Since the early ‘80s, Oracle has been one of the major economic engines of the technology market. Oracle’s innovation and performance in the enterprise database market made it the envy of its competitors—including IBM, Microsoft, and Sybase. However, their efforts to expand their core business outside the enterprise database market may end up costing them their leadership position in that market as well.

In the last month, the CIOs of three midsize companies who built their entire technical infrastructure on Sun and Oracle each asked me to come talk to them about their future platform and development strategy. I expected the conversation to be about conversion to Web services or interoperability with Microsoft .NET. To my surprise, they actually wanted to talk about how to convert their existing Oracle/Java applications to Microsoft’s .NET platform. In this article, we’ll discuss their opinions about Oracle’s future and why they felt the need to begin seriously considering alternatives.

Looking back on Oracle’s success
All three CIOs expressed concern over the continuing decay of Oracle’s three key value propositions: cross-platform functionality, performance, and developer support. A look at Oracle’s performance and business decisions across these three areas over the last 10 years reveals why these strengths have eroded.

Cross-platform functionality
By the mid 1990s, Oracle had vanquished its only serious competition in the cross-platform race (Sybase), and neither Microsoft nor IBM appeared to be able to mount a challenge. But in the last five years, IBM has turned DB2 into a great cross-platform alternative. The collapse of the dot-com market has also raised DB2’s profile as an alternative to Oracle, since most corporations continue to use DB2 on multiple platforms, but Oracle’s market share gains in the dot-com market have all but disappeared.

Additionally, even though Microsoft has no intentions of making SQL Server available on alternative platforms, the industry move to XML interoperability between different systems via Web services instead of proprietary database interfaces makes the database platform decision much less of an issue.

In another delicious twist of irony, IBM—the proprietary mainframe vendor of the ‘80s—has become the open-systems consulting company of this decade. IBM has more Java programmers than Sun and more certified Microsoft engineers than Microsoft.

Performance and price/performance
Three years ago, Oracle owned a minimum of seven of the top 10 performance database benchmarks and at least one of the top 10 price/performance benchmarks, as ranked by the Transaction Performance Processing Council. Those days are long gone. The competition has become fierce between Oracle and DB2 on the high end, Oracle and SQL Server in the middle, and Oracle and Linux/DB2 or Linux/MySQL or Postgres on the low end. In the latest top 10 TPC-C raw performance benchmarks, Oracle could do no better than eighth. Even more damning, Oracle was completely shut out of the top 10 price/performance TPC-C benchmarks.

In addition, the continued declining CPU and memory costs allow competing database vendors to get enough raw performance for their applications using non-Oracle solutions without regard to Oracle’s marketing message of “the fastest database.”

Companies are also beginning to realize that the overall performance of a system has more to do with how the components work together than how the individual components perform. This gives Microsoft and IBM a significant advantage in that their complementary transaction monitors, queuing systems, and underlying operating systems are more tightly bound than those same components from Oracle or its partners.

Oracle had a strong relationship with developers who created applications using their database, their development tools, and their stored procedure language. But by spending the last few years developing and promoting their own financial systems (Oracle Financials) at the expense of one-time partners like SAP and PeopleSoft, they’ve created a scenario where high-end ERP vendors are looking for alternative database platforms. Where developers once learned Oracle skills to support these applications, now they’re focusing on alternative database platforms.

Oracle’s new dependence on Java has eaten away at the revenue stream they used to generate from their development tools. And their constant rewrites of their core application server technology have alienated many of the developers they were counting on to help them build a competitive integrated development platform. For example, two years ago they encouraged developers to port from Oracle Application Server version 4.x to Oracle Application Server 9i version 1.0. This required massive rewrites in both the database procedural code and the application logic. At Sun’s JavaOne conference in June, they announced that their new application server would have another new code base, again requiring application rewrites.

All the CIOs expressed concern over the need to continually rewrite applications that use Oracle technology, not only because of changes that Oracle makes, but because of dependencies on other companies’ technologies (like Java, the underlying operating system, as their choice of application server).

What does Oracle look like going forward?
The biggest concern expressed by these CIOs was that Oracle’s focus on becoming an application hosting company on the high end (with Oracle Financials) and an application hosting technology company at the low end (with major investments like those in NetLedger) would cause it to lose its focus on creating robust database technology. Also, by directly competing with those companies who used to be its major partners, Oracle’s database market share, and therefore, its ability to support the product, would begin to erode. Witness the major falling out between Siebel and Oracle and the inroads that IBM has made in the last year in closing the database market share gap with Oracle.

Another problem that Oracle (and Sun) must both face is the reality that the Java terminal revolution isn’t likely to happen any time soon, if at all. For this strategy to succeed, businesses must have access to inexpensive high-speed Internet connections, and the Java terminal has to be significantly less expensive than full-blown PCs. The rapid disappearance of the competitive local exchange carrier (CLEC) market has made it more difficult for ubiquitous high-speed access to become a reality. The combination of the continuing downward spiral of CPU and memory costs and the increased manageability of Windows 2000 and Windows XP has lowered not only the hardware cost, but also the TCO of full-blown PCs into the range of the dumb terminal. (To be fair, even Microsoft tried to convince its hardware partners to introduce Windows CE-based terminals—albeit for consumers—and this turned out to be a fiasco.)

Within the next 12 months, Oracle is likely to face a choice: become a hosted ERP provider and relegate their database technology to a key component of that strategy or de-emphasize the financials business and refocus on driving the database market with continued innovation. Many companies—like the ones I’ve spoken with recently—don’t intend to wait around until Oracle makes up its mind.

Smart move or foolish choice?

Should Oracle stick with its core business of building database software? Or is it a wise move to diversify? Share your opinion with us on the best direction for Oracle.