Oracle has become far too predictable. The company that used to print money every quarter now routinely misses analyst expectations as it struggles to come to grips with the new realities of enterprise IT: open source and cloud.

Of course, Oracle isn’t alone. As strategy consultant Peter Goldmacher declares, “All the legacy guys are going to die… [but Oracle is] going to die last.”

Are we in the midst of a changing of the enterprise guard?

Oracle misses earnings: Same ol’ same ol’

Oracle doesn’t hit earnings estimates. Not very often, anyway. Over the past few years, it has missed far more than it has hit its numbers.

This time, the company decided to change things up and announce big changes: Larry Ellison is resigning his role as CEO! While superficially true, the change is essentially meaningless: Ellison retains exactly the same control as before, as the two CEOs (Mark Hurd and Safra Catz) have to report to the board, which is chaired by Ellison.

And let’s be clear: Oracle needs to make serious changes. As The Wall Street Journal intones:

“Oracle can no longer count on its near-monopoly in database systems, as the emerging technologies of big data and cloud computing — often available in open-source versions that cost far less to use — along with a customer base eager for alternatives, fragment the market. At the same time, younger competitors such as and Workday are picking off Oracle customers by offering specialized software applications sold by subscription rather than in a large lump sum plus a service contract.”

This shows up pretty clearly in the percentage of Oracle’s revenue that comes from new license sales. Redmonk analyst Stephen O’Grady has assembled the data, and it’s not pretty (Figure A):

Figure A

Percent of Oracle software revenue from new licenses.

As he summarizes, “Oracle’s software revenue growth is increasingly coming not from new customers but from existing customers,” milking those that have already built on Oracle’s database products or other licensed products. Others, as The Wall Street Journal notes, are increasingly tempted by open-source and cloud-based alternatives.

This is not exclusively an Oracle problem, as O’Grady continues:

“[Charging for software is] certainly getting harder for Oracle. And if it’s getting harder for Oracle, which has a technically excellent flagship product, it’s very likely getting harder for all of the other enterprise vendors out there….This is not, in other words, an Oracle problem. It’s an industry problem.”

Unless, of course, you play by different rules.

Amazon: New king of the enterprise jungle

The company that is rewriting the rules of enterprise software — more than any other vendor — is Amazon Web Services (AWS). Compare Oracle’s new license revenue (above) to Amazon’s AWS revenue over the past few years, as deciphered by Andreessen Horowitz’s Benedict Evans (Figure B):

Figure B

Amazon “other” TTM revenue.

Oracle and its ilk are racing to embrace the cloud and, in some cases, open source. But it may be too little, too late. Oracle’s cloud revenue is growing at a 31% clip, but it still accounts for a mere 6% of total sales. Oracle must replace its slowing license revenue with faster-moving cloud revenue, but so far that hasn’t happened. Not fast enough, at any rate.

It’s not a pleasant place to be. I should know. When I was at Novell, our NetWare revenue — once a massively profitable cash cow — was declining at a torrid 11% pace. We acquired SUSE to add a growing product category, which we hoped would also bolster declining revenues from management tools and other software.

It didn’t work. Not enough, anyway.

But Oracle’s task is even bigger than Novell’s, because the competition thrives on entirely different rules. As Evans writes, “With Amazon, [CEO Jeff] Bezos is deferring [a] profit-producing, investor-rewarding day almost indefinitely into the future.” He’s focused on owning the future, even as Oracle and other technology incumbents keep trying to milk monetization of the past so as to keep tens of thousands of employees drawing sizable paychecks.

Can old dogs learn new tricks?

It’s simply not going to work — not when workloads are gravitating to new types of data infrastructure that are far better suited to the high-volume, high-variety data requirements of modern businesses. As Drew Johnson, Aeris’s vice president of engineering, told The Wall Street Journal:

“Fifteen years ago, there was no question: If you want to store data, you go to Oracle. I ask my [software] architects, ‘Build something to handle a billion things.’ When you look at that kind of scale, Oracle is no longer the best.”

Aeris turned to Cassandra, a database born within Facebook. Others look to Hadoop, conceived at Google but turned into a downloadable open-source project by Yahoo!. Few continue looking to the old guard.

AWS keeps rolling out services and no longer seems content to focus on geeky infrastructure. Amazon’s CTO Werner Vogels has declared that Amazon is in the “enterprise pain management” business, with no set boundaries on the kinds of pain it will seek to alleviate. Even Microsoft increasingly has embraced this new way of developing and delivering software, announcing an early version of a document database delivered as a service.

Such new businesses demand a completely different way of thinking about cost structures, delivery models, and business models. Can Oracle and its peers make the change? The company is due to announce its own database service shortly, but why would enterprises turn to Oracle for a cloud-delivered database when the king of clouds, AWS, already offers this?

The forecast seems cloudy with a chance of failure. Do you agree? Why or why not? Share your opinion in the discussion thread below.