Late last year, Computer Associates' Ingres r3 Database became one of the most prominent examples of previously proprietary code to be published along open source guidelines.
Other companies have released the source code of proprietary software but Ingres was unusual as it is a high-end enterprise database being used by over 15,000 customers.
CA's chief technology officer Yogesh Gupta says the software company decided to open source Ingres to fill what it saw as one of the main gaps in the open source stack -- the lack of a high-end enterprise database.
"We thought: 'What can we contribute?'" says Gupta. "We have a database that is mission critical. We also have tens of thousands of CA customers and if we move to open source they will no longer have to pay the licence. But, we are willing to do this, as this is the next big hurdle for open source."
Gupta claims this apparent bout of unprecedented altruism was actually a decision based on extremely commercial foundations. "We believe it is in the best interest of CA's business in the long term -- we hope they [users of the free Ingres database] will come back to us for support, we hope they will buy extra software," he says.
But analysts claim there may have been other forces at work here. James Governor, principal analyst at Red Monk, claims CA's decision to open source Ingres is actually about the company trying to hang on to its existing customer base.
"This claim that 'the open source market needs a high-end enterprise database', I'm not sure that's true," says Governor, "The open source database with the momentum at the moment is MySQL and it's not being adopted because it's seen as a high-end enterprise technology.
Although Governor questions CA's real motivation, he claims the decision to open source Ingres is sound. "There are customers using Ingres and this is good for them. When someone comes along and asks them [the customer's Ingres team] to swap to another database, they can say 'It's open source and we have support'. It's a good defensive move," he explains.
CA should at least be praised for giving free access to Ingres even though it probably won't attract many new customers to the database, says Gary Barnett, research director at Ovum.
"Having garnered a reputation in the 90s for doing very little for the technology they acquired, it was a surprise to many people to see their [CA's] approach to open source," says Barnett. "In my view they didn't get enough credit for open sourcing Ingres. They have open sourced a proper, pucker, enterprise-level database. It is more widely deployed in enterprises than any other open source database. It was a very clear and unambiguous approach."
At the same time as CA released Ingres as open source, the company also launched a contest with $1m in prizes to encourage developers to write software for the database. CA's Gupta says this contest was set up to help foster a sense of community around Ingres and generate interest in the platform.
"We are trying to build a community around Ingres," says Gupta. "But it was also how do we create some excitement? How do we create some buzz? The vast majority of open source projects have less than 10 people contributing. The fact is we have hundreds of people contributing to Ingres."
Since releasing Ingres as open source six months ago, the database has attracted a lot of interest from the community, says Gupta, but the results of the contest, which closed on 5 February, should give some more clues as to exactly how well received the move has been.
"We have had enormous interest from individuals and small companies," he says. "Ingres has had many thousands of downloads and hundreds of contributors. There are lots and lots working on their submission for the contest."
Gupta claims CA has contributed more to the community development model than any other large vendor. As well as the Ingres strategy, CA has also contributed to the Kernel Generalized Event Model which is due to be incorporated into the Linux kernel. The software is designed to improve Linux security and to feed performance information from systems built around the OS to management software.