In-Depth

DBMS Licensing Woes Turn Customers to Open Source Systems

Customers losing patience with negotiating database licenses are increasingly adopting open source DBMSes such as MySQL

• By Stephen Swoyer
• 12/16/2003

The upshot, suggests META Group analyst Charlie Garry, is that by 2006, a majority of customers will conclude that the current vendor-client licensing model is broken and in need of retooling. Some will look to downsize – or “rightsize” – their database infrastructures, even as others consolidate DBMSes, or, increasingly, look to open source software (OSS) database solutions.

There are indications that organizations are rapidly losing patience with the vicissitudes of negotiating with database vendors: META Group says that more than 50 percent of the database-related inquiries it receives are directly related to DBMS licensing issues.

One of the prickliest experiences that any organization can undertake today is to enter into licensing negotiations—for either new or existing database software—with a major DBMS vendor. That’s partly the case because new licensing growth has been slow to non-existent, particularly at the high-end, and all vendors are feeling the pinch. Market research from Gartner Inc. and International Data Corp. (IDC), for example, shows a decline in overall DBMS growth from 2001 to 2002, with Gartner finding a drop of 6.9 percent and IDC a smaller dip of 0.7 percent.

The problem is compounded, META Group says, by the costs associated with switching from one database platform to another, especially for so-called legacy DBMSes. As a result, many IT organizations feel as if their DBMS vendors have them over a barrel, so to speak, when it comes to licensing negotiations. It’s also not surprising that some DBMS vendors—perhaps mindful of this inequity—are restructuring the terms of their licensing and support agreements with existing customers to help realize additional revenues. Not surprisingly, Garry says, this has not been a popular move. “Never-ending pressure to keep up with licensing changes and support agreements that force new negotiations and upgrades is beginning to wear on organizations,” he writes.

Another persnickety area is that of database support and maintenance agreements. Here, too, vendors have restructured existing arrangements and in some cases have started playing hardball with customers. This can result in some nasty surprises, Garry and other experts say.

For example, conventional wisdom says that organizations should push for a cap on the annual maintenance costs associated with a software application or DBMS. As far as it goes, this is generally a good idea, conceded Pat Cicala, president of IT procurement and software asset management practice Cicala & Associates, in an interview earlier this year. But in the midst of a market drought in which new licensing revenue has all but dried up, she pointed out, ISVs are pulling a bait-and-switch of sorts: holding the list prices of their products steady even as they increase the cost of maintenance caps—and on an annual basis, at that.

Not surprisingly, Cicala said, this shouldn’t be happening: “You’ve got to look at those caps, because if vendors are no longer raising their list prices, that means that the caps on maintenance shouldn’t be going up every year. We found a client who every year their price went up by 10 percent, and they ended up saving $2.5 million worth of over-charge.”

Traditionally, this meant that organizations had two options: Pay up or take legal action against an aggressive vendor. But for these and other reasons, speculates META Group’s Garry, organizations will increasingly turn to OSS databases, such as the MySQL DBMS, which are typically licensed with more agreeable usage terms. “Certainly, open source licensing offers some relief from the death spiral many companies seem to [become] engulfed in every few years with proprietary vendors,” he writes, “but the open source movement will need to mature during the next three years to address the current concerns and misconceptions of organizations.”

One area of note in which OSS vendors trail proprietary software companies is that of service and support. But here as well, vendors such as MySQL AB have gone a long way toward reconciling the best of the OSS and proprietary software worlds, Garry says. “[They] are making strides toward striking the balance between the ‘usage’ freedom afforded by licenses”—he cites the GNU General Public License, the standard OSS license—“and the warrantees and support and control offered by proprietary vendors.”

As a result, Garry and META Group predict that organizations will standardize on Intel-based platforms for DBMS architectures, along with other n-tier architectures—such as application and Web servers --- by 2006. This will push proprietary Unix systems even further into the high-end, low-unit-volume market segment. Linux, in particular, will mature and move into mainstream application and DMBS server roles by 2004 or 2005. There it will vie for dominance with Windows-based systems.

Unix- and mainframe-based DBMSes won’t go away, however. After all, most organizations probably won’t do a complete rip-and-replace of their existing DMBS infrastructures with OSS offerings. In all probability, many shops won’t go the open source route in the first place. Garry acknowledges that some organizations will undertake consolidation efforts—i.e., standardizing on one or a few DBMSes and consolidating dozens of distributed databases on to several larger systems—but few of these are cost effective, so on a whole, database consolidation isn’t often done.

Instead, Garry speculates, a significant number of organizations will attempt to "rightsize” their database infrastructures. “There will always be ‘legacy’ infrastructure and what is new today will be considered legacy within two years of its implementation,” he notes. “Rightsizing embraces this concept and entails a conscious effort to enable introduction of changing standards and methods to the organization in a planned manner.” The upshot, Garry says, is that organizations should develop a flexible model for infrastructure planning that balances the need on one hand to reduce costs and, on the other hand, the unique capability of new technologies to promote innovation and change.