Traditionally, Hadoop has been supported on Unix-based operating systems. Installation on Microsoft Windows was tedious and not consistent. It involved installing Unix-based emulators such as Cygwin and carrying out installation steps similar to Hadoop installations on Unix systems. Other alternatives were to run a Linux virtual machine on Windows hosts and install Hadoop on them. But Hadoop was still not natively available on the Microsoft Windows operating system until Hadoop 2.0 arrived.
With all major players moving into the cloud, the Hadoop as a Service (HaaS) offering is becoming popular. It offers an easy and cost-efficient way of analyzing big data on the cloud. Microsoft also joined the cloud bandwagon with the Azure suite of services on the cloud. The Microsoft Azure cloud not only supports Linux Virtual Machines, but also provides Hadoop as a service. Players such as Hortonworks collaborated with Microsoft to bring Hadoop to Windows.
Hadoop's native support on Microsoft Windows becomes immensely important because:
Microsoft Windows is known for having excellent tooling for business intelligence. Microsoft Excel, PowerPivot for Excel, and PowerView are a few examples of tools that can facilitate powerful decision making by analyzing and visualizing data from enterprise data sources. The power of these native Windows tools can be unleashed on big data stored and processed by Hadoop.
SQL Server and related technologies are native Windows database solutions that are widely deployed in many enterprises. They cater to the enterprise needs of storing and managing structured data. With Hadoop on Windows natively, unstructured data can also be added into the mix for insightful decision making. This involves almost zero migration and learning costs for the enterprise.
In this chapter, we will look at single-node deployment of Hadoop on Microsoft Windows.