Big Data

Oracle CEO Hurd thinks the database market is about features. He's dead wrong.

Oracle's CEO thinks he's immune from NoSQL and cloud challengers because he's got more features. That's a poor strategy for the future.

Oracle has got features. So many features. It has so many features, rival databases would be crazy to try to compete. Because, after all, features always win, right?

That seems to be the view of Oracle co-CEO Mark Hurd, who believes that Oracle and Microsoft have "got too many features" for an upstart competitor like MongoDB to match. He's right that MongoDB can't (or won't) reach feature parity with Oracle anytime soon. He's also correct that MongoDB's $100 million in revenue is "a long way to becoming meaningful" relative to Oracle's billions.

But Hurd is absolutely wrong in his belief that "features" provide an indomitable barrier to would-be database competitors like MongoDB or AWS. Dead wrong, and the hubris could cost him.

What could possibly go wrong?

Let's not get ahead of ourselves. For all the bluster of a MongoDB or AWS, neither is yet making a dent in Oracle's database dominance. A few hundred million in revenue here or there is the sort of pocket change Oracle founder Larry Ellison leaves between the sofa cushions on his yacht.

Hurd's primary complaint about supposed rivals like MongoDB and AWS isn't so much about financial performance, however, but ultimately about features. This is where he demonstrates a profound misunderstanding of market dynamics (and a lack of familiarity with The Innovator's Dilemma). In Clayton Christensen's view, it's precisely those vendors that serve their best, most demanding customers with feature-packed, sustaining innovations that lose sight of emerging market needs and over-serve that market with expensive, out-of-touch products.

Which is a pretty good description of Oracle right about now.

SEE: Oracle, NetSuite and the future of SaaS ERP (Tech Pro Research)

The market no longer needs or wants a gargantuan, scale-up approach to data management. Instead, as one JAXenter survey indicated, a world awash in big data is interested in databases that offer the flexible schema and serious scale-out performance of a MongoDB or Amazon Aurora:

jax.png
JAXenter

See Oracle on the list? It barely squeaks by Microsoft SQL Server. Both manage to dominate the "Not interesting at all" space.

Building the future

Not that Oracle is going to be snuffed out anytime soon. Far too much enterprise data resides within Oracle databases for that to happen next year, or even next decade. It's going to be a slow rot, one that Oracle's almost complete non-existence in public cloud isn't helping.

That rot will be fed by developers.

Databases are incredibly sticky, yes, but developers are the new kingmakers and they're not handing out crowns to Oracle. As AWS chief Andy Jassy recently stated, AWS has "hit 40,000 database migrations for AWS Database Migration Service." That's significant, but even more so, he continued, the "Pace [is] quickening." Partly this is because, as Jassy said in a separate interview, "Customers are sick of [Oracle]." It's not hard to find enterprises to corroborate this claim. Companies buy Oracle because they have to keep feeding the beast upon which they built the last two decades of data infrastructure.

SEE: How to build a successful data scientist career (free PDF) (TechRepublic)

Developers, however, have plans for their future data infrastructure, and those basically never include Oracle.

This shows up in new software license revenue, which Redmonk analyst Stephen O'Grady picked apart back in 2013. (Since that time, the picture for Oracle's license revenue has only worsened.)

orcl-new-licneses-wm-e1385069046327.png
Stephen O'Grady (Redmonk)

The punchline? "Oracle's ability to generate new licenses is in decline, and has been for over a decade....Oracle's software revenue growth is increasingly coming not from new customers but from existing customers," O'Grady wrote.

Oracle's revenue has soared over this same period of time, not because of net new demand for its software, but simply because it has become expert at squeezing existing customers for more cash, selling them new applications that it acquires.

This strategy has worked well for giving hidebound CIOs a bit more mileage out of their existing Oracle database investments. However, it does absolutely nothing to address the needs of modern applications, which demand new data infrastructure to handle the high velocity, high volume, and high diversity of data. For such "big data," AWS has a great story, as do MongoDB, Apache Cassandra, and other new school databases. Oracle doesn't, and it's much-vaunted "features" actually prove to be a harmful shackle to Oracle's legacy rather than a springboard into its future.

Hurd must get this. He's a smart guy, albeit one who made his reputation at HP for cutting his way to profitability, rather than growing it through innovation. He is, in short, a great shepherd for Oracle's demise, rather than a prophet for its rebirth. I wouldn't call that a feature.

Also see

oraclesign.jpg
Image: iStockphoto/maybefalse

About Matt Asay

Matt Asay is a veteran technology columnist who has written for CNET, ReadWrite, and other tech media. Asay has also held a variety of executive roles with leading mobile and big data software companies.

Editor's Picks

Free Newsletters, In your Inbox