on January 4, 2001
Messrs. Marcus and Stern have created a book that works for the seasoned sysadmin as well as the HA newbie. Frank depictions of case studies provide helpful insight into the many pitfalls of creating HA solutions. The "key points" sections at the end of each chapter provide quick and easy checkpoints should a reference refresher be required.
Another excellent trait of this book is its readability and realistic approach to the HA "art". The mixture of pure philosophy tempered with daily reality provided me at least, with more certainty in finding the best overall balance of reliability, availability, and manageability vs. things like cost of ownership, etc.
There is an excellent level of technical proficiency treated in the text as well. The hardcore bits and bytes folk won't be dissappointed by Mr. Stern's treatment various scenarios. The attention Mr. Marcus pays to higher level structural detail also makes you say occasionally, "gee i would have never thought of that.". (And even if you had, i doubt you would have been able to elaborate on it as humorously.) :)
It's one of the few books i refer back to occasionally, (not so surprisingly, this goes for Mr. Stern's Managing NIS/NFS book as well). Overall, it's just an extremely invaluable book that's actually a good read as well. (love the "Tales from the field" blurbs). I don't believe there's any stone regarding HA left unturned by this book. Replication, DR, backup and restores...it's all there.
I look forward to their next collaborative effort.
High availability is almost like the Holy Grail - believers know it exists, but getting it can be a quest. Not that it's an elusive, impossible-to-achieve objective - it's just expensive and the path to high availability is filled with challenges, blind alleys and the risk of spending money on the wrong things.
This book covers every conceivable aspect of high availability, from application recovery to the [Zen] of service level agreements. The emphasis is on the underlying technologies, with a good deal of attention paid to processes and business considerations as well.
There are two things that stand out about this book: it is comprehensive and the authors are in-the-trenches technical types who can actually write.
It is obvious that the authors have stuggled with and mastered every technology and technique about which they have written. This is a refreshing change from the plethora of technical books written by professional writers who are assigned book projects because they have writing ability, but not necessarily any real technical background. So, in a world that is littered with books with no substance, this gem is definitely worth buying.
I am not going to rehash the table of contents, which is readily available on this page - I will say that the authors gave each of the topics excellent treatment on a number of levels. First, they have deleved into the mechanics of each topic, they have shown the strengths and weaknesses of each, and in many cases, have provided anecdotes and real-life stories about their experiences with designing and implementing similar solutions. The anecdotal content follows a pattern: it points out problems and how the authors resolved those problems when they encountered them. These alone make the book invaluable, and increase the credibility of the authors.
Who should read this book? Systems engineers and managers.
Systems engineers will benefit from the wealth of technical information provided in the book, and will also benefit from the excellent explanations of underlying mechanics of how the various technologies and solution sets work. I am sure that most systems engineers will find the book's many anecdotes to be entertaining reading that imparts valuable information.
Managers should read this book because it puts into perspective how various pieces of a high availability solution set fits together. It also provides a dose of reality that might just stop some "suit" from decreeing a technical direction without fully understanding the full set of issues and factors. This book will provide the issues and factors to managers in a readable manner.
on March 20, 2000
Finally a book that specifically discusses the very important topic of availability. The book discusses how to address each layer of an infrastructure from the data layer, network, application, etc. with an emphasis on availability requirements. I found it to be an excellent synthesis of all of the components of end-to-end availability. Anyone involved in Web Architecture and design will certainly find value in this book. My only criticism is that there are a few plugs specifically for Veritas but as a fan of their products it didn't bother me.
on June 15, 2003
Anyone who has ever worked with me in the past knows I'm somewhat of an availability freak. It is the mark of a network administrator's excellence if they can say they maintain a 99.75% availability rating. High Availability as a design concept has kind of been brushed off to the side in recent years. This book was written is 2000 and at that time I think I bought about three books on the subject. Nothing new has been written about it. During the year 2000 there was a lot of discussion about availability in the private sector since a company like Amazon.com or any other on-line retailer lost big bucks if they were down for just minutes. Availability rating was the most important benchmark of a company's network. This book is my favorite on the subject because it is easy to understand. Cisco published a book on availability and you had to be a CCIE to understand the Introduction. This book is easy to read and marked with good illustrations to emphasize things like RAID, SANs, redundant routing, etc. Hardware companies have built servers, hard drives, controllers, etc. with high availability in mind so for the network administrator it is no longer critical for them to understand or be able to explain availability. However, high availability as a design concept is still the standard and once you understand how it all works and what the difference between 99% versus 99.75% availability you can take steps to maintain that high availability of your network systems.
on July 31, 2000
I must first confess to being a colleague of sorts of Evan's--I work in the same company (VERITAS Software), but in a different part. I'll also confess to having bought the book (yes, I bought it) more out of curiosity than out of an immediate need for the information it contains.
Having said that, I will go on to say that I am pleased with my purchase. This book is great, not because of the detail it contains in any one area, but because it tells you what all the areas are that you need to be concerned with if you're implementing a highly available system. Lots of books answer questions on how to implement this or that aspect of high availability; what's much less common is a book like this that you can use as a check list to see if you have all the bases covered. It doesn't go into great detail on any single topic, but it gives you enough meat to make informed decisions in the less technical areas (like staff, training, etc.) and to send you searching for more detail in the more technical areas (like backup, RAID, and clustering).
I'll pay Evan and Hal another compliment. I own many more technical books than I have read. I didn't really expect to read this book, just because I don't read most of the books I buy all the way through. What I found, however, is that this book sort of creeps up on you. You read a little, and a little more. You put it down, only to pick it up later. The information is very well organized and chapters are self-contained. You can pretty much open it anywhere and have an informative read for 15 minutes, for an hour, or whatever time you have to give it.
It's obvious that both authors bring lots of hands on experience to this book. The "Tales from the Field" anecdotal sidebars are worth the price of the book by themselves. I wish I could bring these guys to bear when I have HA problems. If I can't, I'll at least have their book at my side.
on December 30, 2003
Hal and Evan have produced an excellent overview of the field of reliable computer systems which is useful to system administrators, system architects and to non-technical people who need a general understanding of the field.
As a system architect with a programming background facing the task of designing a simple highly reliable system this provided an excellent perspective on the different issues and technologies. It left me well prepared to then delve into the product literature of specific products that addressed the issues relevant to the project.
I believe this would also be an excellent book for IT managers who are looking at commissioning a "high availability" system, whether from an external software integrator or an internal company IT department. In particular, it describes what is easily achievable, what is achievable only at great expense, and what is simply not doable. It also emphasises the roles and responsibilities of people involved in ongoing support of HA systems; this book clearly describes how creating such a system is not the end but the start of providing reliable services.
The reading will be challenging for those of non-technical background, but the book keeps its focus on why you would use each technology rather than simply how each is applied, which makes it relevant to a wide range of readers. It should also ensure that this book remains useful even as technologies change.
What this book does not do is provide a formula for setting up a system. It's a great first step, though.
In addition the writing is clear, structure and flow are good and mistakes are few. Not the lightest of reads, but then again this is a serious topic about systems worth serious money. And at 550-odd pages of dense reading material, this book is well worth investing in.
on March 17, 2003
An excellent overview of high availability techniques. Starts with "why and how much HA do you need", and goes all the way through the hardware and design side of HA.
Doesn't describe any product in details, the authors explicitly refrain from doing so. Instead, the book makes you think the right way by pointing at the actual problems and offering actual decisions. Upon reading this book you can easily answer the question "what can we do to make it work", not "what brand of server should we buy".
Covers HA theory, redundant hardware, redundant systems design, failover techniques, replication, backups, procedures, disaster recovery.
The only thing that I didn't like and still can remember (a year later that I've read it), is that in my opinion the authors should stay totally clear off the "how to write a stable software" side of HA. There is like 2 pages of that, and it doesn't sound like anything sane.
Clear language. Solid visual design. Lots of (fun and) real-life samples. See the title of the review.
on May 24, 2000
This is a superb book. Anyone who doesn't concern himself with availability ;-) shouldn't buy it, but the rest of us need the kind of clear-headed, on target, practical guidance that this book offers. Evan and Hal have done a wonderful job of presenting the issues and options for HA. This book isn't a "must have", it's a "must read"!
on February 24, 2001
I'm a German reader, who has to deal with HA, writing and publishing about HA and to be aware of loosing all SCSI devices at once because of an electic accident.
Everyone who hangs on HA will be glad that there is a new book available. Most of them are from the last decade. The book is clearly structured. You can put it away and open it later by reading the conclusions to get the lost thread again.
And it`s humorous: I'm shure the the small hearts may have been helpfull for John, George, Paul and Ringo to play together in 1-to-4 environment. Something that's unusual (in Europe) when you read books about dry hard stuff like HA.
Everyone, who has to manage resilience system should have this book and the "20 keys system design principles" should be plugged on the door - the operational part of the book.
But there is a drop of bitterness: the chart of "Causes of Downtime" (from 1995) and "Costs of Downtime" (1996) should be newer -- as time goes by. At least, I miss some word in the Index. There are new solutions working, which don't have to be based on Sun systems -- like Linux?
But my message to Evan and Hal: go on, we'll need those experiences as printed in the book "tales from the field".
on November 5, 2003
It isn't often that a single book can deliver solutions to so many issues that plague enterprise executives and system administrators alike, but this book delivers the goods.
Most technology related documents focus on the low-level gorp that makes up complex solutions, but this book provides information at all levels. The focus of the information is made clear through the use of real-world examples, and research that must have taken years to assemble. As well, there are mathematical formulas that help one determine how outages can effect a global enterprise or even a small business.
The authors even had the forethought to end the book with a look into future technologies that may solve some of the more complex issues that elude even the most expert of solution architects.