Vertica Gets Hadoop Upgrade
An analytics platform unveiled this week by Hewlett-Packard allows access to data residing on a range of Apache Hadoop distributions.
HP said Nov. 17 its Vertica for SQL on Hadoop platform provides a pathway to data residing on Apache-based data lakes as well as Cloudera, Hortonworks, MapR or any combination the these Hadoop distributions. The new analytics platform reflects HP’s push to integrate the distributed computing technology for exploring data stored on the Hadoop Distributed File System (HDFS).
While large amounts of unstructured data are being stored in data lakes, HP said the broadening of it Vertica platform was prompted by early challenges in analyzing HDFS-based data. The intent was to combine HP analytics platform with Hadoop as its use expands.
HP said its Vertica platform would run natively on any preferred distribution of Hadoop while leveraging certification with visualization tools.
The expanded platform is intended to analyze live data using the Hadoop cluster while supporting complex data types and other features included with SQL on Hadoop.
HP claims its implementation of SQL on Hadoop differs from earlier offerings by allowing data analysts to continue using familiar business intelligence and analytics tools that visualize and auto-generate ANSI SQL code to interact with a variety of Hadoop distributions. Vertica is also said to tap into built-in analytic functions that support joins, complex data types and other capabilities.
The Palo Alto, Calif., company has been expanding its big data offerings this year through its Vertica Marketplace, including free software downloads from HP and its partners. Also in February, it debuted three free add-ons to its column-oriented data store for capabilities in areas like running large R workloads, sentiment analysis and geospatial tracking.
In May of this year, HP and MapR Technologies demonstrated the Vertica analytics platform on MapR’s distribution of SQL on Hadoop. The partners said the combination provided business intelligence and ETL (extract, transform, load) tool support intended to expand analysis of semi-structured data along with traditional structured data.
Also in May, HP announced the next release of the Vertica platform called HP Vertica Dragline. Dragline incorporates technology from HP’s “Project Maverick” designed to improve performance during more frequent and concurrent queries. A dynamic workload manager spots and adapts to the varying complexities of queries, the company said.
The new release also expands SQL-on-Hadoop analysis while eliminating the need to move data. HP also said it supports more formats for data analysis, including Avro, CEF, Parquet and Thrift.
Vertica for SQL on Hadoop is priced on a per-node basis and is available immediately. The company also hinted at an upcoming announcement during a sponsored event in early December in Barcelona, Spain, involving “leading Hadoop distribution partners.”