What can Cloudera do for customers

Big data

The two Hadoop specialists Cloudera and Hortonworks have decided to merge their businesses. The agreement provides for an amalgamation of equal rights. Based on the company's share prices in early October, the new joint venture has a market value of $ 5.2 billion. The Cloudera shareholders will hold around 60 percent of the shares in the joint venture, the Hortonworks shareholders the remaining 40 percent. Both providers want to achieve annual sales of around $ 720 million in 2018 together with more than 2,500 customers. The deal was unanimously approved by the boards of directors. It should be completed in the first quarter of 2019.

Cloudera and Hortonworks grew up in the wake of the Hadoop big data framework, which was published in 2006 and became increasingly popular in the years that followed. In contrast to complex and expensive data warehouse infrastructures, Hadoop, in conjunction with the Hadoop Distributed File System (HDFS) and the MapReduce algorithm, promises simple and fast processing of large amounts of data. Around the framework maintained by the Apache Foundation, a real ecosystem has developed over the years with numerous other tools for data handling - for example the scalable database "HBase" and the in-memory batch processing engine "Spark".

In addition to the specialists, numerous IT giants such as IBM, Microsoft and SAP built interfaces to Hadoop and integrated the solution into their own data management and database suites. Providers such as Cloudera and Hortonworks developed their own distributions of the open source system and enriched their platforms with tools and services.

Next generation data platform

In addition to a free Hadoop version, the "Cloudera Data Hub" (CDH), the company offers its commercial "Enterprise Data Hub" (EDH). Cloudera creates added value for its enterprise customers primarily with additional analytics functions relating to Spark. With its Cloudera Navigator, the Hadoop specialist also offers users a range of tools for handling, managing and optimizing data.

You can find many more articles on the topic of big data here in our online special

With the Hortonworks Data Platform (HDP), Hortonworks also has an open source distribution of Hadoop in its program. Usage is free. If corporate customers want support, they have to pay for it. With the new version HDP 3.0, which was only presented in June of this year, Hortonworks offers its customers automated provisioning tools for its platform in various cloud infrastructures such as AWS or Google. Another innovation is the support of Docker containers. Users can use it to encapsulate individual Hadoop workloads in containers.

With the marriage of convenience, the two companies founded in the USA want to create "a leading global provider of next-generation data platforms," ​​it says. Despite the shared Hadoop history, those responsible do not see any conflicts with regard to the product and solution portfolio. Cloudera CEO Tom Reilly, who will also head the future joint venture, said the two companies complemented each other well. The platforms have developed into different areas of data management over the years. "By merging end-to-end data management from Hortonworks with data warehousing and machine learning capabilities from Cloudera, we will deliver the industry's first enterprise data cloud from the edge to the AI," said the manager.

The next Oracle

Reilly is not lacking in self-confidence: The Cloudera boss told the US broadcaster CNBC that they wanted to be the "next Oracle". That does not mean replacing legacy systems from Oracle. Rather, customers expected new techniques to use machine learning and artificial intelligence to process their ever-growing mountains of data. The appropriate solutions will be offered for this, according to the Cloudera manager.

In order to avoid any uncertainty among existing customers, the existing platforms should continue to be supported in the long term. After the merger, all products will be supported for at least three years, assured Mike Olson, Cloudera's chief strategist. "Every customer who chooses one of the two platforms can be sure of the long-term future of the chosen platform," wrote the manager in a blog post.

However, Olson also made it clear that there should be a uniform platform in the future. It was a lucky coincidence that Hortonworks Version 3 and Cloudera Version 6 were both shipped recently. "We have upgraded our latest products to the latest open source project versions, which means that our lines of development are closer than ever before," states Olson. "We will act quickly after the transaction is complete to bring the two together into a single release." In the long term, this single platform will be expanded to include all functions that both providers have developed separately. The manager promises that customers can safely upgrade from their current Hortonworks or Cloudera installations to the new product.

As part of the merger, Rob Bearden, CEO of Hortonworks, is to be promoted to the board. "Together we are well positioned to continue to grow in the streaming and IoT, data management, data warehousing, machine learning / AI and hybrid cloud markets and to assert ourselves against the competition," said Bearden. First and foremost, it is important to offer a broader range that enables customers to utilize the value of their data.

Both providers are deeply in the red

However, it is by no means certain whether the high expectations will be met. Neither provider has been able to get their Hadoop business rolling profitably until the end of the day. The numbers are similar. Cloudera reported a loss of nearly $ 386 million on annual revenues of $ 367 million for its fiscal year 2018, which ended January 31 of this year. The previous year had sales of $ 261 million and a deficit of $ 187 million. Hortonworks had sales of nearly $ 262 million in 2017 and a net loss of $ 204.5 million. It didn't look any better the year before. In 2016, the revenue of a good 184 million dollars was set against a minus of almost 252 million dollars.

Read more about Hadoop:

Those in charge of both companies admitted mistakes. It was said at the beginning of the year by Clouderas that the focus was too much on customers who represented too little sales. The management of Hortonworks admitted to be too dependent on the US business. The newly merged company is to expand more internationally in the future and align itself globally.

But there are still many question marks. It is not known under which name the merged company will operate or where it will be headquartered. Both providers are currently located close to each other in Silicon Valley.

MapR - the laughing third party?

There is also consolidation in the market. The triad Cloudera, Hortonworks and MapR in the Hadoop concert will become a duet after the merger. MapR, which like Cloudera was founded in 2008, mainly relies on the enterprise edition of its "Converged Data Platform" (CDP). The company is going its own way. With "MapR FS", the platform relies on its own file system and the integrated NoSQL database "MapR DB".

Another core element is an event streaming engine, which should allow users to evaluate streaming data in real time. In addition, with "MapR Edge" there is a much slimmed-down edition that can also run on devices in the Internet of Things (IoT). In addition, the Hadoop specialist has not lost sight of techniques such as Docker containers and machine learning and has integrated them into its own platform.

The MapR management does not want to be worried about the merger of the competitors. There is no need to worry, said co-founder and CEO John Schroeder. "We played a different game than Hortonworks and Cloudera from the start," he told US media. The company has been building its CDP for two years and integrating the individual parts closely with one another. According to Schroeder, the platform was designed from the start to support analytical, operational, and emerging workloads. That is why MapR is already being used productively in Edge and IoT scenarios.

The manager expects some of Cloudera and Hortonworks' current and potential customers to look into the MapR solution. He speculates that the pace of innovation at Cloudera / Hortonworks could slow down in the next few months. "Hortonworks and Cloudera used to be enemies. There could be disputes within the new company - and there will also be layoffs," the MapR boss said.