In reading through IDC's excellent report, "2007 Industry Adoption of Open Source Software, Part 2: Project Adoption," analyst Matt Lawton stumbles across an intriguing observation in open-source software adoption. He apparently believes it is a weakness of the current open-source landscape, but I believe it is a strength.
The observation? That IT departments do most of the services around open source, rather than third-party consulting companies.
The shocking result here is the complete dearth of service providers that are currently being tapped for installation, training, and other services associated with open source software, regardless of where in the software stack that software sits. Less than 1% of the projects have attendant services sourced from service providers.
And these services are needed -- only 21% of the projects did not require attendant services. Service providers, the natural third-party source of attendant services in the proprietary software world, are virtually nonexistent in the open source world. With 68% of open source projects sourcing attendant services from internal IT groups, this looks like a throwback to the 1970s and 1980s, where internal IT groups handled all the system integration requirements for their software deployments.
IDC believes that if open source software is to continue its penetration into mainstream and enterprise environments, the ecosystem must evolve to include many service providers that are knowledgeable about open source software and can handle the integration, implementation, and training needs of end users in particular. (28)
Why is this a bad thing? Enterprises are unshackling themselves from proprietary, expensive licenses and reinvesting that money in the gift that keeps on giving: people. That's how I read the data.
This becomes especially pronounced when one considers two other questions IDC asked. The first is, "Compared to all of your other current IT initiatives (whether Open Source software or not), please indicate the importance to your organization of your top 10 Open Source software projects." The answer? Across the board (Applications, Infrastructure, and Application Deployment and Development) these open-source projects were rated "Critical" or "High Importance" by 73 percent of respondents.
In other words, these projects weren't simply casual afterthoughts that didn't require outside help. They were perhaps some of the most important IT projects the enterprise was deploying. Those may be best kept in-house.
The second question - "Does your organization plan to increase, keep the same, or decrease its activities in the next year for each of your top 10 Open Source software projects?" - is also revealing. 90 percent of respondents are planning to increase or keep the same (very healthy) level of investment in open source.
Clearly, if the projects weren't working out, we'd see this number come in much lower. But apparently this open-source liberation from vendors (Many opted to support themselves rather than purchase commercial support, according to IDC's research) turns out to be a Very Good Thing.
Would the industry be better off with more and better-qualified consultants? Absolutely. Actually, a better way to ask this might be, would the industry be better off with more consulting companies actually owning up to how much open source they're already implementing? The answer to this is a resounding "Yes!" Accenture, SAIC, Cap Gemini, etc. are all using a tremendous amount of open-source software, but generally aren't working with the project leads/maintainers to do so (and I suspect aren't telling their clients about it, either). I know of one that has done well over 100 projects with an open-source infrastructure project, yet has failed to work with the (commercial) project source at all.
Then again, maybe this is a source of strength in open source, too. There's plenty of money in open source for commercial vendors, but there's also plenty of opportunity for customers to build and support their own open source derived projects. And why not? Since when must enterprises lock themselves into vendors forever?
That's the 20th Century. We're in the 21st Century now. The Customer Century.