For years, the open source Hadoop project for managing huge amounts of data has been in the spotlight, but it hasn’t necessarily lived up to its hype. Now it might have finally found its killer app in the Internet of things.
As more companies collect data from devices such as industrial sensors, mobile phones and cars, companies selling Hadoop software can try to swoop in and make sales. During the past couple of quarters, revenue at one of those companies, Hortonworks (HDP), has been growing rapidly—more than doubling on a year-over-year basis—partly because the Internet of things is driving demand for its software, CEO Rob Bearden told Coins2Day.
“Absolutely, beyond a question of a doubt, it is a functioning, seriously monetizable, fast-growing market,” Bearden said. “…If [companies] can get visibility in real time, and act and react in real time, it completely changes the dynamic of that transaction.”
Earlier this summer, Hortonworks helped advance its story for collecting these new data types by acquiring a startup called Onyara. It spun out of the National Security Agency, and was commercializing NSA-built software used for moving data from its source to other systems that might need to process it.
“The reception in our customer base has been astonishingly positive,” Bearden said of the Onyara acquisition. “Some of the biggest enterprise on planet earth are instantly aligning with this.”
Other open source technologies including Apache Kafka and Apache Spark have also helped to make Hadoop better suited to the Internet of things. They have made getting data into Hadoop and then processing it much faster than before.
For companies trying to profit from Hadoop, the rush to connect devices and analyze their data could not be better. Despite the early hype, fewer customers adopted Hadoop than expected, which limited the market for companies hoping to sell them additional software and support (the core software is open source, even in its commercial form).
Annual revenue at the two most well-known companies in the space, Cloudera and Hortonworks, is around $100 million. But that’s tiny compared with the billions brought in every year by legacy data-management companies such as Oracle (ORCL) and even Teradata.
Analysts point to several factors for the underwhelming adoption, including the complaint that the original set of technologies were quite slow, and therefore helpful for only a limited set of uses. Bearden said that although public market investors are beginning to understand the company’s open source business model and the associated revenue flows, other influential voices still do not.
“The flipside is the financial journalists clearly didn’t go to accounting school,” he said. “… We have not been able to get the level of education that we need there.”
If the Internet of things can help Hortonworks, and its peers Cloudera and MapR (which are eyeing initial public offerings), start making more money, perhaps they’ll have to do less explaining to do. The numbers will be able to speak for themselves. Or perhaps everyone will have to come around to Bearden’s view of the world whether they like it or not.
“The enterprise software companies are in meltdown mode, in general,” he said. As more open source companies go public in the next several quarters, he added before referring to critics and the open source business model, “They’ll be forced to learn it and understand it and absorb it and get it right.”
For more about the business value of data analytics, watch this Coins2Day video:
Sign up for Data Sheet, Coins2Day’s daily newsletter about the business of technology.
