Hortonworks, a company created in June 2011 by Yahoo! and Benchmark Capital, has announced the Technical Preview Program of Data Platform based on Hadoop. The company employs many of the core Hadoop contributors and intends to provide support and training.
Shortly after IBM announced the availability of a big data analysis platform based on Hadoop, a relatively new but significant player, Hortonworks, has advertised the Technology Preview Program of their Hortonworks Data Platform (HDP), an open source data analysis system also based on Apache Hadoop.
Besides Hadoop 0.20.205, HDP 1.0 contains a number of other open source projects meant to help with platform’s administration: Ambari, an open source installation and management system, HCatalog, a metadata management system, and the usual members of a Hadoop platform, Pig, Hive, HBase and Zookeeper. Over the next few weeks, Hortonworks plans to release version 2.0 of HDP based on Hadoop 0.23 which implements the Next Generation of MapReduce.
In what seems to be an already crowded marketplace, Hortonworks has an advantage over other competitors. Born out of a Yahoo! initiative, Hortonworks employs many Hadoop architects and source code contributors who previously had worked for Yahoo! in the past and have contributed over 80% of the total source code of the Apache Hadoop project, according to Hortonworks. They have also worked on developing some of the other projects included in the distribution (HCatalog, Ambari, Pig), and have participated in running Hadoop on over 40,000 servers at Yahoo!.
Hortonworks promises not to fork Hadoop but to contribute all future code back to the Apache project. They say it is in their interest that the Hadoop project is not split and grows over time. Their business model is to provide support and training to customers interested in running analytics on big data. As a side note, one of their present customers is Microsoft who announced the intent to become a Hadoop provider for Windows Azure.
Hortonworks Data Platform is currently available as a Technological Preview to a limited number of companies, and will enter Public Preview in early 2012 with the intent to have a public release some time later.
