Open Source is sexy in a cloudy, big data world, but if you want your commercial-grade Hadoop distribution to be speedy, reliable and secure, you’ll be better off parting with a few dollars and buying a proprietary solution like ours.
That’s MapR’s pitch to Enterprises venturing into the land of Big Data.
And though some Hadoop vendors find MapR’s posturing to be divisive, separatist and bad for business (and the Hadoop ecosystem) over the long haul, companies as tech savvy as Google and Amazon think that it’s such a good bet that they've hooked up with MapR themselves.
The big difference between us (and our Hadoop distribution) and the other Hadoop vendors (and their Hadoop distributions) is that we offer high availability, data protection, and dependability,” says Jack Norris, MapR’s VP of Marketing.
Apache Hadoop committers (those who have been selected to have “write” access to the Open Source Apache Hadoop project) no doubt disagree. They also don’t have much respect for developers and/or vendors who write patches for Hadoop but don’t share them. It’s the antithesis of how an Open Source community is supposed to work, they say. But MapR doesn't seem to be bothered by their sentiment.
One-Upmanship in the Hadoop World
Instead at least one manager at the company isn't shy about pointing out that the Hadoop community has not been (or wasn't until recently) able to provide snapshots and mirroring or to eliminate failure in Hadoop’s name node. (Thus implying that MapR can serve up Enterprise grade Hadoop better than the Apache Hadoop Open Source project).
It’s important to note that MapR’s business model is different than that of its primary competitors -- MapR is primarily a software company that offers some services; Cloudera provides its basic version of Apache Hadoop for the enterprise, Cloudera Enterprise Free, free of charge, but sells products and services that make using Hadoop more palatable; and Hortonworks, being the largest contributor to the Apache Hadoop project, provides its enterprise-grade Hortonworks Data Platform, a 100% open source distribution of Apache Hadoop, free of charge -- the company makes its money selling education and services.
MapR Opens its Code
That being said, late last week, MapR, which isn't a big contributor to the Open Source Apache Hadoop Community, did something unexpected -- they announced that the source code for the component packages of the MapR Distribution for Apache Hadoop are now publicly available on GitHub.
When asked why MapR decided to make such a move, Norris answered that it was something customers requested.
OK, we’ll buy that.
But add to that MapR’s next piece of news, that it is partnering with Canonical (the organization behind Ubuntu‘s Open Source Operating System for the desktop, server and the cloud) to build a solution for quickly implementing a Hadoop cluster in any enterprise, we can't help but wonder if this new relationship didn't influence MapR’s decision. After all, Canonical champions Open Source in a very big way.
Consider its promise to developers and other users:
Whether we're right or wrong, this is a big win for Canonical and MapR customers and to Enterprises who plan to leverage Hadoop in a safe, painless way.