Book Image

Learning Hadoop 2

Book Image

Learning Hadoop 2

Overview of this book

Table of Contents (18 chapters)
Learning Hadoop 2
Credits
About the Authors
About the Reviewers
www.PacktPub.com
Preface
Index

Distributions of Apache Hadoop


In the very early days of Hadoop, the burden of installing (often building from source) and managing each component and its dependencies fell on the user. As the system became more popular and the ecosystem of third-party tools and libraries started to grow, the complexity of installing and managing a Hadoop deployment increased dramatically to the point where providing a coherent offer of software packages, documentation, and training built around the core Apache Hadoop has become a business model. Enter the world of distributions for Apache Hadoop.

Hadoop distributions are conceptually similar to how Linux distributions provide a set of integrated software around a common core. They take the burden of bundling and packaging software themselves and provide the user with an easy way to install, manage, and deploy Apache Hadoop and a selected number of third-party libraries. In particular, the distribution releases deliver a series of product versions that are certified to be mutually compatible. Historically, putting together a Hadoop-based platform was often greatly complicated by the various version interdependencies.

Cloudera (http://www.cloudera.com), Hortonworks (http://www.hortonworks.com), and MapR (http://www.mapr.com) are amongst the first to have reached the market, each characterized by different approaches and selling points. Hortonworks positions itself as the open source player; Cloudera is also committed to open source but adds proprietary bits for configuring and managing Hadoop; MapR provides a hybrid open source/proprietary Hadoop distribution characterized by a proprietary NFS layer instead of HDFS and a focus on providing services.

Another strong player in the distributions ecosystem is Amazon, which offers a version of Hadoop called Elastic MapReduce (EMR) on top of the Amazon Web Services (AWS) infrastructure.

With the advent of Hadoop 2, the number of available distributions for Hadoop has increased dramatically, far in excess of the four we mentioned. A possibly incomplete list of software offerings that includes Apache Hadoop can be found at http://wiki.apache.org/hadoop/Distributions%20and%20Commercial%20Support.