Most Helpful Customer Reviews
Cassandra Design Patterns based on 0 ratings. 2 reviews.
Cassandra Design Patterns is very brief and to the point book.This book is not for beginners and is more suitable for architect and developer roles. It starts with a brief history of Cassandra to explain what are the key advantages of Cassandra and how are they achieved under the hood. The Book then makes a short reference to the GoF patterns and introduces the 3V (Volume, Velocity, Variety) Model and explains it in the context of the Big Data. The author then describes in short the common patterns solvable with the Cassandra and gives tips how to implement them. Every presented pattern is described with the problem-intent, context-applicability, forces-motivations, solution summary and an example. I found the description of Complex Event Processing pattern to be very useful. The good part about this book is that it also has a chapter called Patterns and Anti-patterns where the author describes when not to use Cassandra. As developer I used some of the techniques presented in this book especially the reversed indexes described in the chapter about the Needle in a Haystack pattern. All in all i find this book a nice and decent read for any Cassandra developer or architect.
To me the most useful part of this somewhat skimpy book is the table of features of Cassandra. Usefully placed at the start, in chapter 1. It shows at a glance the items, with their antecedants in Google BigTable and Amazon Dynamo. The chapter provides a succinct recap of the basic ideas behind those 2 commercial approaches. We get the big picture of the CAP theorem. This provides the theoretical context for understanding BigTable, Dynamo and Cassandra itself. Granted, some readers will be frustrated with the modica of details furnished in the chapter. You may have to use this to extract search terms and look online for more information. The remainder of the text delves into the Cassandra patterns for efficient writes and reads. Well, it is a change from most books on general purpose design patterns. Those cited here are indeed specialised. This is the book's value. It's not about generic patterns. One particular pattern is intriguing. How to integrate with higher level analytics. Cassandra is motivated primarily for real time transactions. Analytics is typically and necessarily batch oriented. Too computationally intensive to do in anything approaching real time. Well, turns out one pattern lets you link to Hadoop, which is strong at those batch analytics. But the discussion of this is frustratingly sparse. More could have been written here.