Hortonworks Data Platform update flicks on containerisation
Extends cloudy deals with IBM, Microsoft and Google
Data management firm Hortonworks has enabled containerisation in the latest release of its Data Platform, while announcing a set of extended cloud deals with Microsoft, Google and IBM.
CTO Scott Gnau said the overall aim of the announcements, made today at its 11th annual DataWorks Summit in San Jose, was to offer customers a "seamless experience on premise and in the cloud", supporting hybrid and agile deployments.
The firm has been pushing the idea of giving customers a single view of all their data for some time now. The theme of the European DataWorks Summit back in April was that every company's business strategy should be its data strategy, which should also be its cloud strategy.
"In an Internet of Things-centred world, there's going to be this notion of hybrid footprints where data may live in multiple cloud footprints, it may live on premise," Gnau told The Register ahead of today's announcement.
He said the firm was focused on customers "being able to connect all of that information together and take on the whole silo effect". The idea is for Hortonworks' core components to run in any of those instances, to provide standardised footprints and allow application mobility and portability across them.
Gnau said that the Hortonworks Data Platform 3.0 release – based on Apache Hadoop 3.1, which was just released in April – was a "huge step forward, architecturally".
The highlight, he said, was containerisation, giving customers the ability to use containers for agile deployment of new analytics and launch apps quickly to save time and resources. It will also support deep learning applications that depend on substantial and expensive GPU resources.
Gnau added that the new release represented a "coming of age" for real-time database capabilities as part of the data warehouse tool Apache Hive, which offers an SQL-like interface for querying data stored in Hadoop.
"In our previous releases we talk a lot about enhancing the experience for Hive, whether that's overall performance, ACID compliance, the notion of indexing," he said. "And really, it's related to HDP 3.0, moving from a collection of SQL on Hadoop to a really integrated database that contains all the things you’d expect."
The Hadoop-flinger has also announced three extended cloud deals, including greater integration with Google Cloud Storage and a new IBM Hosted Analytics with Hortonworks offering, an integrated solution – comprising HDP, IBM DB2 Big SQL and the IBM Data Science Experience – available on the IBM Cloud.
The third is an extension of a six-year partnership with Microsoft that will allow customers to deploy Hortonworks Data Platform, DataFlow and DataPlane Services products natively on Microsoft's Azure IaaS.
"The notion behind this expanded relationship... is to make sure we give customers a common experience, regardless of whether it is on premise or in the cloud. This enhancement obviously creates some more options and opportunities for customers leveraging Microsoft cloud products."
HDP 3.0 will be generally available in Q3 of 2018. ®