Oracle’s is in Big Trouble : Big Data is to Blame

In Wikibon’s March newsletter, analyst Jeff Kelly takes a look at the factors that have transformed the database market  into a fiercely competitive ecosystem where small startups are standing up to legacy vendors – and winning.

It wasn’t always like this.

Up until recently, Larry Ellison’s Oracle dominated the database market with an appliance-led business model that was (and still is) meant to lock the customer in. This status quo reigned for decades up until mobile became a trend, and enterprises started waking up to the fact that relational databases cannot address today’s new set of business requirements.

Kelly describes this phenomenon, dubbed by some as the Big Data explosion, as a “trend towards scale-out platforms and non-relational databases that use commodity hardware and open source software supported by some level of proprietary components.”  The heart of this movement is Hadoop, the open-source data crunching platform that is being distributed commercially by half a dozen startups.

Hadoop is a driving force behind the Big Data trend, but it’s only one contributing factor in the disruption of the traditional database market. NoSQL solutions such as MongoDB, Cassandra and Accumulo are also being rapidly adopted by the enterprise, and across all segments.

“The economics don’t look great for Oracle. According to analysis by Wikibon’s David Floyer (and highlighted in the Wall Street Journal), the NoSQL database market is expected to grow at a compound annual growth rate of nearly 60% between 2011 and 2017. The SQL slice of the Big Data market, in contrast, will grow at just a 26% CAGR during that same time period.

Changes are certainly afoot in the database market. If you were to ask someone today to describe the database market, you’d hear response like “exciting,” “innovative,” and “NoSQL.” A lot can change in just 10 years.

Maria Deutscher

Maria Deutscher is a staff writer for SiliconANGLE covering all things enterprise and fresh. Her work takes her from the bowels of the corporate network up to the great free ranges of the open-source ecosystem and back on a daily basis, with the occasional pit stop in the world of end-users. She is especially passionate about cloud computing and data analytics, although she also has a soft spot for stories that diverge from the beaten track to provide a more unique perspective on the complexities of the industry.

15 Comments

  1. imbigdata I would agree with merv that this article is blowing hot air

  2. Relational databases aren’t going anywhere soon, And i think Oracle has the resources to adapt. Sure, there a major trend (not shift) towards Hadoop for Big Data challenges (Current employer, Home Depot SSC is big on it). But I remember a time when Cassandra was supposed to be the next big thing. There’ll always be room for appliance-led model, after IBM’s mainframe model is still in using, a lot if companies are not deviating from VB or Unix, and those have been around forever. That said, the new trends are helpful across the board, for companies looking to alternative “cheaper” solutions and professionals who rapidly gain the new skills brought about by this mobile iOS age of scaled-out/hybrid platforms. The real winners in all this are the memory/storage companies. #emc #ibm

  3. Oracle has the industry leading Big Data Appliance to run NOSQL & Map Reduce jobs 5x faster than commodity in triplicate.  Do your homework before posting flagrant nonsense!

  4. @Millyy What does “industry leading” mean? Please provide a little data in your refutation, please.

  5. No doubt relational DBs aren’t going anywhere, and neither is Oracle. Big Data is not an existential threat to Oracle, but my point in the research note was that Oracle is not well-positioned to capitalize on the move towards open, scale-out platforms. Start-ups and increasingly established enterprises are recognizing that you can handle new Big Data workloads and (more and more) “traditional” data workloads alot cheaper with NoSQL and Hadoop than you can with a mighty expensive, scale-up Oracle Exadata box.

  6. With the size of NoSQL market as it is now and the size of RDBMS market (which is the right term rather than SQL market – these days everyone not too lazy does SQL layers on top of NoSQ) is such that even with 60% growth over the next 5 years, it will be far far behind of RDBMS market with “modest” 25% growth.

  7. @alexgorbachev Very true Alex. And at some point, “Big Data” databases (aka NoSQL) and relational DBs will hardly be a distinction worth mentioning as the two will converge. But for the next five to ten years, the growth opportunity for DB vendors is in the NoSQL space, IMO.

  8. Even with their existing MapR & NoSql solutions, do you really think Larry is that worried about a few startups? He’ll just open up his personal checkbook, buy like 5, combine the best in breed and go to market with a far superior solution.

  9. I just love how you can throw in something like “the fact that relational databases cannot address today’s new set of business requirements” without caring to back that “fact” with anything.
    As far as I can see, the NoSQL dbs allow people to store a lot of data, usually in the cloud. Most enterprises care about their data and would not outsource its storage to an outside cloud. So their choice will be to either build large private clouds, or use large traditional RDBMSs. With enterprise RDBMS vendors (I know for sure about Microsoft and Oracle, the others are probably following the trend) offering their own solutions, enterprises will have little incentive to switch to NoSQL.

  10. that was to read “their own *cloud* solutions”

  11. In the words of Gregor Hohpe “your local coffee shop doesn’t use two phase commits”

  12. @chrisgeorge4 Well, I think Oracle is not dismissing NoSQL and Big Data trends but paying all attention and does what needs to be done to stay relevant and lead the database market. Whether they will do right or wrong thing is another discussion but I don’t think Oracle is dismissive.

  13. @dvellante , neither does a local coffee shop use distributed cloud technology. I agree with using the right tool for the right job, what I don’t agree with is that enterprises are suddenly going to start replacing their RDBMSs with Hadoop.

  14. @Sergey Kazachenko my point was not on of TAM rather one of style of developing apps. Starbucks initiates a transaction using a FIFO queue but then proceeds to process the request in a manner that is optimized for fast delivery of results in total. They’ll make a latte at different speeds than teas or smoothies and deliver whichever result is completed first. How do they maintain integrity? By writing “Dave” on my cup!

  15. RDBMS’ aren’t going anywhere. They have a valuable purpose for their use in commercial apps. NoSQL dbms’ are a good complement to the use of RDBMS’ in apps where RDBMS’ are not optimized, e.g. social networks, unstructured data, many to many relationships, etc.
    I agree with the comment that Oracle will buy NoSQL technologies to improve their market presence in this space. For example, they purchased BerkeleyDB which Oracle has repackaged as their “NoSQL DB” (it’s a key value store).
    Regarding Millyy’s comment referencing Oracle’s Exadata system, yes it’s superfast, but a lot of up front work (such as ETL) has to be done to make it fast and it’s super expensive. As a result, Exadata is having a challenge in the market acceptance department.

Submit a Comment

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>