With the recent announced retirement of IBMs Janet Perna, most associated with their DB2 database franchise, it got me thinking about IBMs future in the relational database market.
I, like many others, got my first taste of relational technology back in the mid-1980s supporting DB2 on MVS.
I remember attending many of the early International DB2 User Group (IDUG) conferences, which were both rowdy and informative.
The momentum and energy around DB2 was growing and positive, clearly the front runner.
Not many at that time considered Unix a threat, a toy really, and even if it did become a credible enterprise platform, no doubt DB2 would dominate that market as well.
By the early 1990s, I had left the end-user side for the vendor side, but working in field support only connected me more with what different organizations were doing.
By 1993, I began to see the shift toward acceptance of distributed database platforms.
By 1995, that trend was unstoppable. By then, I hadnt touched a mainframe for over two years, having switched attention to Oracle, Sybase and Informix.
But where, I asked, was DB2? What had they done to lose their leadership position in such a growing market segment?
Some may remember IBMs push around OS/2 and subsequently its releases of DB2/2 to complement it.
They also pitched connectivity between DB2/2, DB2, SQL/DS and OS/400 databases. Sound familiar?
Of course OS/2 and VSE failed to win market share and the improved price and performance of AS/400 was eclipsed by rapid improvements in RISC-based and now with Intel-based chip architectures.
The mainframe remains a solid cash cow, but certainly not a growth engine. So perhaps while other competitors foresaw the implications of Moores Law on the database market and focused on their database software, IBM took its eye off the ball.
By 1995, Oracle had solidified its position as the market leader in the distributed database market.
Sybase and Informix had unknowingly reached their zeniths, and Microsoft was just beginning its assault.
By 2000, IBM seemed to be a player in the distributed relational database market only by default as Informix and Sybase had self-destructed.
By 2004, IBM was no longer viewed as a database company but as a company with many databases.
The economic downturn in 2000 forced organizations to focus on software costs once again.
Oracle was having difficulties with its licensing model (remember Power Units), and IBM had released DB2 V7 for Unix and Windows, which many thought to be IBMs first truly competitive version for the distributed market.
With users upset with Oracle, many (myself included) thought the time was right for DB2 to finally take market share away from Oracle.
Based on my interaction with clients over the past five years and by looking at some of the market share numbers a bit closer, my conclusion is that IBM never succeeded in making much headway with DB2 relative to its competitors on Unix, Windows or Linux platforms, and that perhaps their window of opportunity had indeed passed them by.
Problems with DB2
Here are some reasons why.
A key problem with DB2 was that they chose to follow an options pricing model similar to Oracles versus a bundled model that Microsoft embraced.
This is important because IBM essentially loses its value proposition versus Oracle when you start adding on optional features such as cube views, OLAP and data mining.
The separate options approach is required by Oracle because it is so dependent on its database revenue and needs opportunities to keep coming back to existing accounts.
IBM on the other hand, is more akin to Microsoft in its revenue diversity.
It could have bundled much more functionality at a price that undercut Oracle significantly, but it chose not to.
It also would have presented the market with a clearly simplified licensing model, which the market desperately wants, especially as a counterpoint to Oracle.
Once again, IBM demonstrated its complete lack of understanding of a market it essentially created.
By early 2002 it was becoming obvious that Linux was growing exponentially in importance due in no small part to IBMs own support of the platform. DB2 had already lost the battle for market share on Unix platforms and needed desperately to win the battle for the Linux market.
However, it was Oracle that was willing to make the bold move by announcing at Oracle World its intention to convert its entire back office infrastructure to run Linux.
Oracle also began its unbreakable Linux campaign, and subsequently Oracle has never looked back owning more than 80 percent of the Linux database market to IBMs 16.5 percent, according to Gartners most recent numbers.
So while Oracle was willing to mortally wound a major partner (Sun) in its pursuit of market dominance, IBM couldnt even decide to de-emphasize its AIX business, likely in deference to the margins it was making on its large pSeries servers.
While IBMs revenue is growing on Linux, Oracles is growing almost six times faster.
I dont believe there is much IBM or any other vendor can do long-term on Windows to stem the growth of Microsoft SQL Server.
However, with Windows currently commanding 40 percent of the server market and with estimates pointing to a greater than 50 percent share by 2010, it is a key platform in which to remain relevant.
Certainly there will still be a significant number or organizations that will prefer to hedge their bets by utilizing database software that will run on multiple operating systems.
The problem for IBM is that DB2 is actually losing share on Windows (-4.4 percent in 2004), according to Gartner.
From my dealings with clients and through briefings with IBM itself, its pretty clear that IBM is challenged to get mind share within its largest channel partner, its own consulting organization.
While attempts have been made to educate the consulting side on DB2 for Unix, Linux, Windows old habits die hard.
The result is that a IBM consultant recommends something other than DB2 more than 70 percent of the time, according to at least one senior IBM consultant I spoke with last year and confirmed by a number of clients.
Those organizations that did choose DB2 often reported that IBM consultants knowledge of the DB2 platform on Unix, Windows or Linux was uneven at best.
My bottom line is that I believe the dominant platforms for databases will be Windows and Linux by the end of this decade.
The two major winners will be Oracle on Linux and Microsoft SQL Server on Windows. I believe Microsoft will actually become the leading commercial database vendor by 2015, irregardless of the adoption of SQL Server 2005, which I dont see happening in earnest until 2007.
The fact that DB2 will be a non-player within the relational database market in the future is not a reflection on the databases capabilities, which are strong.
I do believe, however, that the market no longer has room for three premium priced commercial databases, and that open-source databases, most likely MySQL, will replace IBM as the third player in the market.
This will be difficult to tell, as industry counters like Gartner and IDC will be hard-pressed to account for free or near-free software in an equitable manner.
I suggest users start to watch third-party tools and application vendors for early signs of this change.
Charles Garry is an independent industry analyst based in Simsbury, Conn. He is a former vice president with META Groups Technology Research Services. He can be reached at [email protected]
Check out eWEEK.coms for the latest database news, reviews and analysis.