIP Network: An Availability Analysis".

This book is essential reading, and is nicely augmented by "High Availability Networking with Cisco" by Vincent C. Jones (ISBN 0201704552) because that book takes the fundamentals and expands them into complex scenarios. If you're new to networking get this book before getting Jones' book, but if you're experienced, get both.

Rating:
Summary: Very elementary introduction to Cisco HA
Review: Reliability engineering has been around for quite some time now, and has taken on even more visibility with the rise of the Internet and the subsequent role of eCommerce. The financial stake in reliable networks has driven the need for mathematical analysis that will shed light on how resilient a network is to downtime. This book gives a very elementary introduction to this analysis, and is written for those who do not want to imbed themselves in mathematical constructions beyond the level of elementary algebra. Readers who need a more advanced overview, and one that is more general in not being specific to Cisco products, can consult the references given in the book.

The first chapter of the book introduces the concept of HA, why it is important, and introduces the arithmetic needed to quantify HA, such as the MTBF and the MTTR, the percentage method, and the defects per million method. The author clarifies the difference between MTBF and MTTF, and explains his decision to use MTBF instead of MTTF throughout the book. Due to the level of mathematics to be used in the book, the author does not discuss partial outages in any detail.

Chapter 2 introduces the mathematics to be used in the book to quantify high availability. Eschewing completely the use of calculus in the book, the author sticks to elementary arithmetic and algebra throughout. Many examples are given throughout the chapter on how to compute availability for both the serial and parallel cases. No details of fail-over mechanisms in parallel systems are given since the author wants to stay away from probability theory. N + M redundacy is not discussed in detail as it too is deemed too mathematically complicated.

In the next (very short) chapter serial, parallel, and serial/parallel network topologies are discussed, and an elementary example of how to use "divide-and-conquer" to calculate availability is given.

Chapter 4 overviews the different factors that will affect availability, such as hardware, software, the environment, human factors, and design considerations. The discussion on the prediction of software availability brings in the concept of a "lognormal" distribution from probablity theory, but the author explains it fairly well. Although the discussion on software availability is very elementary, it serves as a good introduction for further reading on the subject. The author is careful to note that the mathematics he is using and the assumptions he is making are geared toward simplifying the mathematics.

In chapter 5, the author gives a more detailed overview of the divide-and-conquer algorithm for calculating availability, restricting his attention to hardware and software failures. He illustrates, with many examples, the construction of reliability block diagrams.

Chapter 6 discusses three "real-world" examples of network availabilty prediction dealing with the uBR924, uBR7246, and 12000 devices. The chapter gives the reader more practice on how to calculate availability using the various components of these devices.

The availability analysis becomes even more useful in chapter 7, where it is done for a small Internet service provider. Use is made of the accompanying CD to the book, which employs Excel spreadsheets to perform the computations and report the results. These results are generalized to a small Enterprise network in chapter 8, and again the author makes use of the CD SHARC spreadsheet to calculate the availability.

The last chapter of the book is an availability analysis of a large voice-over-IP network. All five major contributors to network downtime are considered, and the reader gets a good taste of the complexity involved in analyzing availability in large networks.

Rating:
Summary: Unbelievable what this book has in it.
Review: Sometime when I pick a book to review I have no idea what information is contained within the pages. So much to my surprise I was extremely pleased to find this 240 page book so chock full of useful information.

What I found after reading this book is that the author must be very experienced in the High Availability networking and the information was detailed, but also written at the level for the beginner or intermediate technician.

Taking the reader from the topologies and hardware, to environment and design, the author gives you 10 scenarios from real world situations. Finally there was a cd included with a System High Availability Calculator (SHARC) to calculate reliability and availability.

Overall I have most certainly learned something new from a well organized and well documented book - great job!

Rating:
Summary: A must if you care about your data infrastructure
Review: This book provides an excellent model to evaluate the availability of your data network. The tools provided in this book are straight forward and easily understood. If your network is critical to business, and especially if your voice traffic is running over it, you need this book. This book was written as something to use, not just to read.

Rating:
Summary: A must if you care about your data infrastructure
Review: This book provides an excellent model to evaluate the availability of your data network. The tools provided in this book are straight forward and easily understood. If your network is critical to business, and especially if your voice traffic is running over it, you need this book. This book was written as something to use, not just to read.