Shaun Connolly, VP Corporate Strategy for Hortonworks, discussed the latest Hadoop and Big Data trends, as well as the company’s business model, with theCUBE co-hosts Dave Vellante and John Furrier, live at #BigDataNYC 2013 in New York.

“There’s been a first generation particularly in the Hadoop workloads,” Connolly explained. “We’re now in the next step of innovation.” Hadoop can now run interactive, online streaming. “There are are more changes to go.”

Innovation in monetizing open source tech

Hortonworks’ strategy was always 100 percent open source technology, and the business makes money from “software subscriptions primarily,” Connolly was quick to explain. “We are an enterprise software vendor. This market opportunity around Hadoop and Big Data is so large, our focus has to be on making the market function. Don’t hold back the manageability, it will slow adoption.”

Commenting on the company’s software services business model, Connolly said the business currently has about 200 customers. “Out of that, 70 percent of our business is on software subscriptions, customers using our data platform.” It is a “classic software model, we don’t charge for licenses upfront. Our goal is 80 percent” in software subscriptions in the near future. The company does not charge for a license upfront, but instead charges for maintenance.

Hortonworks not “religious” in the NoSQL space

Debating real time technologies, such as MapReduce, Accumulo, and real time streaming in the Apache software foundation, Apache Storm, Connolly said “it can provide that engine for sensor and machine data” powering use cases in the Telco and healthcare industries.

Accumulo’s value proposition, although similar to other NoSQL databases, appeals to government institutions and insurance and hospitality use cases. Security is part of that value proposition. “We tend not to be religious in that NoSQL space,” a market that’s valued at over $1 billion. “MongoDB’s doing a great job and has a great lead,” but it’s still a fragmented market.

Maturing the open source model

Commenting on Hortonworks’ view on open source, Connolly said “the Apache projects are upstream innovation, then we package stable solutions. Hadoop’s influence on the larger Big Data market is around 50 percent.” It’s not only software, it’s also hardware, infrastructure.

Hortonworks’ app story, Connolly said, was “clearly focused on being a data platform, in many aspects we’re horizontal. We want to publicize and work with customers to get vertical solutions.”

Usign the Apache Storm investment as an example, he added that the company didin’t “want to chase the complex event process market horizontally. We want to target sensor machine data processing on verticals” such as Telco or insurance, going on to say that “very concrete use cases are what we’re interested in.”

The Hortonworks strategy is to listen to the market, listen to the early vertical movement and wrap it up for enterprise use.

Asked about the size and maturity of the company’s deployments, Connolly said they “range currently from tens of nodes up into thousands of nodes” for clients such as Yahoo who has running the Hadoop 2 stack in production for about a year. Renewals show that a year later there is a 3-4x growth in the clusters customers have started with. That puts Hortonworks renewals to over 100 percent. “If we scare away a potential user of the technology, we have lost something,” Connolly added.

Talking about the evolution of the market, Connolly stated that “we’ll see traditional vendors getting very concrete around solution architecture. Hadoop is just one data system in the overall modern data architecture.” More and more data processing engines will run natively in Hadoop.