Lincoln to Open SQL
March 6, 2006 Hesh Wiener
In 1865, John Wilkes Booth shot Abraham Lincoln. Today, one of his relatives, Cherie Booth (Blair), is the First Lady of the United Kingdom. In its youth, IBM was insanely jealous of its rivals. But by the mid-1970s, the mature IBM was so secure in its dominance of information processing that it felt it could, without risk, tell the world one all about one of its ideas, a database language called SEQUEL, later dubbed SQL. Today, competitors’ implementations of relational database engines built around SQL threaten to do to even more harm to IBM than the nearly fatal PC.
IBM and other DBMS suppliers are under increasing pressure from open source SQL systems, particularly MySQL, which is on course to be certified for by SAP. The process could take another year or two, but even now the potential has rattled the big three of proprietary DBMS packages, IBM, Oracle, and Microsoft. Developers can get working DBMS systems for free, smaller companies can get software licenses quite affordably, and large enterprises are finding that the deals they are offered either directly by software suppliers or indirectly as part of total hardware-software bids seem to be a lot more attractive than they were only a short time ago.
A prominent example, but not the only one, is IBM’s introduction of a DB2 co-processor for its high end mainframes, a development that will boil down to a big reduction in total cost for the giant enterprises that are its most loyal and most profitable customers. A more interesting example may be that of Microsoft, which includes in its rich SQL Server package components that IBM and Oracle use to not merely increase but to actually multiply the cost of their premium DBMS offerings. Microsoft would love to double the fees it gets for its enterprise package, but even at its current, far more modest, pricing levels, it knows it must pay more attention to products coming up from below than those pressing down on it from above. Unless conditions in the software market change quite radically and most unexpectedly, Microsoft is more concerned about whether it will have to give away core DBMS technology as it does Web browser software than it is about hitting up big users for the kind of money IBM and Oracle have been able to get.
As it stands, the main constraint on Microsoft in the DBMS market is its insistence on confining SQL Server to Windows. If it ever offered versions of its DBMS for Linux, Unix, OS/400, and z/OS, chances are it would slaughter IBM and Oracle in a very short time. But Microsoft often plays by the rules IBM invented and then, perhaps by accident or error, abandoned. Microsoft values control at least as much as market share, and knows how to define and redefine its own market in ways that no other software provider can match.
IBM, when at its peak in the 1970s and 1980s, as it vanquished competitors and exploited a favorable change in the political climate surrounding antitrust enforcement, ultimately suffered as its self-confidence metastasize into hubris. By the time its management got a first, crude understanding of the impact Unix, Windows, then Linux, and ultimately open source in general would have on its software business, by the time it could begin adjusting to the pressure X86 chips put on its hardware business, by the time its army of ambassadors–the typewriters–were replaced by DOS and then Windows on desktops, its only escape from doom lay in its ability to replace customers’ computing professionals with its own employees. This insight, the brilliant retreat into services devised by Lou Gerstner, allowed IBM to claim the one-third of computing expenditures it had formerly disdained.
IBM figured out that it could no longer dominate the hardware third, that its growth in the software third was severely constrained, but also that no worthy rivals were able to grab the remaining third of customers’ budgets, the portion allotted to personnel, until IBM, still in possession of its political capital in the corporate world, made outsourcing not merely legitimate, but fashionable.
Nevertheless, IBM, along with Oracle and Microsoft, are threatened by a trend fueled by the kind of optimism Lincoln expressed in his Gettysburg vision: A nation of the people, by the people, and for the people. It’s not a certainty that open software will ultimately prevail. There is no protection guaranteed by the delightful and generous spirit of open source developers, any more than a democracy can be protected from a tyranny of the majority, as de Tocqueville wisely if cynically observed.
The battle for DBMS market share is far from over. If anything, it is growing more heated. Data from Gartner based on surveys compiled in 2004 but released in mid-2005 suggest that IBM and Oracle each got about $2.6 billion in revenue from DBMS software, giving them each about a third of the total market. Microsoft, despite confining itself to the Windows world, had about $1.5 billion in DBMS revenue, or 20 percent of the market. Because the open packages are essentially free to acquire and because the Gartner data don’t measure support revenue, MySQL, PostgreSQL, and their ilk remain below Gartner’s radar. But when small businesses build Web storefronts, it’s the open software that often does the DBMS work, and the aggregate amount of commerce the countless small enterprises on the Internet produce is a significant part of the online economy.
There is no question the high-end proprietary DBMS packages from IBM and Oracle provide features and functionality that less costly programs cannot offer. Nor is there little doubt that SQL Server is a richer offering than the open packages, if thinner than that of the Big Two. But where DBMS cost savings translate into a vital competitive advantage, users, however reluctantly, will ultimately sacrifice software quality for business advantage. Let the programmers cry if they want, but where the choice is between lean and dead, a business is going to take the route of survival.
Additionally, the computing environment available to the parsimonious users isn’t as stark as it used to be, and it promises to become more fertile with each successive release of open packages. The upshot has been a move toward light or free versions of the proprietary DBMS core offerings on the part of the three top vendors, accompanied by an increased emphasis on software that adds value to DBMS software that can no longer be stupendously profitable. This is why the big software vendors are buying or building application suites with more gusto than they are adding bells and whistles to their core DBMS offerings.
They don’t have to know much about Lincoln to realize that they can’t fool all of the customers all of the time.
The executives whose careers at IBM, Oracle, and Microsoft depend on DBMS profits may regret the changing rules of the game, but they ought to be grateful, not angry, because the developments that are forcing them to change are occurring largely in the post-industrial nations. As tough as things may be, it would be a lot harder on the DBMS leaders if the emerging alternative was a proprietary offering from India or China rather than open products from the cultures in which they have learned to prosper.
It’s hard to say whether IBM, Oracle, Microsoft, and the other big software vendors appreciate that the open software movement (or phenomenon if you prefer) is not only a cultural reaction to the computer industry’s software oligarchy, but also an expression of Western civilization’s survival instinct.
Sure, the big players would like things to remain as they used to be. Who in their position wouldn’t?
But they can’t.