MapR has unveiled a converged cluster plan to do away with emerging silos in big data.
The NoSQL startup today announced Streams as the newest part of what the firm has branded the MapR Converged Data Platform. MapR joins fellow Hadoop spinners Hortonworks and Cloudera in the handy platform branding stakes, with Hortonworks Data Platform (HDP) and Cloudera Distributed Hadoop (CDH).
Streams, which will become generally available in early 2016, is an event-streaming architecture pitched at those gathering and analyzing data in real-time.
It’ll form part of CDP, which combines MapR’s file, database and analytics.
Overall, MapR is pitching CDP as the way to run an integrated big-data stack that sidesteps the need to run separate nodes for Kafka, Tibco, Spark, Storm or Hadoop.
“Spark and Storm are excellent but there’s more analytics and processing than just the stream analytics,” director of product marketing Will Ochandarena told The Reg.
CDP supports the Kafka API.
“This will make Spark and Storm more... flexible,” Ochandarena said.
MapR’s goal is to create single system of record for steams of data and one place for analysis.
It cuts down the time and cost normally associated with setting up separate nodes for aspects of data upload, streaming and analysis.
“We are starting to see data silos emerge as different processing requirements with different clusters for data in motion versus data at rest and separate cluster for special analytics,” Ochandarena said.
“Customers are looking at speeding the process between when data is collected and when an action is taken.” ®