Edmunds is an early Hadoop adopter, moving from Oracle RAC. Hbase allows Edmunds to combine structured and unstructured data into a single architecture. “The other advantage is we can load and query our data much faster, so our business analytics are much more productive,” Rokita told Wikibon.com Co-Founder and Chief Analyst David Vellante and SiliconAngle.com Founder and CEO John Furrier in the Cube from the floor of HadoopWorld 2011 on Nov. 9.

Edwards captures about 2 Tbytes of data a month, he said. Hadoop generates reports and query results much faster than Oracle could and includes much more data. In some cases it will allow Edmunds to run queries that simply stalled out and never finished on the Oracle system. “Once we got past a certain amount of data these queries just wouldn’t run,” he said.

And of course it then also allows new kinds of queries that combine the structured information with unstructured data based on clicks on the Edmunds’ site. This allows them to see, for instance, which car models and prices generate the most interest or what information customers access most.

Hbase also allows integration of other components of the Edmunds’ infrastructure such as its job coordination package more tightly into the architecture.

He is not concerned about the proprietary management console that Cloudera has released. “We don’t use any of that software in any case,” he says. “We have an in-house provisioning system that lets us do that.”

Actually, he says, Edmunds has yet to actually buy anything from Cloudera, as opposed to simply downloading its open-source Hbase software. “One problem with making your open-source products too easy to use is that then the users have less incentive to move to your proprietary products,” he says.

However, he says, “I am sure Cloudera will find ways to provide additional value to people that they will be willing to pay for. So they will find ways to generate income.”


Watch live video from SiliconANGLE.com on Justin.tv