IBM's proposal to acquire Sun Microsystems Inc. may, for now, be off the table.
But several experts say IBM should still strive to complete the deal for one key prize — open-source database MySQL.
Despite being the inventor of the relational database, IBM has played second fiddle to Oracle Corp. in the market for many years, apart from a short blip earlier this decade.
According to industry research firm IDC, Oracle had $8.34 billion in database revenue in 2007, giving it a 37.6% share of the market. IBM was second, with $4.88 billion, or 22.%, primarily from DB2 and Informix revenues. It was barely ahead of Microsoft Corp. (21%), which has been catching up to IBM for the past decade, with its SQL Server database likely already having more users.
IBM could recapture the relational database market by injecting MySQL, already wildly popular among Web 2.0 firms and start-ups, with its vast, storied portfolio of database patents, said Paul Vallee, executive chairman of database support services provider The Pythian Group.
This would involve making as much as 40 years of database R&D and product development open source in order to quickly transform MySQL into a full-fledged enterprise database credible to the largest of customers.
At the same time, IBM would maintain MySQL's popular business model (free to users, except for enterprises).
The move would be an unmitigated boon for enterprise database users, who would gain access to a beefed-up MySQL that would continue to vastly undercut Oracle and Microsoft on price.
If done right, it would lop billions of dollars off the $22.1 billion database market, Vallee said.
“It's an aggressive strategy that would actually change the marketplace completely over five to 10 years,” he said.
Miriam Tuerk, CEO of Infobright Inc., a Sun-backed start-up that makes a storage engine for MySQL, “agrees 100%.”
“MySQL is already grabbing significant market share from Oracle, and with IBM's brand, R&D capabilities and customer relationships, this may turn out to be the best part of an acquisition of Sun,” she wrote in an e-mail late last week before the acquisition talks turned sour. “I already know of many opportunities which would instantly convert to us and/or MySQL should this transaction go through.”
Sun lacks database know-how to execute this strategy
Vallee conceded that translating IBM's R&D into actual MySQL features will take some engineering work but that it shouldn't be a big problem.
“Patents exist because they protect innovations that are otherwise easy to implement,” he said.
Sun couldn't execute on such a strategy because it lacks the database know-how and the deep financial pockets, Vallee said. “IBM could do this way better than Sun,” he said.
IBM also has experience with such “scorched earth” strategies. When IBM threw its support behind the Apache HTTP Server in the late 1990s, it helped the free, open-source Web server quickly overtake the Netscape Web server, which was Netscape's main source of revenue.
On the other hand, IBM's attempts to pry users from Microsoft Office by offering Lotus Symphony for little or no cost have yet to bear fruit.
Would there be cannibalization?
Upgrading MySQL into an enterprise database could hurt IBM's flagship DB2. Vallee, however, said IBM has long treated DB2 as a “cash cow,” milking longtime users for maintenance fees and upgrades.
A stronger MySQL wouldn't hurt DB2's new customers, since there aren't many, Vallee said, nor would it cause many existing customers to consider a lengthy, tricky migration.
MySQL could hurt Informix, which IBM acquired in 2001 to help it regain the database crown, albeit for only one year.
IBM hasn't put much marketing muscle behind Informix for years, though its user base remains strong and loyal.
What MySQL would cannibalize is DB2 Express, the free database IBM created in reaction to MySQL and other open-source databases.
But “it's not like DB2 Express does IBM much good or harm; it's just there,” Vallee said.
Independent database analyst Curt Monash also likes the strategy, though he is dubious about whether MySQL could ever be bulked up enough to become a true “Oracle-killer.”
“If IBM wanted to produce a true high-end, open-source DBMS, it would be easier via PostgreSQL than MySQL,” he wrote in an e-mail.
If IBM is not interested in a scorched-earth strategy, Vallee said it could aim for a more modest but sure-fire plan: Wring more money out of its paying corporate users.
Sun tried to do that last year but, after an outcry, backed off.
“I think IBM would have the guts to enforce certain rules that Sun didn't try,” Vallee said. For instance, Sun doesn't require support customers to purchase subscriptions for all of their MySQL databases, something IBM would have the leverage to demand and enforce.
“I think IBM is in a better position to say, 'Either we support all of your databases or none of them,' ” Vallee said.