We've generally used the Cloudera distribution for Hadoop in this book, but have attempted to keep the coverage distribution independent as much as possible. We've also mentioned the Hortonworks Data Platform (HDP) throughout this book but these are certainly not the only distribution choices available to you.

Before taking a look around, let's consider whether you need a distribution at all. It is completely possible to go to the Apache website, download the source tarballs of the projects in which you are interested, then work to build them all together. However, given version dependencies, this is likely to consume more time than you would expect. Potentially, vastly more so. In addition, the end product will likely lack some polish in terms of tools or scripts for operational deployment and management. For most users, these areas are why employing an existing Hadoop distribution is the natural choice.

A note on free and commercial extensions—being an open source...