Inspired by Linux, Pivotal to open-source bulk of big data suite

Pivotal will make the majority of its big data suite open source, drawing inspiration from the Linux concept.

The decision will see “core” code for HAWQ, Greenplum Database and Pivotal GemFire released to anyone who cares to footle with it, but plenty of the footling looks like it will be done by a new group calling itself the Open Data Platform (ODP).

The ODP's Platinum members are Pivotal, GE, Hortonworks, IBM, Infosys, SAS and “a large international telecommunications firm”. AltiScale, Capgemini, CenturyLink, EMC, Splunk, Verizon Enterprise Solutions, Teradata, and VMware are signed up as Gold members.

Sundeep Madra, Pivotal's CTO and veep for matters mobile, said the ODP is going to try to do for Hadoop what Linux did for Unix: give it a unifying force that makes it easier for organisations to adopt.

To that end the ODP will create “a tested reference core of Apache Hadoop, Apache Ambari and related Apache source artifacts”, an effort intended to make it easier for users to know where to start with Hadoop.

Open-sourcing its products, Madra added, makes sure Pivotal is aboard the open source train before its momentum moves the market away from the company’s current offerings.

Organisations are making technology adoption decisions based in part on newsgroup activity, he said, as posts can be taken as an indicator of innovation.

The thinking behind the decision to open source the tools seems to owe a lot to the 1970s maxim of “If you love something set it free. If it returns, it is yours. If it does not, it never was” as Pivotal feels that by open-sourcing the applications listed above things will return asking for consulting help and commercial products.

It's also not giving away all its code – the exact language it's using is that the products listed above will henceforth be “based on an open source core”. So it will still have plenty to sell and support. And if the community gives the products' “core” some new useful features, Pivotal will be there to clean up.

Madra said the ODP is open to new members. Which, pardon the pun given it focuses on Hadoop, is the elephant in the room, as Intel and Cloudera aren't on the organisation's roster. Cloudera's aims are rather similar to the ODP's, namely to create a cut of Hadoop most people can agree on as a foundation for future big data work.

Cloudera's got bushels of Intel cash on its side. ODP has a membership roster that probably represents a sufficient counterweight to make this market even more interesting.

And in the background of all those moves is Google's mid-2014 decision to move away from Map Reduce, Hadoop's foundational trick, in favour of a new alternative. ®