March 20, 2013
The world’s love affair with Hadoop seems to be continuing apace. Both in a funding for startups way and a support from big vendors way. On the funding side, yesterday it was MapR, the Hadoop company pushing a couple of different Hadoop distributions that was announcing a $30M VC round, today it’s the turn of Concurrent who has also closed a round – this time to the tune of $4M. From the large vendor support angle it was Microsoft announcing that Windows Azure now supports the deployment management and use of Hadoop clusters running on Azure.
MapR closed its $30M C round which was led by the Mayfield fund along with participation from Lightspeed, NEA and Redpoint. MapR is reporting that half of their customer base comes from web and cloud based vendors while half come from more traditional sectors – financial, telcos and manufacturing. This is an interesting development as it shows Hadoop having real adoption beyond the usual suspects. MapR’s strategy has diverged from the usual Open Source approach – rather than primarily deriving revenue from support and consulting functions, MapR gets most of its revenue from license fees. It’s fair to say that this approach has created some tension within the Hadoop community – vendors with more traditional revenue streams such as Cloudera and Hortonworks are a little dismissive of vendors that sell a solution with some proprietary components – in the case of MapR this is its proprietary file system that the company claims is much faster than the HDFS that most of its competitors use.
Concurrent is another, less well known vendor. Its product, Cascading, is claimed to be the most widely used and deployed application framework for Hadoop – it has achieved the not insignificant figure of 75000 downloads per month. Concurrent is boasting a customer base including companies like eBay and Twitter – both organizations with issues around data processing and filtering. Concurrent has also recently launched Lingual, an open source project that delivers ANSI-standard SQL technology to build new and integrate existing applications onto Hadoop. Along with its new funding, Concurrent is announcing the appointment of a new CEO, Gary Nakamura, formerly an SVP at Terracotta Inc.
EMC Greenplum rolled out its own Hadoop offering that fuses Hadoop with its own massively parallel processing analytic database. Pivotal HD as the product is called is made up of the Greenplum Database with its own file system replaced by HDFS.
Not to be outdone, Intel also announced its own Hadoop distribution. While the company parsed the decision in terms of creating a highly stable hardware/software solution. The truth is somewhat more nuanced than that and Intel likely sees the benefit that software can bring to a hardware business that is increasingly threatened by commoditization.
Hadoop is of undeniable important to modern organizations – that said there are too many vendors competing for a large, but limited market opportunity – I expect to see some consolidation occur, as well as some larger vendors who have entered the Hadoop world but are not really passionate about the opportunity to downsize their Hadoop initiatives when the next big thing rears its head on the horizon.