OpenWorld is less than two weeks away and speculation is increasing that Oracle is going to make a Big Data-related announcement at the show.

Big Data presents a bit of a quandary for Oracle. Oracle is touting its Exadata appliance as the answer to its customers’ Big Data challenges. While Exadata does bring a number of benefits to Oracle customers (namely a highly optimized analytics appliance that requires little if any hardware tuning), the fact remains that Exadata is not a cost-effective or time-effective Big Data solution.

In short, Exadata is a traditional – if highly optimized – data warehouse appliance little better at processing and analyzing large volumes of distributed, loosely structured data than any other traditional database. (See Wikibon CTO David Floyer’s excellent side-by-side comparison of Exadata v. an MPP-style database for crunching Big Data workloads.)

All this means that Oracle needs to make a substantive move to stay competitive in the nascent but fast growing Big Data market. As I see it, Oracle has a handful of options. I describe them below, in order from least to most likely.

1. Develop and release its own Hadoop distribution. This is the least likely scenario, from my perspective. As far as I know, Oracle doesn’t have any engineers or developers working on or contributing back to the open source Apache Hadoop project. It would simply take too long not just for Oracle to develop a competitive distribution from a technical perspective, but too long for Oracle to build up the open source credibility it would need to spur adoption. Not to mention, Oracle isn’t too keen on sharing its intellectual property with competitors.
2. Acquire a Hadoop distribution vendor. Slightly more plausible, but still unlikely. There simply aren’t any mature Hadoop distribution vendors on the market at the moment, so if Oracle were to acquire Cloudera or MapR, it (Oracle) would have to let the newly acquired company continue developing its distribution and contributing back to the community. As mentioned, Oracle isn’t big on sharing. Plus, I don’t really see any reason for Cloudera or MapR to sell at this point. The Big Data market is only going to get hotter and both companies should be able to fetch a much higher price in a few years time.
3. Acquire an MPP-style database vendor. In the last year, the three leading MPP database vendors were bought up by Oracle rivals, leaving a thin field for Oracle to choose from. But not too thin that it won’t make a move. The most likely acquisition target left on the market is ParAccel. The company’s flagship analytics database is natively columnar, boasts significant data compression capabilities and, according to the company, can easily integrate data to and from HDFS.
4. Partner with a Hadoop distribution vendor. This is the most likely scenario leading up to OpenWorld. I expect Oracle to announce a partnership with Cloudera, MapR or perhaps another Hadoop distribution vendor in which the Hadoop distro is tightly integrated with Exadata, Oracle 11g and other Oracle products so Oracle customers can easily transfer data from Hadoop to Oracle and back again. It would provide important functionality for Oracle customers but not require Oracle to get involved in the Apache community itself. This option also allows Oracle to save a bit of face and continue touting Exadata for Big Data scenarios, at least as Oracle defines Big Data.

ServicesANGLE

Whichever option Oracle chooses, it will have an impact on its services business. Specifically, as my Wikibon and SiliconANGLE colleagues and I have pointed out before, working with Hadoop and other Big Data technologies and approaches requires significant expertise in a multitude of domains. Unfortunately, there just aren’t enough highly trained Big Data developers, admins and Data Scientists to go around. For Oracle, this means the company will have to beef up its services offerings, whether internally or through third-party providers, to include Big Data consulting and development. For Oracle customers, be sure to question the company aggressively about how it plans to support you as Big Data workloads become more common and the tools and technologies to address them become more diffuse.