If a large corporation unloaded its open source database project to some high-profile investors, would anybody care?
The question became relevant this week when Computer Associates (CA) spun the open source Ingres database off into an independent entity called Ingres Corp., to be controlled by Garnett & Helfrich Capital, an investment firm based in Menlo Park, Calif.
CA has kept a stake in Ingres, noting that it will cooperate on product development, industry partnerships and marketing. But industry experts said that the answer could still be "nobody" cares.
Noel Yuhanna, a senior analyst with Cambridge, Mass.-based Forrester Research Inc., believed the news was probably a death knell for the fledgling database, which CA open sourced in 2004.
"Ingres didn't have much of a presence in the market," Yuhanna said. "People had had difficulty dealing with Ingres software; almost a fear, and that's why they open sourced it last year."
When Ingres went open source in November 2004, the database's founder and chief architect Mike Stonebraker supported the move, but said a product that contained good technology didn't necessarily make it a commercial success.
Ingres had been alive and well since the early '90s he said, but "benign neglect" on the part of CA could doom the database. The open sourcing of Ingres was a sign he believed was the end of the abandonment.
If Yuhanna's analysis of Ingres over the past year is any indication, however, it would appear that Stonebraker's worst case scenario has come to fruition.
"Spinning off can help in certain degrees, but if CA cannot really manage Ingres then there is no other company that can manage Ingres," Yuhanna said. "[CA] has depth and knowledge far more than anyone today."
What went wrong?
The causes of Ingres' slow demise were many, but Yuhanna attributed the most significant decline to ineffective or non-existent marketing from CA.
"It takes a while for an open source DBMS to become widely adopted. Just look to MySQL; it has been there for almost a decade, and just now it has the market penetration. Five or six years ago, if you asked, no one cared about MySQL."
Yuhanna, who was once a user of Ingres, said when customers look to implement an open source database, they are looking for new innovation in their applications and don't want any heavy baggage.
"The technology is mature, but the problem is that when you bring old technology to an open source community, the question always is, 'How well will it work with the community?' and 'If we need to change the source code of 20-year-old technology will that really be modular in design, and can we work with that?'"
Simply opening a product to the open source community is another myth this announcement disproved, Yuhanna said.
"Just because it is free does not mean you should use it," he said. "It has been the same thing with Cloudscape. We haven't seen great deal of penetration, and it hasn't had penetration given that IBM was unable to develop great marketing around it."
Marc Fleury, the outspoken founder of open source vendor JBoss, said as much during a session at the recent Open Source Business Conference in Boston.
"People come to me almost as a last ditch effort and say their company is dying and they are thinking about going open source. Well, that's the bullet in your head right now because they are going to cut the only revenue stream they have left," he said.
In a month where Ingres was criticized by analyst and competitor alike, the future of the open source database market could never look brighter.
MySQL version 5.0 was released, as was an update to PostgreSQL that included over 120 enhancements and changes in query syntax and utility commands.
"Whenever people speak about an open source DBMS, they are talking about MySQL," Yuhanna said.
The open source industry is also in a period of transition where the smaller database companies were moving up the chain to be supported by larger vendors. Yuhanna cited Austin, Texas-based Pervasive Software Inc. as an example of one company that has pledged support for PostgreSQL.