Cloudera Soars on a Mountain of Cash and Opportunity

The Hadoop firm has attracted mindboggling funding on the back of its promise to unpick Big Data

Even in today’s wacky world of tech financing, Cloudera raised eyebrows to nearly unprecedented heights when it took $740m in funding from Intel for an 18% stake. Add in the $160m the commercial Hadoop tools developer and seller raised in the same round and you have about $900m — and that’s not counting the relative chump change ($140m or so) that came previously. That is an enormous sum for a startup and begs the question: what’s so hot about Cloudera?

To find out, I talked to chief strategy officer Mike Olson by phone from the company’s HQ in Palo Alto, California. First up: that great lump of cash…

“It’s $900m and I understand that’s a jaw-dropping number but it’s really not about the money,” he says. Instead, Olson insists, it’s a statement to enterprise buyers that Cloudera has been ‘de-risked’ and there’s a major platform available for those setting up on the Hadoop stage and seeking answers to questions mined from information seams via Big Data.

Olson frankly confesses that one advantage of the deal was that it was with Intel. As part of the agreement, the chip maker took its own Hadoop distribution off the market and gave Cloudera the inroads it had made into India and China.

Intel also helps Cloudera open up a new partner ecosystem “because there are lots of people who don’t like software companies but there’s nobody who doesn’t like Intel.”

Well, that might only be true up to a point (ask AMD, nVidia or other silicon rivals how much they love INTC) but the basic argument is sound: Intel is an intersection for the ICT industry, the provider of silicon brains, communications, reference architectures, graphics engines and more to the rest of the sector.

Thirdly, Intel provides direct access to what’s happening in core hardware.

“The lines between CPU, disk, memory and networking are going to shift through the silicon that changes how we build racks of servers,” says Olson. And when it comes to taking advantage of the new directions for bits and bytes, “we will be the authors of those enhancements. This agreement will drive adoption of our platform and concomitant adoption of servers.”

There are other, obvious opportunities too to accelerate growth and extend reach. Or, as Olson puts it: “900 million bucks wasn’t what was driving us but the ability to do mergers and acquisitions, hire people and spread geographically were.”

Cloudera has other backers providing more than just dough too, and they include the company that invented MapReduce [Google] and [indirectly, via Michael Dell’s MSD Capital] the world’s largest server maker, Dell.

Taken together then, this is a plan to make Cloudera the 800-pound gorilla of Hadoop and de facto choice for risk-averse CIOs who like the appeal of Hadoop but need to a big company standing behind it. And surely, it will also help to pin down Hortonworks, often regarded as the Tweedledum of Hadoop to Cloudera’s Tweedledee (interview here). However, Olson downplays this.

“There are many, many companies shipping competitive distributions, just as previous technology [breakthroughs] like SQL and search spawned a bunch of copycats,” he says. “Our plan is to innovate and build our own products on the platform. It’s not a battle between two companies.”

Olson defines Hortonworks as offering the “low-cost” Hadoop option while IBM, Pivotal and others offer proprietary “cobbled together stacks”. He’s scathing about companies that are “flat-out absent and not developing code for the open source community”. Cloudera’s position, by contrast, will be to offer the most innovative open-source solutions for CIOs who don’t want to risk lock-in, he says.

The prize is very big indeed but nobody knows how big.

“Data is inundating every organisation and that’s a huge challenge because it’s hard to capture,” Olson says. “But if you can explore risk, opportunity, fraud, sentiment, you can make better decisions, you can save cost and drive the top line.”

Olson was CEO at Sleepycat Software, one of the first sizeable open source database companies to sell to an established brand (Oracle), and he compares the Hadoop segment to the early days of relational databases.

“It’s a young market that’s emerging faster than I’ve ever seen. It could be even bigger than RDBMS. When we started [in 2009] nobody had heard of us and we were the evangelists. We were right that this Google idea is destined for enterprise dominance. This is the only bet we’re making and we’re all in. If Oracle can make [tens of billions of dollars] per year on simple numerical data store tables then driving consumer desire from ingesting more data is at least as large.”

A common criticism of Hadoop is that it is short of core features and tools, never mind bells and whistles, but Olson believes that just as SQL was so powerful that it became ubiquitous very quickly, Hadoop can make the same sort of rapid progress without bureaucracy, politics or standardisation processes getting in the way.

“The old standardisation model is based on a flawed assumption,” he says. “You don’t need a special committee. If you look at HDFS, Impala and so on, the communications are crystal clear. If you want to know how something works, just go look at the repository.”

Another point of comparison: Olson talks about the development of the Mosaic web browser, its roots in the consumer internet and compares that to what’s happening in Hadoop. But he says he and his fellow board members have no intention of copying the progress of other companies.

“We’re not trying to build what IBM, Oracle or Microsoft has. We don’t want to pattern this company off another one. We want to build the next Cloudera.”

The promise of Big Data is enormous. There will be discoveries that have been longed for over decades but there will also be the ability “to ask stupid questions” and stumble up on remarkable findings, he predicts. The real value of Hadoop and companies like Cloudera won’t lie in VC stories or eye-popping valuations but in answering some of the most complex questions the world is asking today. These systems will help us not only run better businesses, but discover cures to diseases, understand our planet better and… well, pretty much anything you might want to do. The truth is out there and, with a newfound ability to thresh so much information, we might just find it.


Martin Veitch is Editorial Director at IDG Connect