Pivotal, the EMC-VMware spinoff that aims to bring a re-architected, battle-tested Hadoop to the enterprise, has incorporated the latest perks of the Apache Hadoop 2.2 into its new Pivotal HD 2.0 release, along with the broad release of GemFire XD, the in-memory database that integrates with it.
What Pivotal hopes to do with this stable of technologies, is to offer big companies a way to build virtuous feedback cycles from the data their products generate every minute. A cell phone carrier, for example, could use the big data and analytics capability of PivotalHD and GemFire HD to determine the fastest routing of a call and — should that call degrade or fail — feed that information back into the process to fix the problem . And, in a perfect world, issue a timely apology or make-good to the customers affected.
The new benefits, courtesy of that Apache Hadoop 2.2 update, is that Pivotal HD 2.0 will now support NFS and snapshotting, meaning enterprise customers can roll back their clusters in the event of a problem.
Pivotal may not be the first commercial entity to embrace the Apache updates, but it claims to do most rigorous testing of the open-source code. “We take the the Apache distribution and harden it doing our own QA and regression testing but also roll it out on a 1000-node cluster so our testing is done at scale,” said Michael Cucchi, senior director of product marketing for Pivotal.
The new Pivotal HD 2.0 release also integrates with GemFire HD for real-time analytics but also with GraphLab graph analytics and adds an improved HAWQ SQL query engine for analytic workloads.
None of this is surprising — Pivotal is knitting together technologies from a series of EMC and VMware acquisitions into software stack that should make it easier for corporate customers to deploy Hadoop at scale and get real value out of it. The HAWQ SQL query tool, for example, comes from EMC’s 2010 acquisition of Green Plum; Gemfire from VMware’s buy of GemStone that same year
Initially Pivotal HD software is available on its own to run on bare-metal or in VMware environments or bundled with hardware as an appliance, but over time users who want to run Pivotal HD on Amazon Web Services or other public clouds will be able to do that as well, Cucchi said.
This is indeed a huge opportunity, and one that is not lost on legacy IT giants including IBM as well as newer look Hadoop powers like Cloudera and HortonWorks, so there will be plenty of contention for those enterprise accounts Pivotal is targeting.
Watch for Pivotal CEO Paul Maritz to talk more about this product and what usage it will enable Wednesday morning at our Structure Data conference.