Studio DMM Photography, Designs & Art/Shutterstock
To the steadily growing list of big companies making strategic investments in the open source big data platform known as Hadoop, you can now add computing giant Hewlett-Packard.
Re/code has learned that HP today will announce a strategic partnership under which it will invest $50 million in Hortonworks, the Hadoop startup that spun out of Yahoo in 2011. HP will join the previously announced investment round in Hortonworks led by the investment firms BlackRock and Passport Capital. Martin Fink, HP’s executive vice president and CTO, will join the Hortonworks board of directors as part of the deal.
That funding round will grow to $150 million and will bring Hortonworks’ total capital raised to $248 million. When it was announced in March, the round was said to value Hortonworks at north of $1 billion and viewed as a possible step toward an IPO in 2015. At the time, Hortonworks CEO Rob Bearden told me that the round “could have been bigger, but frankly we don’t need it.”
The deal may not be as much about HP’s money as it is about its deep relationships with large companies struggling to get their big data problems under control. Hortonworks and HP will jointly develop Hadoop to run on HP’s Haven big data platform. And HP, as one of the world’s largest suppliers of hardware, software and services to large corporations, will throw its weight squarely behind Hortonworks, and among other things, provide service and support.
Colin Mahony, general manager of HP’s Vertica business analytics software unit, told me that about half of HP’s customers have been running Hadoop alongside Vertica.
“When we first started selling Vertica in 2007 we had been hearing about Hadoop when it was just getting started,” he said. “We track all the business intelligence tools and within a year we started seeing it. Now about half our customers are deploying it side by side with Vertica. The size of deployments will differ, but it’s real and growing rapidly.”
Hadoop is open source technology designed to make it easier to work with large collections of data in a distributed computing environment. Once a collection of data has been poured into a Hadoop cluster, companies can run analytics tools on it in order to glean useful insights. It’s a big part of the larger narrative around “big data” we hear so often these days. The idea is basically that you can learn a lot by asking intelligent questions of whatever data you happen to collect. The trick is figuring out what questions to ask. EBay and Twitter are two big users of Hadoop technology.
Hadoop itself is an open source project run by the Apache Foundation. The Yahoo team that created it went on to start Hortonworks, with the aim of building a commercial business by supporting companies deploying Hadoop.
This deal marks another strategic collaboration for Hortonworks. One with Red Hat was announced in February. Hortonworks also works with Microsoft — its distribution of Hadoop runs on Windows as well as Linux — and it has reseller agreements with Teradata and SAP.
The deal also brings the total of investments in Hadoop companies this year to nearly $1.2 billion: Chip giant Intel led a $900 million round in Cloudera in March, and last month MapR landed $110 million in a round led by Google Capital.
The HP-Hortonworks collaboration, though much smaller, is similar in some respects to the huge stake Intel took in Cloudera earlier this year. For its money, Intel got control of about 18 percent of Cloudera, while Cloudera got visibility into future Intel chip designs and a partner that could help sell into large enterprise accounts.
HP will have access to many of those same customers, but will be able to offer a combination of Hadoop plus Vertica software, hardware to run it all, and service and support.
Prior Hortonworks investors include Tenaya Capital, Index Ventures, Benchmark, Dragoneer Investment Group and Yahoo.