Reaction to Hortonworks-Cloudera Mega Merger
“I didn’t see this coming.” That was a common reaction to yesterday’s news that Hortonworks and Cloudera are combining forces in a blockbuster $5.2-billion merger. Sentiment was mostly positive, especially among people who worked with the two vendors, but questions remain about the merger’s impact.
The two Hadoop distributors announced the unprecedented merger Wednesday afternoon after the stock markets closed. The combined company will have $720 million in combined revenue, 2,500+ customers, and more than $500 million in cash, the companies said.
While the merger is being touted as a combination of equals, it’s apparent that Cloudera will have a larger stake. The combined company will be called Cloudera, and Cloudera board members will occupy five of the nine seats (the 10th member will be selected after the deal is closed, slated for the first quarter of 2019). Tom Reilly, CEO of the current Cloudera, will be CEO of the post-merger Cloudera.
News spread quickly across the Internet, in particular on Twitter, where big data leaders shared their opinions. “This makes a ton of sense,” wrote Jay Kreps, the co-creator of Apache Kafka and CEO of Confluent. Kreps previously has been critical of Hadoop, and told Datanami in 2017 that it was “a very complicated stack to build on.”
“We’re putting the band back together,” wrote Doug Cutting, Cloudera’s chief architect, who co-created Hadoop and then joined Yahoo! to put the fledgling clustering solution into production for the first time in the mid-2000s.
“Interesting development,” wrote Uber’s chief architect Suresh Srinivas, who is a co-founder of Hortonworks and a core contributor and committer to the Apache Hadoop project. “A lot of overlapping products and project consolidation to unfold. Should be good for open source projects with some winners and losers.”
“The merger is great news for consolidation of the market,” says Tendü Yoğurtçu, the CTO of Syncsort, which is partners with both Hadoop distributors and is a major contributor to the open source Apache Hadoop project.
The merger will “bring together the strengths of each company into one single entity that will be able to advance its solutions and the industry’s maturation far faster than either could alone,” Yoğurtçu says. “Hortonworks clearly put focus on IoT and streaming use cases whereas Cloudera put focus on data science use cases, strengthening the platform with machine learning and artificial intelligence.”
“We are not surprised by this merger,” says Ashish Thusoo, CEO and co-founder of Hadoop-as-a-service provider Qubole. “We see it as something of a swan song for the two declining legacy players in the market, and proof that the on-prem data world is becoming obsolete. The market is evolving away from the Hadoop vendors – who haven’t been able to fulfill their promise to customers – toward cloud native options that are much more nimble and better suited to solve the big data challenges of businesses.”
We’re putting the band back together.
— Doug Cutting (@cutting) October 3, 2018
But not everybody was thrilled with the news. “I can’t find any innovation benefits to customers in this merger,” said John Schroeder, CEO and chairman of MapR Technologies, which develops a proprietary platform that’s based in part on Apache Hadoop and competes directly with Cloudera and Hortonworks.
Schroeder says the merger is all about cost cutting and rationalization. “This means their customers will suffer,” he continues. “MapR has been innovating and delivering a better data platform for years, and we continue to see Cloudera and Hortonworks customers move to MapR. We support a broader set of workloads from analytics to ML and AI, already provide hybrid and multi-cloud, and deliver containerization with Kubernetes. Our customers are delivering their vision today with MapR.”
Milind Bhandarkar, who was a member of the original team at Yahoo that implemented Apache Hadoop and the founder and CEO of in-memory computing platform maker Ampool, had some good analysis of the ramifications of the merger.
The merger was “inevitable” as both companies’ business models were based on packing ASF projects, Bhandarkar wrote on his Twitter page.
“There will be short term pain for all teams at CLDR & HDP. But, if done right, this will be viewed as the best thing that ever happened to Hadoop,” he wrote. “This will kill the ambiguity in the market, and be detrimental to those who were superficially bridging the gap between these two vendors by creating a unified layer as the only differentiation.”
The merger raises interesting questions about what the Hadoop stack will look like in the future, as Hortonworks and Cloudera backed different sub-projects and products. For example, Hortonworks backed Hadoop projects like Ranger, Hive, Ambari, Atlas, and NiFi for security, SQL processing, management, governance, and stream processing, respectively, whereas Cloudera backed Sentry, Impala, its own Cloudera Manager product and other partner products for those functions.
Over the years, Cloudera and Hortonworks had battled fiercely over many things, including which contributed more to the various open source projects that make up Hadoop and which company adhered closer to open source principles in general. Both companies employed critical open source contributors to various projects. While Cloudera had some proprietary offerings, Hortonworks had none.
Wall Street seems to like the merger. Industry analyst Jack Andrews of Needham wrote that the two companies go together “like peanut butter and jelly,” according to a story in Barron’s. “This deal makes all the sense in the world and frankly should have happened a year or two ago,” Rosenblatt analyst Marshall Senk said, according to the Barron’s story.
Stock prices for both Hortonworks (NASDAQ: HDP) and Cloudera (NYSE: CLDR) were up nearly 10% today, adding hundreds of millions of dollars to the two companies’ market capitalizations (currently $2.83 billion for Cloudera and $1.95 billion for Hortonworks). That’s a good sign that investors have given their blessing to the proposed merger, which still must clear shareholder votes and gain regulatory approval.