The theory goes something like this: an open source project is somewhat like an unclaimed turf in a gang war. The gang that can claim the turf will essentially control all flow of cash and illegal goods in and out of the area. Similarly, the company that controls an open source project stands to gain the most when the dust settles. As the turf war for Apache Hadoop distribution heats up, several experts have weighed in with their predictions of winners.
The debate began with Matt Asay, SVP of business development at Nodeable, who argued on The Register that open source markets tend to be “winner-take all”. He likened the current battle for Hadoop to that of Linux, which he says is now firmly in the grasp of Red Hat. There is no room for two major Hadoop vendors, he asserts, and at this time, the two major contenders are Hortonworks and Cloudera.
John Furrier, Founder and Editor of SiliconANGLE responded to Asay’s assertion, arguing that the Hadoop market is considerably different from the Linux market of several years ago. It is moving much faster and grew out of a more organic environment in big data, rather than from the dissatisfaction with proprietary Unix vendors that spawned the Linux revolution.
While Asay argues that there can only be one winner and that Cloudera has a significant lead in the race, Furrier contends that Hortonworks has made noticeable strides and employs some of the top Hadoop contributors. Moreover, Hortonworks has adopted a 100% open source model to its distribution that has helped it grow strong. Rather than the necessity of a single gang to rule the turf, Furrier believes there is room for two at the top.
Hadoop is still very young, especially for an open source project, so it is probably too early to actually start making bets on any clear winner. Furthermore, as Furrier noted, there is no real reason to assume the Hadoop wars will only have one winner who will contribute the most code and control the direction of development, as one could argue Red Hat does for Linux. Regardless of the outcome, however, it will definitely propel big data development in general and Apache Hadoop development in particular.