Give open source databases (OSDBs) two years or less, and they'll be ready to take on the enterprise. That's not just some conclusion I came to out of my head; that's the latest word on the subject from AMR Research, a research firm that specializes in providing analysis on enterprise software. The report has not yet attracted the attention it deserves--but it will. Even IBM and Oracle may start feeling the pressure in the next year, if AMR's view on this can be trusted.
It's hard to beat anything that's open source on the cost issue. If a company actually wants to spend money on open source, it can buy maintenance and support contracts--no doubt a source of comfort to those who are more familiar with proprietary software. Even with such contracts, OSDBs should cost a company less than their proprietary counterparts, since there's no initial licensing fee. As for the ease of administration--well, that raised an interesting point. It seems a database with fewer features (assuming all the important ones are included) is easier to administer. Currently, OSDBs include fewer features than proprietary databases, which is just one factor in their ease of administration. Many, if not most, of the IT people interviewed by AMR seemed confident that OSDBs would continue to increase in quality while possibly avoiding the "feature bloat" plaguing proprietary systems.
I admit I'm extrapolating. Here's the relevant quote: "Even among those not using an OSDB...25% believe OSDBs are ready for mission-critical systems now, with only 13% believing that they will never be ready. Those that are using OSDBs are more resolute: 80% believe they will be ready for mission-critical use either now or within 24 months." Those who have to maintain mission-critical databases take a very hard look at what's involved, and you can bet--if you don't already know from experience--that the features of the database play a large role, because they have a major effect on the ease of maintaining the database. You don't want software that's suffering from feature bloat to be a mission-critical element in your network.
So if OSDBs aren't quite ready for prime time yet, what do companies use them for? According to the report, "about 60% of current OSDB use is for custom systems, websites, or systems exposing data to a company's outside customers." Everyone knows that DevShed.com uses an OSDB--MySQL by name--to serve information at its web site. But did you know that C|Net Networks uses the very same OSDB to serve information at its web site? They have offices in 12 countries and something like 65 million folks checking them out regularly. And their whole point of existence is delivering information. You can possibly get more mainstream and mission-critical than that, but it isn't easy.