MapR Technologies, the San Jose, Calif.-based startup that sells it own Hadoop distribution for analyzing large volumes of unstructured data, has raised a $20 million Series B round from Redpoint Ventures, Lightspeed Venture Partners and NEA. MapR has been red-hot in the big data space since launching in the spring, positioning itself as an alternative to Cloudera, and the new money proves that MapR has some believers.
Prior to MapR, the Hadoop space pretty much consisted of the core Apache Hadoop distribution (along with various related projects) and Cloudera’s enterprise-hardened, and all-inclusive, version of the same code. Cloudera used the relative lack of competition to establish itself as the Hadoop company, but everyone knew Cloudera wouldn’t have the spotlight to itself forever with interest in big data, and especially Hadoop, growing so fast.
Before MapR launched, Cloudera also took the lion’s share of VC capital going toward Hadoop, raising $36 million since 2009. Its $25 million round that closed in October 2010 was the most a Hadoop startup has seen by a long shot, until MapR’s $20 million today.
MapR has garnered so much praise because its Hadoop distribution aims to solve many of the performance problems that bog down the tech. The company actually offers two distributions — M3 and M5 — with the big difference being advanced reliability and storage features in the paid M5 edition. Storage giant EMC believes so much in MapR’s approach to Hadoop that its Enterprise Edition Hadoop distribution is based almost entirely on MapR’s M5 offering.
MapR Co-Founder and CEO John Schroeder told me he thinks his company already has a couple-year lead on the Hadoop distribution market in terms of technology, and it will use the new money to expand on that lead. Whether its technology lead is that big is debatable, but creating its own IP certainly puts MapR in an interesting position because it’s not beholden to Apache for new capabilities as Cloudera is.
The emergence of Yahoo spinoff Hortonworks no doubt will help advance the pace of Hadoop innovation within Apache, but even that work will float all boats. Because Apache Hadoop is open source, MapR is free to use — and does use — that code just like everyone else. And, in fact, MapR is an Apache Hadoop contributor.
We’ll have to see if MapR is ultimately able to make a dent the significant mindshare lead that Cloudera has spent years building up. Cloudera has been an avid Hadoop evangelist, and has integration partnerships spanning the IT landscape from hardware vendors to BI software. But if VC funds and early buzz are any indication, MapR looks poised to do just fine.
Image courtesy of Flickr user zzzack.