Follow Datanami:
November 7, 2018

MapR Targets Cloudera-Hortonworks Customers with ‘Clarity’ Release

MapR Technologies went on the offensive today against its Hadoop rivals Hortonworks and Cloudera with the launch of a “Clarity” release of its MapR Converged Data Platform, as well as a new offer for a free big data assessment from its professional services team.

Cloudera and Hortonworks agreed to a blockbuster merger last month that will result in a single company with about 2,500 customers and $730 million in annual revenue. After the merger is complete, which is expected in the first quarter of 2019, the new company will deliver a pair of “Unity” releases to rationalize the disparate products and open source projects that comprise their respective Hadoop distributions.

It was only a matter of time before MapR Technologies, which has traditionally been viewed as the third leg of the Hadoop distribution stool, went on the offensive and attempted to capitalize on the merger of its two rivals and the impact that it will have on its offerings. That time is apparently now.

Bill Peterson, vice president of industry solutions at MapR, says Cloudera and Hortonworks customers are concerned about the disruption that the merger may have on their data platforms.

“We have very high confidence on the early stuff we’re hearing form sales reps in the field, that there is some confusion and concern in their customer base,” Peterson tells Datanami. “It creates a bit of a storm for customers, and we see that as an opportunity for us, that organizations need to be able to find a clear path forward. That’s what we’re announcing.”

MapR hopes that its “Clarity” release resonates with HDP and CDH customers who are feeling fear, uncertainty, and doubt about the “Unity” release. And while MapR isn’t above flinging some FUD while its rivals’ plans are still up in the air, there is actually some real meat to this announcement.

For starters, on the AI front, the company is announcing that it now “directly supports” deep learning libraries, including TensorFlow, Caffe, and PyTorch. That will bolster support for the new RAPIDS software that MapR launched last month with its GPU partner, NVidia.

On the cloud front, the new Clarity release offers an S3-compatible API. This will enable MapR customers to access data on their clusters using the S3 API, to go along with APIs for HDFS, POISX, NFS, and Kafka. (It’s previous support for S3 allowed cloud-based versions of the MapR platform to access S3-compatible data stores, Peterson says.)

On the IoT and streaming data front, the Clarity release adds support for Kafka SQL, which allows customers to use SQL to query data streaming across the Kafka bus. “Developers can use that data and launch a streaming application using Kafka SQL, right from within MapR,” Peterson says.

Lastly, the company has launched a new “step up” data assessment, whereby MapR will send a team of technical experts to a Hortonworks or Cloudera customer (or a prospect) site for a week.

There’s nothing in the new MapR release about containers. However, MapR has supported a Kubernetes volume driver for some time, which Peterson points out is further along than Hortonworks and Cloudera have gotten. Executives with those two companies have stated their intentions to get their big data distributions running in Kubernetes, but admit there is probably one to two years of engineering work to do first.

MapR plans to make hay on that point. “The vendors say they want to get to a ‘containerized architecture for big data management,'” Peterson says. “We’ve had that for over two years. Let’s provide these customers the ability to go with us right now, because there’s going to be a lot of engineering work that needs to happen to rationalize all this stuff.”

As far as choosing among competing CDH and HDP products – such as Ambari versus Cloudera Manager and Ranger versus Sentry – MapR also plans to capitalize on uncertainty in the market to cast its own proprietary platform in a better light.

“There’s a lot to get done inside of a year and there’s a lot of questions,” Peterson says. “If I’m a customer who chose Ranger for security and you’re telling me that I have to change? I don’t want to change. That’s why I bought it because I was happy with it and it’s running my security, which is a big deal.”

Peterson cast doubt on Cloudera’s and Hortonworks’ plan to start the rationalization process within a year with the “Unity” release.

“I haven’t talked to anybody who thinks it’s remotely possible for next year,” he says. “If you think about it from an engineering point of view, it’s going to be really difficult to do innovation in that time until that Unity release goes out the door because there are so many competing issues to deal with.”

Related Items:

New Cloudera Plots a Course Toward a Unified Future

Is Hadoop Officially Dead?

MapR Customers Share Big Data Experiences