With its latest appliance, Oracle has officially embraced big data, including Hadoop and NoSQL.
As my colleague Derrick Harris reported last week, the Oracle Big Data Appliance, the latest engineered system in Oracle’s (s ORCL) lineup, harnesses an open-source distribution of Apache Hadoop, the Oracle NoSQL Database, and the R statistical analysis software.
The form factor is reminiscent of the existing Exadata appliance, said Thomas Kurian, Oracle EVP, speaking at Oracle OpenWorld 2011 today. Layered atop that is the new Oracle Loader for Hadoop, and a collection of tools including Oracle’s ETL (extract, transform, load) technology, he said.
“ETL can deploy on the Hadoop cluster and you can model that using Oracle Integrator ETL tool and then deploy that on Hadoop MapReduce platform,” said Kurian. “We provide load balancing and after preprocessing is done, [the loader moves] the data set into Oracle.” The finished data set then can be piped into Exalytics for analytic dashboards and reports,” Kurian said. Oracle CEO Larry Ellison announced the Exalytics analytics appliance Sunday night.
Oracle said this device will integrate tightly with the company’s existing Exadata database machine, Exalogic application server, and upcoming Exalytics analytics appliance.
By creating the Big Data and Exalytics appliances, Oracle now has all the tools needed to call itself a big data player that can compete with early leaders IBM (s IBM) and EMC (s EMC). It always had the analytic database with Exadata, but now it has integrated Hadoop for unstructured data; R Enterprise for running statistical analysis in the same vein as SAS or SPSS; and the real-time BI functions of Exalytics. Oracle’s NoSQL data is just icing on the cake, really.
Of course, the burning question is who will opt to fork out a premium for Oracle appliances and/or software when they can run any number of open-source Hadoop, R and NoSQL implementations on commodity hardware.
“I would look at this if I were looking at Greenplum and I had a lot of money. These [special appliances] really focus more on data warehousing applications; you need a very targeted reason to run them. I might try Oracle appliances but I would be more likely to use a Hadoop cluster or something on my own hardware. Oracle makes it sexy by prepackaging it on the hardware, but you need a lot of money,” said Pete Sclafani, CIO and co-founder of data center specialist 6Connect.
Availability and pricing of the Big Data Appliance were not disclosed. It’s worth noting, however, that the new software included — Oracle NoSQL Database, Oracle Data Integrator Application Adapter for Hadoop, Oracle Loader for Hadoop, and Oracle R Enterprise — will all be available as standalone software products, too.
Derrick Harris contributed to this story.