Not only is there more data than ever that needs to be managed, but much of that data is more distributed than ever. To make it simpler to manage distributed computing environments based on Hadoop, Hortonworks today unveiled a centralized management platform through which data and application workloads deployed on multiple instances of Hadoop can be centrally provisioned and managed.
Arun Murthy, chief product officer for Hortonworks, says the Hortonworks DataPlane Service (DPS) can be deployed on a public cloud or in on-premises environments to create a distributed data fabric across an open source computing environment. Built into Hortonworks DPS are tools based on the open source Apache Atlas metadata framework that IT organizations can employ to apply data governance policies, says Murthy.
Also included in Hortonworks DPS are tools to register and create data sources to allow consolidated access as well as a data services catalog of utilities to configure and manage new services. In addition, security access controls are included.
Murthy says the goal is to make it simpler for organizations to manage connected data at scale on a Hadoop cluster rather than individual components.
“Organizations don’t want to be required to manage snowflakes or individual crystals,” says Murthy.
At the same time, Murthy says Hortonworks DPS is designed as an open environment. The first third-party vendor to pledge to plug a data governance framework into Hortonworks DPS is IBM. The IBM Unified Governance Platform will be integrated with Hortonworks DPS sometime next year.
Big Data applications are now being routinely deployed across the enterprise. The challenge IT organizations face is figuring out how to manage them across an IT environment that in the age of the cloud becomes more distributed with each passing day.