Yahoo spinoff unveils Hadoop data analysis system

IDG News Service - Yahoo spinoff company Hortonworks has released a preview edition of what will be a fully open source distribution of the Apache Hadoop data analysis platform, called the Hortonworks Data Platform, the company announced Tuesday.

The company has also started offering commercial support and training programs, designed for system integrators and independent software vendors planning to deploy Hadoop on behalf of their customers.

"We'll use our deep-domain knowledge to empower our partners to solve their customers' problems with Hadoop," said Eric Baldeschwieler, CEO of Hortonworks. He likened Hortonworks business model to Linux system software provider Red Hat, which maintains open source versions of all its software and charges for maintenance and support.

The current pre-production version of the Hortonworks Data Platform (HDP) is now being distributed through the company's Technology Preview Program, available to a select number of users. Hortonworks plans to publicly release the pre-production version of the software within the first three months of 2012.

HDP will be a fully open source distribution, including all the components used for a typical Hadoop deployment, including Hadoop Distributed File System (HDFS), MapReduce, Pig, Hive, HBase and Zookeeper. The package will also include a number of newer enterprise-ready components as well, all open source. One is HCatalog, a metadata management service that can aid in connecting Hadoop with other enterprise information systems. The package will also include the Ambari open source installation and management system for Hadoop clusters.

Last June, Yahoo, along with Benchmark Capital, set up Hortonworks as a stand-alone company dedicated to facilitating the enterprise use of Hadoop. Yahoo played a pivotal role in the early development of Hadoop, hiring Hadoop creator Doug Cutting in order to refine the software, which it used for large-scale data analysis.

About 20 Yahoo engineers, all with considerable Hadoop skills, were shifted to Hortonworks when the company was created (though Cutting himself is now at Cloudera). The company is hoping that its in-house expertise will make its services a viable choice alongside earlier entrants into the field of Hadoop support and software, including Cloudera and IBM.

Also, unlike their competitors, Hortonworks will keep its distribution fully open source, Baldeschwieler said. Other companies "have proprietary components where they replaced parts of the Hadoop stack with their own proprietary components," Baldeschwieler said. Hortonworks plans to fill in the missing pieces of the Hadoop stack, such as management tools, with fully open source components.

Hortonworks also plans to aggressively pursue partnerships with other software and system companies. Last month, the company teamed with Microsoft to bring Hadoop to Windows Server and Microsoft's Azure cloud service. "We're developing to build an ecosystem of partners," Baldeschwieler said.

Joab Jackson covers enterprise software and general technology breaking news for The IDG News Service. Follow Joab on Twitter at @Joab_Jackson. Joab's e-mail address is Joab_Jackson@idg.com