When it comes to Hadoop, there are varying opinions about its relevance and compatibility with enterprise systems. SiliconANGLE held a panel at BigDataNYC 2015 to discuss the Open Data Platform initiative (ODPi) from the perspective of the enterprise.

The panel, moderated by Dave Vellante, of SiliconANGLE and Wikibon, included Shaun Connolly, vice president of corporate strategy at Hortonworks, Inc.; John Fanelli, vice president of marketing at DataTorrent, Inc.; and Michael Maxey, senior director of corporate development and strategy at Pivotal, Inc.

Adoption numbers

Connolly started the conversation by saying, “From an adoption perspective, things are progressing well for Hadoop in the enterprise.” He sees new solutions developing but as he put it, there is a “data dial tone” and there are still kinks that need work. But, for him, the platform is very relevant.

In a recent study by Gartner, Inc., 44 percent of 300 people polled are planning or investing in Hadoop. Maxey sees this as the standard market adoption curve, and he feels that this is the point where things stabilize.

Pricing structures

The topic then turned to pricing when Vellante referenced Wikibon’s George Gilbert’s quote that there is a “slow motion collapse in infrastructure software pricing.” The question: How are you evolving in this environment?

Maxey responded by saying that while prices are going down, his company is growing revenue in other ways. He said, “Consumption is not where it should be, but when prices are down at the lower layer of the stack it drives innovation above.”

Customer feedback

What issues are customers voicing in the marketplace? Connolly detailed some of the problems that are delaying enterprise adoption. He explained that his customers want ease of use and the ability to build applications. He continued by saying that even if you build an app, there is a need to create excitement about it.

The road to making open source ready for the enterprise is still a bit bumpy as there are procurement, certification and compatibility issues. Connelly said, “There needs to be a focus on enabling apps.”

One size does not fit all

When asked if Hadoop will disappear or become invisible, the consensus was that there is a need to tap into the value of data in Hadoop storage and make it consumable for all communities.

Spark, lauded by some as the next big thing in analytics during the event, is also causing some debate. The panelists all agreed that you should not put all your eggs in the Spark basket. Fanelli said, “There is not a one size fits all in the process. You make design decisions that are tradeoffs. Enterprise will run multiple paradigms and engines should all be compatible.”

Connelly felt that there is an interesting user community around Spark as an analytic engine. He said, “I like that from an application and enable perspective.” However, he also believes it will take many strategies, not just one.

The general response to the debate was that there has to be more than one solution to enable enterprise to derive value from its systems.

Revenue opportunities

Vellante, who was quoted early on during BigDataNYC 2015 as saying that the ecosystem is “crowded, over funded, profitless but with great potential,” clarified his point by saying there is a silver lining by underscoring “great potential.” He ascertained that while the quote seemed pessimistic, he truly believes that profitability is forthcoming.

Fanelli stated, “It is still a good business, and everyone is betting on the potential.” He considers the investment made will change the way enterprise does business.

The market is moving quickly. According to Connelly, “From a cost angle, you can save money on Hadoop deployment, but that is not interesting.” He considers the other innovations that will come from using Hadoop down the line.

Maxey simply said, “If we build it, we will figure it out later,” explaining the drive for innovation will eventually turn profitable.

Making things work

There were a few questions from the audience, and all had the same underlying theme. The main question was: How are you going to take open source and make things work?

The answers varied, but to sum up the general consensus, Fanelli said, “Open source is for developers, and eventually the work they do becomes deliverable to customers. As a stack, it all works together. There are many solutions that work, and you can see plenty of examples at this show.”

Watch the full panel discussion below, and be sure to check out more of SiliconANGLE and theCUBE’s coverage of BigDataNYC 2015.

Photo by SiliconANGLE