Users might want containerisation to separate compute and storage for data analytics but few fully understand it, according to MapR, which has launched a set of integrations with Kubernetes.
The firm has squeezed out additions to its data platform to allow users to deploy native Apache Spark and Drill applications in Kubernetes, the open-source container orchestration software, separate from Hadoop.
This means that, as well as pushing customers to containers, today's launch edges MapR another step away from its former life solely as a Hadoop-flinger. Similar efforts are being taken by Cloudera.
And while Cloudera is focused on rebranding itself the "enterprise data cloud", MapR is attempting to sell the idea that it is the one offering AI and analytics. "MapR: End-to-end AI in one platform," shouts its website. "Cloudera: No platform for AI."
You can never have too much AI! MapR shoves more in data platform in bid to fill 'critical gaps'READ MORE
MapR made a commitment to Kubernetes last year, and added a Container Storage Interface plugin for it earlier in 2019.
However, MapR product bod Suzy Visvanathan said today that few customers fully understand the concept of Kubernetes and were "quite befuddled" by the complexity around the tech.
The firm wants to turn this to its advantage by offering easy-to-use integrations on its platform, so people can use the hyped tech at arm's length from the complexities of Kubernetes.
"One of the biggest focuses for us is to make it easy and simple, so down the road, an end user shouldn't even have to feel the need to understand all the components and complexities in Kubernetes; we hide it and then they should only understand the existing concepts they already know."
Visvanathan added that the aim is to separate compute and storage to allow workloads to be provisioned depending on the use case. She outlined three main use cases: data engineers building next-gen apps, architects and admins building data lakes, and for line-of-business architects developing apps in a multi-cloud environment.
At the moment, the platform will support Spark and Drill, allowing them to be deployed in containers orchestrated by Kubernetes. The longer-term plan is to make it so more components and analytics frameworks can be natively deployed on Kubernetes.
The latest release will be generally available in Q2, although Visvanathan said about a dozen companies have used the features in its proof-of-concept stages. ®