MapR Tries To Separate From Hadoop Pack With New Streams Product

MapR is one of several companies built on the open source Hadoop platform, and as such it has a bit of competition in the space. Today, possibly in an effort to create some separation from its better heeled rivals, it announced a new product called MapR Streams.

This new product takes a constant stream of data like feeding consumer data to advertisers to create custom offers or distributing health data to medical professionals to tailor medication or treatment options — all of this in near real-time.

Streams let customers share data sources with people or machines that need to make use of that information in a subscription-style model. A maintenance program could subscribe to the data coming from the shop floor of a manufacturer and learn about usage, production, bottlenecks and wear and tear, or IT could subscribe to a data stream with log information looking for anomalies that signal maintenance issues or a security breach.

One particularly interesting aspect of MapR’s streaming product is that it can act as a system of record, creating a persistent record that you can even rewind like a recording to any point in time and review what happened. That means it’s fully auditable and that could prove useful, especially in regulated industries that have to keep track of all transactions as they happen.

The company also announced a new converged platform, providing a way to process multiple kinds of data including files, databases and analytics on a single platform. The platform makes use of not only Hadoop, but also Apache Spark, the open source big data and analytics platform developed at the University of California at Berkeley in 2010.

These moves are an attempt to take the company beyond its roots as a pure Hadoop vendor, says CEO John Schroeder, “The way we have presented our product is enterprise grade Hadoop. That’s still important to us, but we’ve expanded beyond that,” he said.

That could be because of the increasingly competitive Hadoop market.

To give you a sense of the competition fellow Hadoop vendor Hortonworks went public last year. Then there is Cloudera, which got almost a billion dollars in funding in March 2014 and came away with a $4 billion valuation. That kind of action could put pressure on competitors (although Cloudera CEO Tom Reilly told me recently in an interview at the Intel Capital Global Summit that all Hadoop vendors benefit from a healthy ecosystem because no one vendor can support Hadoop alone.)

Regardless, Schroeder rejects the notion that he’s chasing the competition with these updates, saying it’s more a result of customer requirements. “I always run my company based on what customers need, not what the competitors are doing,” he said.

MapR has been around since 2009. It produces its own flavor of the Hadoop big data platform, and makes it available as a free open source product, while charging for other products and services. The company has raised $174 million. Its most recent round was $80 million in private equity in June 2014.