If you thought the Open Data Platform (ODP) was just a bunch of hype, think again.
This morning at the Hadoop Summit in Brussels, three big data crunching giants — Hortonworks, IBM and Pivotal Software — are making big news.
The three will announce that their Hadoop platforms are all standardized on identical versions of Apache Hadoop 2.6, as well as Apache Ambari, and available on the market. The aforementioned configuration is known as ODP.
“We’re all shipping the same version,” Anjul Bhambhri, vice president, Big Data and Streams at IBM, said during a pre-release interview.
Bhambhri went on to explain that this is an important milestone for enterprises and the greater Hadoop ecosystem, noting, “By standardizing on ODP, customers won’t have to worry about whose Hadoop they will use because now there is only one Hadoop.”
Not only that, but vendors around the ecosystem will only have to validate their products against one core, she added.
That core consists of Apache Hadoop 2.6, inclusive of HDFS, YARN and MapReduce as well as Apache Ambari software for managing Hadoop environments at scale.
There are at least 16 vendors/companies who support ODP including GE, Hortonworks, Infosys, International TELCO, Pivotal, SAS, Altiscale, Capgemini, Century Link, EMC, PLDT, Splunk, Teradata, VMware and WANdisco.
Who’s missing from the list? Two primary Hadoop distro providers: Cloudera and MapR.
“ODP by all appearances seems to be a marketing tactic, a distraction,” Matt Brandwein, Director Product Marketing at Cloudera, told us in an interview shortly after Pivotal Software announced ODP in February.
We couldn’t contact Cloudera for comment on this news because it is under embargo. But needless to say, just 56 days after the initial ODP announcement was made, it looks quite real.
The Door is Open
“We’re still hoping that Cloudera will join us,” said Shaun Connolly, vice president of corporate strategy at Hortonworks, noting that it benefits customers and the greater Hadoop ecosystem if vendors put their individual agendas aside.
Leo Spiegel, senior vice president of corporate development and strategy at Pivotal, concurred.
“To move forward, Pivotal believes that the big data market requires a standard, predictable, and mature Hadoop-based core platform for modern management solutions,” he said. He added that the quick momentum around ODP is already benefiting customers by taking the guesswork out of fragmented and duplicative processes of what works and what doesn’t.
“The only question companies now need to ask is 'Are you ODP compliant,’” he explained.
If the answer is yes, then they can pick and choose between validated solutions.
Bhambhri went a step further, noting that this eases the way in which enterprises can make choices. “ODP gives customers a rich technology palette to work with,” she said, explaining that they could get core Hadoop from Vendor A, a solution for data wrangling from Vendor B, and a machine learning or SQL product from vendor C, and that they would all work together.
“There’s still competition,” she explained, noting that providers could build on top of, or around the core. “And there’s increased customer choice."
“Companies can now mix and match without frustration,” she continued — something they couldn’t do when everyone was working with different versions of Hadoop.
A Matter of Choice
What do customers think about ODP?
“They’re excited,” said Bhambhri. “It removes their weariness around lock-in," which could happen with non-ODP compliant Hadoop distros.
To Hadoop vendors who fear that working with a common core will eliminate their ability to differentiate themselves and win business, Connolly has one word of advice. Relax.
He doubts things will go that way. “If we ease Hadoop adoption, the market will grow and the pie that we all share will be bigger,” he said.
As to Hadoop distro providers and other members of the greater ecosystem that don’t build on ODP, “they’ll cause fragmentation and hold their customers hostage,” said Bhambhri.