Here is a quote from the preface: "Threading typically relies on a scheduler built into the language". If the author was writing about Windows 3.0 I would probably agree. But threading has been handled in the OS and hardware for ages. The book relies entirely on Python, and perhaps the author's statement is true for Python; but the author's technical inaccuracies lead me to be skeptical about his other claims.
This main content of the book is split up into three parts.
The first part discusses the boto library, which is a Pyhton layer on top of the Amazon API that the author wrote. This is the first strike against the book - I tend to dislike this type of bait-and-switch style. Second, the book focuses almost entirely on Amazon's cloud products - the author does mention other clouds briefly but 99% of the examples are Amazon specific, using boto of course.
The second part describes the various pattern, all of which are very obvious to anyone who has done any enterprise-level application deployment.
The third part goes into detail on a blog application, one written in boto, the other written using Marajo for Google's AppEngine. The problem with this section, and with the code examples for the prior section is that the application of the pattern(s) is never clear. The author spends a lot of time explaining the application and how it works, but very little, if any, time explaining how the application illustrates the pattern(s).
I came away feeling as if the author meant to write a book about boto, realized it wouldn't sell, so converted it into a book on cloud application patterns to give the book broader appeal. I don't think he succeeded.
Disclosure: I received a copy of this book from the publisher in exchange for writing this review on behalf of CodeRanch.