Recent news reports have repeatedly shined a dim light on the financial performance and prospects of leading database server vendors. Should customers and users panic or be patient? Are database vendors, or at least their server products, a dying breed?
After gloating through Informixıs well-publicized financial and management mishaps, Oracle saw its stock price tumble by 30 percent following its early December announcement that fiscal second-quarter revenues and profits failed to meet expectations (they did increase). In particular, sales of Oracleıs database servers registered only a slight increase of about 3 to 4 percent, much less than expected. Sybase, which seemed to be leaving its woes behind, opened the new year by announcing that its estimated fourth-quarter earnings would also fall below expectations, possibly into the red zone. Sybaseıs stock promptly plummeted by 33 percent. (Sybaseıs actual earnings report was scheduled to be published in mid-January after press time.) Many analysts who explained the troubles plaguing Informix, Oracle, and Sybase cited market saturation and the increasing commoditization of database servers as the culprits.
Itıs easy to agree with both rationales, but I cannot help but question them. The saturation argument seems compelling: relational databases are an aging product line, and most companies that want them have several already. But what about the frequent reports of rapid growth in data warehouses and marts? Data warehouses require separate servers from the operational systems. In addition, todayıs data warehouse and data mart architectures actually call for several servers to house the data staging platform, which is usually separate from the queryable warehouse. The rise of data marts is likely to increase appetites for database servers because multiple data marts are usually distributed throughout a company, though each one is probably accessed by fewer users.
Apart from data warehousing, packaged applications are also fueling the demand for database servers. Many of these licenses are sold as part of a bundle, not directly from the database vendor, and per-user license fees may be discounted. This stresses the vendorsı revenue line, but it doesnıt mean that database servers are a dying breed. I also strongly suspect that electronic commerce will eventually add to the demand for database servers. Online transaction rates may still seem comparatively low, but the growth curve is steep, and many new "e-companies" are sprouting.
Unlike the early days of a productıs life cycle, a saturated product usually has accumulated a large and experienced cadre of professionals who can fill developer and administrator positions. Yet todayıs job market is very hot, especially for Oracle developers and DBAs. Demand for qualified professionals exceeds supply. Perhaps itıs just a timing problem and demand will soon slow down, but I see another inconsistency here.
I do agree that database servers are much more of a commodity now than they were years ago, but that could change. For most corporate applications that do not have highly specialized requirements, any of the leading database servers will usually do the job. Itıs a lot like choosing a car to drive to work as opposed to racing the Indy 500. Yes, each car has its differences, but they all get you to the office. Itıs a matter of degree rather than kind. But even here, vendors are working to differentiate their database servers, especially in how they manage complex data. Database servers five years from now will probably diverge significantly, but customers seem to be taking their time upgrading to new features. So differentiation may not lead to steep growth in new sales or upgrades, but it does tend to limit defections.
For managers and executives who must sign purchase orders, the financial health of a vendor is a serious concern. Itıs easy to overreact to bad news, but jumping ship quickly is not always a good idea. If you jumped from Informix to Oracle this year, you probably had a feeling of dıjı vu when Oracle started stumbling. On an even larger scale, IBMıs resurgence under CEO Louis Gerstner shows that even a massive company can recover.
There is no denying that major database server vendors have had some bad days. But it may be too soon to forecast doom.
Congratulations to Dave Kalman, the publisher of DBMS, and his wife Audrey on the birth of their first child, Harrison Samuel Kalman, on January 8. Dave, Audrey, and Harrison are all doing well. Harrisonıs Web site is home.earthlink.net/~kalmans.