There has been speculation for a while now that Oracle (s orcl) might someday make its foray into the Hadoop and NoSQL spaces, and next week looks like that time.
With regard to Hadoop, CEO Larry Ellison made it clear during last week’s earnings call that the company is working on a connector that will let customers load unstructured data from Hadoop into their Oracle Exadata appliances. Now we have proof — and Oracle’s big data plans don’t stop with Hadoop.
Updated: I’ve seen some Oracle-produced content highlighting the company’s plans for a big data platform, apparently slated for launch in the second half of 2012, that not only includes the Hadoop connector — called Oracle Loader for Hadoop — but also a NoSQL database. The goal, it seems, is to let customers acquire data from whatever sources they please and then feed it into an Oracle Exadata data warehouse system. Once there, data can be analyzed via number of means, including existing Oracle technologies such as in-database MapReduce, mining and statistical analysis with R.
Reaffirming this information Thursday, Larry Dignan at ZDNet (s cbs) highlighted that the Oracle Loader for Hadoop is the topic of multiple sessions at next week’s Oracle OpenWorld conference as is “Oracle NoSQL Database.”
What remains to be seen, though, is how heavy Oracle — which has enabled Hadoop integration for some time, actually — will actually invest in Hadoop and NoSQL now that it appears interested in productizing them. Will Oracle sell a physical Hadoop appliance, as Ellison alluded to and as competitor EMC is doing (s emc), or is the connector as far as it goes? Will Oracle support any Hadoop or NoSQL distributions, or will it create its own like it did with Linux?
That database kingpin Oracle is getting into these markets is great validation for the technologies — and certainly will please Oracle customers wanting formal support for Oracle-Hadoop-NoSQL environments — but how it decides to do business within these spaces could be even more meaningful. Oracle buying up a company or two would be a very big deal to everyone else in these spaces, as it would both add and eliminate competition in one fell stroke. On the other hand, Oracle going too proprietary could limit its effectiveness in spaces dominated by open source technologies with plenty of hype and investment to thrive on their own.