Book Image

Learning Apache Cassandra - Second Edition

Book Image

Learning Apache Cassandra - Second Edition

Overview of this book

Cassandra is a distributed database that stands out thanks to its robust feature set and intuitive interface, while providing high availability and scalability of a distributed data store. This book will introduce you to the rich feature set offered by Cassandra, and empower you to create and manage a highly scalable, performant and fault-tolerant database layer. The book starts by explaining the new features implemented in Cassandra 3.x and get you set up with Cassandra. Then you’ll walk through data modeling in Cassandra and the rich feature set available to design a flexible schema. Next you’ll learn to create tables with composite partition keys, collections and user-defined types and get to know different methods to avoid denormalization of data. You will then proceed to create user-defined functions and aggregates in Cassandra. Then, you will set up a multi node cluster and see how the dynamics of Cassandra change with it. Finally, you will implement some application-level optimizations using a Java client. By the end of this book, you'll be fully equipped to build powerful, scalable Cassandra database layers for your applications.
Table of Contents (14 chapters)

Lightweight transactions and their cost


Lightweight transactions allow us to maintain data integrity in the face of concurrent updates, but they don't do it for free. Because of Cassandra's distributed architecture, it actually goes to great lengths to guarantee that the data is in a certain state before modifying it, as all the machines that store that piece of data need to be in agreement.

Note

Accordingly, there's a performance penalty in using lightweight transactions. As a result, you shouldn't use them in situations where you don't need to.

 

When lightweight transactions aren't necessary

As we discussed earlier, concurrent insertions aren't a concern for any table that uses UUIDs that are generated by the application or by Cassandra at row-creation time. Simply using UUIDs guarantees that we'll never have a row key collision.

Another scenario in which we can skip conditional inserts is when we have a globally unique natural key. For instance, if you're building a Rich Site Summary (RSS...