It comes as no surprise that Larry Ellison is stepping down as CEO of Oracle and becoming chairman of the board. As his tennis exploits and much-publicized preference for America's Cup over his own Oracle OpenWorld event suggest, Ellison's interests have run in other directions for some time now.
Yet note that the board has also formally "appointed" Ellison as CTO. According to Oracle, all engineering functions already reported to him, so probably the change is in name only (though I have no idea how hands-on he has been or will be). But hey, just imagine, what if he actually intended to take that CTO position and run with it? He's an awfully youthful 70-year-old and Oracle technology development needs shaking up.
Technology-wise, Oracle has painted itself into a profitable, but ultimately shrinking, corner. True, you won't see governments or giant corporations ripping and replacing their huge Oracle Database installations anytime soon, nor will they suddenly replace PeopleSoft with SaaS financials-by-subscription. There's tons of money in upgrades and maintenance fees to be squeezed from those big customers -- just as, for example, IBM's mainframe business keeps chugging along. But reliance on past conquests has made Oracle's future increasingly hazy.
Ellison must be frustrated to see Oracle lose its technology leadership position, though in part he has himself to blame. Echoes of his famous rant ridiculing cloud computing haunted the introduction of Oracle's cloud less than three years later. Plus, although Oracle dipped its toe into NoSQL with a key-value pair offering three years ago, that product seems mainly intended to be a companion to the Oracle database. Finally, Oracle's venture into in-memory databases last year seemed especially wan, as if trying to ride the coattails of SAP's marketing success with its in-memory HANA analytics solution.
Meanwhile, the once sleepy database realm has exploded. The cluster of new technologies around big data, including Hadoop and a wild array of new NoSQL databases, represents the biggest leap forward in data management and analytics since the 1980s. These emerging technologies are already delivering business value: In deep insights about customer behavior, in faster app dev cycles, in the ability to use commodity hardware, and in reduced software licensing costs, because almost all these new technologies are open source.
Database decisions have moved from friendly golf course chats between the CIO and the Oracle rep to, much further down the hierarchy, developers spinning up open source NoSQL solutions on the fly. Developers are picking the back ends they need -- mainly for Web and mobile applications -- because they scale out easily, offer flexible data modeling, and cost far, far less than an Oracle database.
But this sudden burst of variety has also been difficult for operations to handle, and many senior manager types remain skeptical of what they consider to be "immature" NoSQL solutions. When the RDBMS reigned supreme, choices were relatively simple. Today, CIOs are losing track of exactly what developers are running this week, and they're lamenting the loss of control.
This is Ellison's big opportunity. What if he worked toward a grand, unified database offering that incorporated all the special performance and scale-out features of various Hadoop and NoSQL databases alongside Oracle and wrapped it all in the perception of Oracle as the ironclad database solution? I'm willing to bet many large companies, at least, would turn to Oracle for reassurance and simplification. Yes, even commercial open source is inexpensive -- particularly compared with today's Oracle licensing -- but finding the skills to develop and maintain a host of different NoSQL solutions costs real money.
So far, Oracle has mainly made gestures, such as the recent announcement of JSON support for the Oracle database. Some fundamental architectural matters need to be rethought for Oracle to take the lead once again. Ellison has always had the reputation of being technically savvy. What if maybe, just maybe, he really planned to lead Oracle technology development into a new era?
Assuming Ellison were able pull that off, however, Oracle's pricing model would also have to change drastically. It simply can't stretch to fit big data and NoSQL-style concurrency. Also, one huge benefit of scale-out NoSQL architecture is that it can be done on commodity hardware -- a business Oracle Sun seems unlikely to pursue.
I know it's a pipe dream, but the monumental task of moving Oracle technology forward -- and changing the way the company aggressively markets and sells software -- would be a terrific final act for Ellison, the guy who corned the database market. But come to think of it, the second of those two tasks would probably be a lot harder than the first.