HP Announces New Big Data Analytics Platform
It features speedier queries, improved query-dependent allocation of resources and better SQL-on-Hadoop functionality.
Dragline, the name of Hewlett-Packard Development Company L.P.'s new Vertica Analytics Platform for Big Data, features speedier queries, improved query-dependent allocation of resources and better SQL-on-Hadoop functionality, the company announced yesterday.
The new HP Vertica Dragline release is one component of the company's HAVEn Big Data analytics platform, HP said.
"Our goal continues to be maximizing the value of every organization's most strategic asset -- its data," said company exec Colin Mahony in a statement. "HP Vertica Dragline lets organizations store data in a cost-optimized manner, exploring it quickly and effectively using well-known SQL-based tools, and, most importantly, providing the customer insights and operational advantages to everyone with blazing-fast speed, accuracy and security."
HP's Jeff Healey yesterday wrote a blog post in which he outlined the main improvements of Dragline.
First, Healey said, was technology from what the company calls "Project Maverick." It has several elements, Healey said, including Live Aggregate Projections, which speeds up analytic queries about individual, discrete individuals or devices by up to 10x.
These "live lookups" analyze data as it is received, allowing individual targeted and personalized analytics without the need for programming accelerator layers. Healey said this functionality would be ideal for utilities seeking to provide smart metering reports for individual customers, comparing their energy usage to neighbors, for example.
Project Maverick also allows the dynamic allocation of resources depending on the complexity of queries, functionality already extant in numerous data warehousing solutions. This Dynamic Mixed Workload Management feature can customize resource allocation depending on whether queries are simple one-offs or more advanced and long-running, Healey said.
Dragline also helps customers manage multiple storage tiers. Bulk data is often stored in the Hadoop Distributed File System (HDFS) in Big Data analytic setups, Healey noted, and Dragline allows SQL queries on that data without having to move it elsewhere or use connectors to access it. Data can be moved into Vertica for faster in-depth queries.
Healey said that while Vertica supports all major Hadoop distributions, running Vertica and the MapR distribution on the same cluster "is something special."
MapR also noted the new Vertica announcement on its site. "Together, HP Vertica and MapR deliver a comprehensive, scalable, open standards-based solution for big data deployments," MapR said. "Running the HP Vertica Analytics Platform directly on the MapR Big Data Platform provides an optimized, best-of-breed solution for SQL-on-Hadoop to unlock insights from all your data, with a smaller data center footprint than other RDBMS and Hadoop hybrid architectures."
David Ramel is editor in chief of Visual Studio Magazine and Application Development Trends Magazine.