Hortonworks, Pivotal team to better manage Hadoop
In the hot market for big-data products and services, sometimes even competitors must work together for the common good.
Prominent Hadoop software and service companies Hortonworks and Pivotal have partnered to further develop software called Ambari, which would make it easier for enterprises to manage Hadoop distributions.
“At the end of the day, customers want to deploy Hadoop cleanly into their environment,” said Shaun Connolly, Hortonworks vice president of corporate strategy. “Getting vendors like Hortonworks and Pivotal to work together is important for the market to continue to mature.”
The market for Hadoop software and services is expected to reach US$13.9 billion by 2017, according to industry research firm Markets and Markets.
Ambari can be used to provision, manage and monitor Apache Hadoop clusters. It provides a dashboard to monitor the operational health of a Hadoop cluster. It can also be used to stop, start and reconfigure a Hadoop installation.
Thus far, most software released to manage Hadoop clusters has been proprietary, Connolly said. StackIQ, for instance, offers management software. Apache also oversees Mesos, another open-source application for managing Hadoop clusters.
The two companies will combine engineering efforts on Ambari so they will not duplicate their work, Connolly said. They will strengthen the program interfaces so they can be used by a wider variety of system management products, such as Microsoft System Center, or open-source configuration management software such as Puppet or Chef.
The companies will modify Ambari so that it can better manage very large clusters. Today, Ambari can handle up to thousands of nodes, but this number could be increased, Connolly said. The two companies will also look at ways that Ambari could be modified to more easily run Hadoop as a hosted service.
Hortonworks and Pivotal compete for the same jobs at times, but “from the perspective of furthering of enterprise Hadoop, it opens the door for direct collaboration,” he said. Open-source Hadoop management software can be used by all Hadoop vendors to provide a common platform for customers.
Founded by a number of engineers who worked on the original Yahoo implementation of Hadoop, Hortonworks offers a fully open-source Hadoop distribution, called the Hortonworks Data Platform. Last week, HP invested $50 million into Hortonworks and will roll the Hortonworks Data Platform into its data-analysis systems.
Pivotal was formed by VMware and VMware’s parent company EMC, and offers a set of data-analysis services and software, including a Hadoop distribution called Pivotal HD.
Other companies that contribute to Ambari include Red Hat, IBM, Microsoft and Teradata.