In recent weeks, the folks at Cloudera, which provides support for implementations of Hadoop in the enterprise, have lined up an impressive number of allies for the increasingly popular ‘Big Data’ platform.
Among the vendors promising to support Hadoop via alliances with Cloudera are Teradata, Netezza, Quest Software and, most recently, the Greenplum unit of EMC.
Quest Software is more interested in helping to manage Hadoop in the enterprise, while Teredata, Netezza and Greeplum all have database engines that can be invoked to improve the performance of Hadoop applications.
The question all these alliances bring up is when to best use Hadoop. According to Ben Werther, director of product strategy for the Greenplum/Data Computing Products division of EMC, Hadoop is really a programming model for developers trying to build applications involving huge data sets. In contrast, applications involving lots of database queries or statistics will still be better off running natively on a massively parallel database such as Greenplum.
The real question, of course, is what defines the term ‘Big Data.’ It seems like there are plenty of scenarios where regardless of the size of the data, Hadoop could be applied. And given its relatively low cost of adoption as open source software and its rising popularity in the Web 2.0 community, you can’t help but wonder if Hadoop will find its way into more mainstream enterprise applications because of the flexibility of the environment.
Of course, there might be better tools suited for any number of tasks. But there’s no accounting for popularity and momentum, especially when it comes to Web 2.0 technologies. So don’t be surprised if we start to see Hadoop more broadly applied as the number of Hadoop evangelists increases. Because even if Hadoop isn’t always the perfect tool for every application, it’s becoming a tool that developers want to learn how to use. And once that happens, they’ll start to find any number of reasons to use it.
So in all probability, Hadoop is coming to your enterprise one way or another.
