You know the kid on the block who claims victory even though the game is still being played — and when you ask him why he thinks he’s won, his answer is “because I said so?"
It seems that Tom Reilly, CEO of Hadoop distro provider Cloudera, might have borrowed a play from that book, inspiring Gigaom’s Derrick Harris to write the headline, “Cloudera CEO declares victory over big data competition.”
And while some might conclude that Reilly was talking only about Pivotal — which open sourced its big data platform last month and is now partnering with Hortonworks on its HDP distro — there's more to the story. There are at least two other Hadoop distro providers in the space whose market share is still growing, perhaps as fast or even faster than Cloudera’s. Analysts like them just as much, if not more, too.
Lots of Players
Consider that Hortonworks, in its latest quarterly earnings report, claimed it added 99 new support subscription customers. Given that the company was founded in June 2011 and has 332 total support subscription customers to date, the last three months of 2014 is impressive by a customer acquisition metric.
MapR, which has yet to go public, doesn’t disclose its numbers. But according to Gartner’s Magic Quadrant for Data Warehouse and Data Management Solutions for Analytics, which was just released (look for our summary next week), it and Cloudera sit neck to neck. MapR rates slightly higher in terms of completeness of vision and Cloudera rates slightly higher in execution.
Pivotal betters both of them by a substantial margin in completeness of vision. And Actian, whose SQL on Hadoop Vortex, competes with Cloudera’s Impala, is closer than any other vendor mentioned thus far, to landing in the Leaders quadrant.
Forrester, in its Wave around Hadoop that was released last year, places both MapR (in terms of strength) and Hortonworks (in terms of both strength and strategy) in a better position than Cloudera — even though the margins aren’t that big.
So is Cloudera doing well? Yes. Has it blown away everyone else on the map? There’s evidence to suggest that that’s not the case.
'Light Years Ahead'
In fact Jack Norris, MapR’s CMO pointed out that according to some highly respected analysts, “We have the top ranked Hadoop, the top ranked NoSQL and the top ranked SQL on Hadoop.” He added that MapR’s vision is broader than Cloudera’s. “We’re helping companies leverage data and to transform it as things happen,” he said. “We’re not looking to be an Enterprise data hub,” which is Cloudera’s stated aim.
Norris claimed, “We’re light years ahead of Cloudera and Hortonworks,” in executing on a greater big data strategy. “There’s a huge gap between what we deliver and they have to offer,” he explained.
And while we’re not looking to start a slug fest between vendors, Hortonworks President Herb Cunitz offered that “It’s hard to see how Cloudera can claim victory when they’re adding fewer customers than we are”.
Though Cloudera’s total customer count is greater than Hortonworks’ at the moment, it’s an older company and therefore has had more years to establish itself and win business. Not only that, but Cunitz also points out that Hortonworks is now selling subscriptions faster, and that there’s proof in the numbers.
He bases his conclusion on Hortonworks’ performance records, which are publicly available (it’s traded on the NASDAQ) and Cloudera’s self-reported, unaudited business results which it released last month for its own reasons (since it is yet to public).`
It’s worth noting too, as Cunitz points out, that when it comes to Hadoop 2 (YARN), which Enterprises are adopting now, Hortonworks is an obvious position of strength— its engineers wrote most of the code.
The Big 'But'
This is not to undermine Cloudera’s contributions, but simply a fact, Cunitz explained. And it’s one that is vital to Hortonworks business strategy given that it sells subscription support for HDP its 100 percent open source Enterprise grade Hadoop distro.
Why does this matter? According to Cunitz, it’s because if you have a problem with Hadoop, you want to get help from the provider who built the product (and that’s Hortonworks). There are also other factors such as governance, security, and scalability to consider, especially as IoT data comes streaming in. Distro providers will build, buy and adopt projects and products. and how they do that and what they deliver to customers will determine who does or doesn’t win business.
Not only that, but with subscription and open source models, vendors have to be able to keep the business they win. And that’s not a given. in fact Hortonworks can point us to Cloudera customers they’ve won over. It’s hard to think that Cloudera and MapR might be able to do the same.
Either way, it’s not game over.