"It's still in the pipeline," Chief Executive Marten Mickos said of the plan to hold an initial public offering of his company's stock. He declined to discuss when the company planned to go public, but said, "We're making good progress, doing all the things we need to get done."
During the days of dot-com mania, companies would go public without being profitable and in some cases without much in the way of revenue, and with investor enthusiasm bubbling at the time, many of them raised millions of dollars during their IPOs. MySQL, though, is working on building its business first.
The company garnered about $50 million in revenue in 2006, Mickos said in an interview at the MySQL Conference and Expo here. That compares with $6.5 million in 2002 and about $34 million in 2005, according to earlier figures Mickos cited in a speech two years earlier.
Of the company's bottom line, Mickos said, "Profitability isn't a specific goal yet, but we aren't burning cash. We go a bit above breakeven, a bit below breakeven."
Those with IPO experience advised Mickos to "do it as late as you can because it's so hard to be a public company," but there are good reasons nonetheless, he said.
"It removes one obstacle that some customers may have dealing with us. Conservative customers want to deal with public companies," Mickos said. "And by going public you get the currency to do acquisitions."
MySQL has become a considerable power in the open-source software realm. Its rise was due in part to its integral position in the popular "LAMP" stack of software that powers many Web sites: the Linux operating system, the Apache Web server, MySQL and the PHP language that can construct Web pages on the fly.
While MySQL was borne of the Web, it's posing an increasing threat to existing database powers. In competitive bids, MySQL most often goes up against Oracle and Microsoft, Mickos said. But the company isn't trying to directly attack Oracle's core business; instead, it's angling to head the rival off in growth markets.
"We're not trying to eat Oracle's lunch. We're trying to eat their dessert," Mickos said. "We love to compete with them and win customers from them. At the same time, we don't want to fall into this David-vs.-Goliath strategy where everyone says 'Go, MySQL, go! Go kill Oracle!' There's an old Chinese (proverb): if you focus on one competitor, ultimately you become like them, and we don't want to do that."
Oracle attempted to acquire MySQL, Mickos said last year.
In a keynote address, Mickos described areas where he wants MySQL to succeed, and none of them are the big, back-end databases where Oracle has its stronghold, powering activities such corporate inventory, accounting and customer behavior.
"What we really want to be known for is the online database," Mickos said in his speech. The company wants to power operations such as traditional Web sites, the more interactive generation of "Web 2.0" companies, telecommunications industry services, software as a service, and business applications accessed through Web pages.
There are a number of open-source database rivals--PostgreSQL, Ingres, EnterpriseDB (which builds proprietary software off an open-source product)--but Mickos said he sees those competitors infrequently. Of Ingres, he said, "The team is doing the right things. I must respect them for that, but I still don't see them" in customer bids. Ingres is aiming for different customers, he said.
IBM, with its DB2 software, is another major database company. It's also a business partner, though. On Wednesday, IBM announced a reseller relationship under which it will provide MySQL to customers of its System i servers, higher-end IBM machines geared for midsize companies.
The IBM move will mean software applications written for MySQL will run on the System i machines, but DB2 will still be involved behind the scenes as the MySQL back-end transactional engine. MySQL historically has used open-source software called InnoDB for the purpose, but MySQL has been developing its own engine after Oracle acquired InnoDB in 2005.
MySQL sells support subscriptions for the database; the intermediate "gold" level of support costs $3,000 per server per year. In January, the company launched a new pricing option: $40,000 per year for gold-level support of as many installations of MySQL as desired in a company.
The new option, though it doesn't cover higher-level platinum support or versions of MySQL with some high-reliability "clustering" features, has won over some new customers. Among them are Bookings.nl, Dada, ESPN, Grant Street, IMMI, Instaclick, iStockphoto, Netshop.com, The New York Times Co., Qsent, the Swedish National Police, Trans Union and Uptilt, the company said.
In the past, MySQL development has been restricted largely to a tight group of company employees, though the company boasted of tight links to customers when it came to issues such as new version testing, feature requests and debugging.
Now the company is trying to bring in more outside contributions, Mickos said.
"We have changed our model. We have learned," Mickos said. "We had a lot of contributions before, but we weren't as appreciative, and we didn't ask for them. Now we are."
One major potential ally--Google--appears eager to help out. This week, Google published its own MySQL modifications that it uses internally and expressed hope that the changes would be incorporated in the main project.