Big Data: Principles and best practices of scalable realtime data systems

I don't have much background in large scale system and big data. I bought this book because some one online recommended it. I went non-stop for a few chapters and will definitely continue reading until I am done. Given limited background, I don't have problems grasping the concepts in the book at all. It's a very good book with theory clearly conveyed and examples beautifully demonstrated. I really hope all books in the distributed system area can do such a good job to make learning experience a lot smoother.

Great explanation of both the theory and practice of the lambda architecture. While the practice chapters are nice, it's the theory chapters that really shine. The book explains down to the byte level why components are implemented the way they are. For example, there's an immense amount of detail as to why using a db that doesn't support random writes allows for an application to query the batch layer's results without locking.The only downside to the book is that the architecture and exosystem is so new that there's not really a lot of pragmatic solutions. For example, the theory describes a query layer that can merge the results of batch and real time processing for client applications. However, in real life there are no pragmatic solutions for doing this so you'd have to write your own.It'll be interesting to see how the lambda architecture matures and to see future editions of this book. Hopefully, future editions will be as well written and have a better ecosystem for practice chapters.

In all honesty, the book has simplified big data architecture and its general premise in an eye opening way. Starting from the batch layer and spending a good amount of time addressing different aspects of it gave me a valuable lesson as a developer in understanding the complexity as well as the necessity of evaluating my data entries and their impact in the future formation of worthy analytics/results.My girlfriend and I enjoyed every chapter in this book. I guarantee you that you won't regret buying this book. I am looking forward to another book from you guys on the topic because its the first time where I couldn't wait to pick up the book and get to the end of it.

If you are looking for a survey of different approaches of handling big data, you want to read "ELEMENTS OF SCALE: COMPOSING AND SCALING DATA PLATFORMS". ([...]) This book is dedicated to Lambda Architecture (one that is surveyed in the above article.)The book is very organized. Introduction in chapter 1 will be the road map of the whole book. Motivating with a simple web application based on RDBMS, the author showed how the approach to scale it becomes undesirable. After enumerating a list of desired properties, he proposed Lambda architecture, an approach in contrast to fully incremental architecture (with RDBMS).The Lambda architecture is partitioned into three layers:1. batch layer that computes different views on big data2. serving layer that answers user queries using views from the batch layer and speed layer.3. speed layer that compensates an approximate answer over a period time when the batch layer is working on the complete answers.In the remaining chapters, the author dive deep into the rationale and requirements of all the different pieces of Lambda Architecture.To understand the context of Lambda Architecture, also refer to the wikipedia for crticism.

I feel really sorry for those who gave 5 stars for this book. I purchased the book and started reading it eagerly as soon as I received it. It got my attention until I got to page 20 with a statement saying "...... If anything ever goes wrong, you can discard the state for the entire speed layer, and everything will be back to normal within a few hours." Within a few hours? No high-traffic production sites can afford a few hours down-time. At that point, I decided to return the book, which I did.I did scan through the rest of the book, though. First, the so-called lambda architecture might sound like a new term, but many high concurrency websites already work that way. For a high concurrency web site, the first-layer would be memcached-based, which gives O(1) low latency on all queries. The second layer would be a clustered app-server layer. The third layer could be a high-concurrency, extremely low-latency layer like a NoSQL cluster. The far backend could be Hadoop- or Spark-based for batch jobs. This is the known architecture in production for high traffic websites that need to support millions of concurrent users.Secondly, the bulk of the book is actually about Hadoop in the so-called batch layer. Hadoop once generated some excitement, but has lost its steam due to the new kid in the spot named Spark, which can do whatever Hadoop can do, but 10x - 20x faster with a fractional cost.

This book serves as a guide to building the lambda architecture from scratch, processing and serving big data. It is not a book to teach you about big data technologies though. If you are looking to learn and implement a system with batch and real time layers, this is the best book there is.

This book has a bad binding. I bought this book and opened it only twice and it is already broken

I just received this book. Content is great but as I started to turn the pages, they started to fall off. I buy a lot of books and it has been a very long time since I saw such a bad quality in the book physical structure. 5 stars on the content. 0 stars on the book physical structure.