UPDATED 08:00 EDT / FEBRUARY 26 2015

Harmony or Hail Mary? Experts debate need for the Open Data Platform

hadoop_elephantWill the Open Data Platform (ODP) association bring harmony to the Big Data market or dive a wedge into it? Experts have been debating that point since the blockbuster announcement last week, with the consensus being that, at best, ODP has a long and difficult road ahead.

The new consortium, which is led by Pivotal Software, Inc. and Hortonworks, Inc., says its goal is to provide a big data kernel in the form of a tested reference core of Apache Hadoop, plus related Apache open-source software like Apache Ambari.

The ODP is backed by a roster of heavyweight names like General Electric Co. and IBM, but some key companies – most notably leading Hadoop pure-play Cloudera, Inc. – are refusing to throw their hats into the ring. Industry consortia can succeed without the support of everyone, but at this point the Hadoop community appears to be deeply divided over the need for this organization.

Simplifying Hadoop

 

ODP’s initial focus will be to build a core kernel that members will use to build, package and distribute their offerings, said Sundeep Madra, vice president and general manager of the Data Product Group at Pivotal, in an interview on CIO.com. The idea is to simplify upstream and downstream qualification efforts — creating a “test once, use everywhere” core platform that could eliminate fragmentation in the space. Applications and tools built on the ODP kernel should integrate with and run on any compliant system.

The ODP is also an effort by a number of large enterprise software vendors to accelerate the adoption of Hadoop, said Jeff Kelly, principal analyst at Wikibon. Because the Hadoop platform offers little value without analytics and applications running atop it, ODP’s main backers are hoping to agree on a single reference architecture so they can focus on making their money elsewhere, he said.

“The big players – IBM, Pivotal, Teradata and EMC- recognize they will make money up the stack,” Kelly said. “It’s in their interest to coalesce around a common Hadoop ‘kernel’ that instills confidence in enterprise practitioners, reassuring them that they won’t invest in one version of Hadoop only to have a different version emerge as the standard.”

Hortonworks and Pivotal representatives further say ODP will help foster innovation on both the customer and vendor sides by eliminating the risk of platform fragmentation. In a special segment on theCUBE, Leo Spiegel, senior vice president of corporate development at Pivotal, told host John Furrier the aim was to bring together “a bunch of companies to help accelerate the big data space.”

“We’re going to develop a standard CI suite that people can download and use,” Siegel said. “We’re also going to work on developing the ecosystem and we’re going to work on education.”

Shaun Connolly, vice president of corporate strategy at Horntonworks, said the ODP will also help to foster innovation because of its “common substrate” into which customers can plug whatever “additional elements” they need. “It’s more of a consumption thing than a production thing,” he said.

ODP does have an impressive list of founding backers. Altiscale, Inc., Cap Gemini S.A., CenturyLink, Inc., General Electric Co., IBM, Infosys International Inc., SAS Institute, Inc., Splunk, Inc., Teradata Corp., Verizon Communications, Inc. and VMware, Inc. have all signed up as founding members of the consortium.

But some even more important players have opted out, chief among them Cloudera and MapR Technologies, Inc., both of which compete directly with Hortonworks. Also missing are big cloud services providers like Amazon Web Services, Google Cloud Platform and Microsoft Azure; Hadoop hardware providers like Hewlett-Packard Co., and Dell Inc.; as well as NoSQL database providers like MongoDB, Inc..

Standards, what standards?

 

mike-olson-cloudera

Cloudera’s Olson: “There is no problem ODP solves.”

Cloudera co-founder and Chief Strategy Officer CEO Mike Olson minces no words in outlining why he thinks OPD is a bad idea. Olson, who wrote a blistering blog post shortly after last week’s announcement, said ODP was built purely to serve the vendors involved in it. In an interview with SiliconANGLE, he charged that the group will actually cause more problems than it solves.

ODP is focused on standardizing some of the most stable and well-established components of Hadoop, including the Hadoop and MapReduce cores, which are both over 15 years old. He said there’s no need to create a new standards body for these components when the Apache Software Foundation (ASF) is already doing a perfectly good job.

“Go find a problem that ODP solves,” he said. “No partner tells us that standards in MapReduce is a problem. There is no problem ODP solves.”

Asked if there was any possible scenario under which Cloudera might have joined the ODP, Olson reiterated that the effort is duplicative and unnecessary, a desperation play by Pivotal to buy its way into a market leadership position without earning it through participation. “This is an attempt by guys who don’t participate in the open source community to wield their money,” he said. “The ODP intends to usurp some of the activity of the Project Management Community [within ASF] in driving a standard. “A better thing would be to get involved in that.”

While it’s no surprise that Cloudera is less than enthusiastic about the creation of a second standards body in the Hadoop industry, it isn’t alone. Analysts have also questioned the wisdom of the idea

‘Declare victory and retreat’

 

Gartner Inc.’s Merv Adrian and Nick Heudecker agreed that ODP could just be a way for Pivotal to “reduce its investment in a failing effort to build a proprietary way to capture a slice of this trend. Declare victory and retreat.”

They may have a point given Cloudera’s and MapR’s absence from the ODP party. Cloudera has forged a lead in the Hadoop race, announcing revenues of $100 million last year and saying that its recurring subscription revenue rose by 100 percent. While the private company’s claims are impossible to verify, Olson told SiliconANGLE that more than half of that revenue came from licenses, rather than services. In contrast, Hortonworks recorded $46 million in 2014 revenue in its earnings report released on Tuesday, but said growth prospects are strong.

“ODP positions Hortonworks as the Hadoop arms dealer for the other players,” wrote the Gartner analysts. “Basing an open data platform on a single vendor’s packaging casts some doubt on ‘open.’” The analysts concluded that ODP isn’t designed so much for the customers but as a way to help Hadoop’s also-rans stay relevant.

That opinion was shared by analyst Curt Monash, who dismissed the effort as “a face-saving way to admit that IBM’s and Pivotal’s insistence on having their own Hadoop distributions has been silly.”

Kelly agreed that since the ASF already takes responsibility for developing the Hadoop kernel, its questionable whether another industry consortium is needed to do more or less the same thing.

“The ODP says fracturing of Apache Hadoop around approaches to SQL-on-Hadoop, for example, is slowing enterprise adoption,” Kelly said. “But critics of the ODP say such ‘fracturing’ is simply the way an open source community develops and determines the best approaches to a given challenge. Time will tell who is right.”

Given all this, it’s understandable while many people say the ODP is mainly a competitive play by Cloudera’s rivals to deal that company and its principal backer – Intel Corp. – a setback.

“Cloudera itself recognizes it will make the majority of its money with value-add analytics and database software while Intel makes money on the hardware side,” Kelly said. “The ODP is an attempt by its members to isolate them.”

But despite the criticism levelled against it, Pivotal’s Spiegel insisted the ODP is more about finding ways to work together on behalf of customers. As Big Data becomes more strategic, multiple versions of key technologies are sprouting up and creating complexity, he said. ODP will bring consistency to the way these technologies integrate.

“If you’re an application vendor, you want to know the next version of your product doesn’t have to be tested on all of these different distributions,” Spiegel said. “Things need to go faster, so why don’t we find a way to work together?”

Cooperation might well help to speed things up in the Hadoop ecosystem, but without Cloudera’s and MapR’s involvement, ODP will be challenged to bring about the consistency it claims to seek.

“This simply institutionalizes a dichotomy in favor of a few favored players,” Gartner’s Adrian and Heudecker wrote. “It’s ironic that Hortonworks is one of the founders of an organization that wants to add an anchor slowing innovation in the open source free-for-all it has been the flag-bearer for.”


A message from John Furrier, co-founder of SiliconANGLE:

Your vote of support is important to us and it helps us keep the content FREE.

One click below supports our mission to provide free, deep, and relevant content.  

Join our community on YouTube

Join the community that includes more than 15,000 #CubeAlumni experts, including Amazon.com CEO Andy Jassy, Dell Technologies founder and CEO Michael Dell, Intel CEO Pat Gelsinger, and many more luminaries and experts.

“TheCUBE is an important partner to the industry. You guys really are a part of our events and we really appreciate you coming and I know people appreciate the content you create as well” – Andy Jassy

THANK YOU