Hortonworks, Mirantis and Red Hat Partner on Project Savanna

Two of the biggest players in the OpenStack community and a top Hadoop provider announced plans yesterday to join forces to advance the "Hadoop on OpenStack" project known as Savanna. OpenStack systems integrator Mirantis Inc., the company that started Project Savanna, will be working with Hortonworks Inc., the top commercial distributor of Apache Hadoop, and Red Hat Inc., the current leading OpenStack contributor, the three companies said today.

"We're rallying around this notion of Apache Hadoop as the killer application for OpenStack," Shaun Connolly, Hortonworks VP of corporate strategy, told ADTmag.com. "Hadoop is clearly an important technology in the big data space, and it stands to benefit from being deployed on a cloud platform."

Also known as "Elastic Hadoop on OpenStack," Project Savanna aims to provide a means to easily provision and manage Hadoop clusters on the OpenStack cloud infrastructure. Hadoop, which started as an implementation of the MapReduce paradigm, has evolved into a platform for distributed computing with a growing number of projects built on top of it. The Savanna project specifies Hadoop parameters, such as version, cluster topology and nodes hardware details.

"You can view Savanna as that elastic cloud controller that will make it easy to deploy and spin up Hadoop clusters on demand within an OpenStack cloud," Connolly said.

The collaboration will provide an integration point for third-party Hadoop provisioning and management frameworks, the companies said. One example they point to is the open source Apache Ambari project, which provides an intuitive Hadoop management Web UI backed by its RESTful APIs. The companies say they will have a demonstration of this technology ready for the upcoming Hadoop Summit, scheduled for June in San Jose, Calif.

"We are bringing engineers to bear on this project," Connolly emphasized. "This is not a marketing relationship. It's about real engineers writing real code within a community-driven foundation project."

Mountain View, Calif.-based Mirantis, which specializes in building OpenStack-based, open source cloud platforms, originated Project Savanna, and just recently contributed the source code to the OpenStack community. The project grew out of the "megatrend" of big data initiatives, said company President and CEO Adrian Ionel, and a recognition of a growing demand among its customers for an integrated solution.

"They wanted a unified infrastructure that gives them a more flexible approach to providing resources to applications when they need them, instead of managing separate clusters throughout the enterprise," Ionel said. "We saw a great opportunity to take Apache Hadoop and OpenStack and blend them together to make Hadoop an elastic component on top of OpenStack, which would provide a drastically improved experience in terms of being able to spin up a Hadoop cluster within OpenStack on demand, and then manage it from there and integrate it with other workloads in the enterprise."

Currently in its seventh release (code-named "Grizzly"), OpenStack is made up of several interrelated projects focused on delivering various components for a cloud infrastructure solution. As the community Web site describes it, the project aims to deliver "solutions for all types of clouds by being simple to implement, massively scalable, and feature-rich." More than 180 companies participate in the OpenStack project, including Advanced Micro Devices Inc., Cisco Systems Inc., Citrix Systems Inc., Dell Inc., Hewlett-Packard Co., Intel Corp. and Microsoft.

The three companies made the announcement at the OpenStack Design Summit, which is underway this week in Portland, Ore.

Posted by John K. Waters on April 17, 2013