| Craig S. Mullins
Death Valley Days for the DBMS?
By Craig S. Mullins
Reports of the death of the DBMS and database vendors have been greatly exaggerated. It is true that several of the large database vendors have stumbled as of late. Sybase, Informix, and Oracle all have recently felt the sting of bad news.
Informix is reeling not only from aggressive accounting practices but from a product strategy that includes multiple code bases for its DBMS and over-marketing of its object-relational marriage with Illustra that resulted in Universal Server. Informix is working on righting its wrongs. It has recently announced plans to consolidate all of its code bases on a single product named Informix Dynamic Server. This is goodness. Likewise, its marketing message is harkening back to the Informix glory days when the message was performance, performance, performance. However, all of this "goodness" still takes a back seat in the public's mind (as well as in the mind of the investment community evidently) to the revenue problems Informix is experiencing. Informix's share price has deteriorated from a high of $24 in March of 1997 to a low of around $4 in December 1997. The cause of this erosion is the recent restatement of earnings by Informix (see chart). So, to say the least, it has been a rocky few months for Informix.
Sybase has yet to fully recover from several down years which saw it slide from its perch as a technology innovator to somewhat of a late adopter. They have made strides to rectify this situation in the latest release by adding object-relational capabilities called adaptive components, as well as improving functionality and adding features. And Sybase too has renamed their product to Adaptive Server Enterprise. I guess, when the going gets tough, the tough rename their product!
And, for Sybase the going has been quite tough financially, too. It lost almost 20% of its share price in a single trading day (closing price on 12/31/97 was 13 5/16 while the closing price on 1/2/98 was 9 15/16) when Sybase announced they too would be restating earnings. In late January 1998, Sybase reported fourth quarter revenues of $223.2 million and the net loss for the period was $25.5 million; full year revenues for 1997 were restated down for the first three quarters resulting in a total of $903.9 million and a net loss for 1997 of $55.4 million.
Even mighty Oracle recently announced less-than-expected earnings for its fiscal second quarter. For the quarter ended November 30, 1997 overall revenues increased 23% to $1,614 million from $1,311 million in the same period last year with net income for the period up $187 million versus net income of $179 million in the second quarter of fiscal 1997. This sounds good though, doesn't it? But the stock dropped almost 10 points on December 9, 1997 in reaction to Oracle making this news public. The problem is that Oracle's announcement calculated out to earnings per share of $0.19-which was 4 cents below expectations. What a difference 4 cents can make!
Of course, all of these problems are the result of the Asia-Pacific crisis or weaker than expected international results? Right? That always seems to be the excuse, but I don't think that is so-not in this case.
What Does All of This Mean?
But does all of this suggest that the DBMS will no longer be the lynchpin of the modern application system. The answer to that should be an emphatic "no!" Databases still drive most new applications being developed today. The problem is that the new features of the latest and greatest DBMS versions are not yet compelling enough to drive new and upgrade revenue. Believe me, this will change.
The database industry is going through a phase in which databases are morphing into something more than just a store house of corporate "data." Instead, databases are storing data and the procedures that access the data. Additionally, databases are becoming more flexible, but with the flexibility comes complexity. And people fear adding complexity without a valid reason to do so. This is just sound judgment. The object revolution will change their minds. This phase is different from the relational phase when older DBMS technology was replaced with new relational technology. This time the old technology (relational) is being adapted to work with the new technology (object orientation). These things take time.
Usage of objects and components is increasing and the eventual place corporations will store these must be in the database. Databases provide the atomicity, concurrency, integrity, and durability (ACID) required of persistent data. As customers require object support, they will migrate to newer versions of each respective DBMS, thereby driving increased revenue.
Another factor to consider is that eventually the DBMS vendors will discontinue support on the database versions that are most popular today. When Oracle7 and Sybase SQL Server 11, for example, are no longer supported, customers will migrate to the new releases. They will have to. The fact is, customers should begin to migrate before support is discontinued just to ensure the continuity of their production applications. Moving to a new release requires training, conversion, modification of standards, and a migration plan, to name just a few issues. And who wants to be forced to do all of that within a limited time frame?
The Other Players
And what of the other major DBMS players: IBM and Microsoft? Both are doing extremely well. IBM's DB2 Universal Database (yet another new product name) provides object-relational support as well as any in the business. Likewise, its performance and architecture is sound, IBM is finally putting its marketing muscle behind the product, and IBM has an increasing number of customer success stories for DB2 UDB.
Microsoft is enjoying similar success. They are basically riding the wave of Windows NT success with SQL Server 6.5, with 7.0 soon to come. However, according to Dataquest, Oracle still sells more database servers on NT than Microsoft. But, knowing Bill Gates who really would be surprised to see them give away SQL Server as part of Windows NT? And what would that do to the dynamics of the market?
Microsoft still has some work to do to make SQL Server and Windows NT full-scale enterprise level players. But, they will accomplish that feat, perhaps sooner than most people expect.
Of course, both IBM and Microsoft are much larger than any of the other major DBMS players. So, they can ride out a dry spell in new DBMS revenue. However, given their current success, it does not look like they will have to.
So things may look somewhat bleak, but the game is far from over. The database is still the primary underpinning of the bulk of applications running and being developed. The era of lightening fast growth may be behind us, but sustained growth will continue for quite some time.
From Computing News and Review, March 1998.
© 1999 Mullins Consulting, Inc. All rights reserved.