Making good on its promise from earlier in the year, Pivotal has released as open source the distributed in-memory database that powers GemFire, a featured part of Pivotal's Big Data Suite Hadoop product.
It's another step on Pivotal's road toward building an open source base for its Big Data Suite rather than keeping them on a proprietary leash. However, Pivotal still sees ways it could monetize its Hadoop products -- even as advances in open source squeeze companies with proprietary offerings.
Pivotal announced back in February that it would open-source all of Big Data Suite, which includes GemFire, the Greenplum analytics database, and the Hawq data-querying system. That said, GemFire will continue to exist as a commercial project.
Pivotal's plan is to offer a commercial distribution of GemFire, with cost-plus features of interest to enterprises, such as replication between multiple data centers, or stream-processing technology. Geode, the new open source project derived from GemFire's core technology, will be offered to the ASF (Apache Software Foundation) for incubation. (Pivotal also plans to ramp up its membership in the ASF.)
The audience for the full range of GemFire's features consists mainly of paying enterprise customers, so Pivotal stands to lose little upfront by turning the core of GemFire into an open source project. But the company is clearly aware that today's paying feature may not persist as tomorrow's revenue stream.
Roman Shaposhnik, director of open source strategy at Pivotal, described this movement as a "forcing function" that aids enterprise adoption of open source-based products as much as it encourages further innovation of proprietary products based on them.
"You may have a killer feature today," he said in a phone conversation, "but if the bulk of the core of your product is actually coming from an open source project, you can be guaranteed that killer feature will be re-implemented by the open source community at some point."
Shaposhnik also claims Pivotal "has more of an appetite" for true open source engagement than might be found elsewhere in the industry -- to compete at the product level with rivals, but to still be transparent and committed to the open source collaboration model.
For a contrasting example, Shaposhnik offered Cloudera, where he previously worked and which he described as "trying to be more controlling" with its Hadoop projects. He characterized Cloudera's Impala as "read-only open source, where the company publishes source code on GitHub, but is not really inviting any collaboration on that source code."
Pivotal is also trying to make its stance on open source clear by way of the Open Data Platform, an initiative to create a common-core version of Hadoop that can be reused by other vendors. Not everyone -- Cloudera in particular, but others too -- is on board with that project, meaning Pivotal may find more direct success with open source by opening and donating its own projects over time.
[An earlier version of this article incorrectly identified Fuel as a Cloudera project.]