Having had a day to ruminate about the Open Source Business Conference 2008, a few key takeaways suggest themselves. It was by far the best OSBC yet, with a far more diverse audience and speaking faculty that we've had before. This naturally leads to a diverse set of "conclusions" arising from the event:
- Enterprises love open source but the business models necessary to fuel both their happiness and that of the vendors still need a lot of work.
Jon Williams of Kaplan Test suggested in his keynote, as Dirk Hohndel captures, that the more happy he is with his commercial open-source software, the less likely he will be to pay for it. Why? Because his developers will acquire the expertise over time to support themselves and because the product will mature to the point that support will be less necessary.
The vendor can respond in two ways: Innovation and proprietization (made up word). By innovation, Jon suggested that continual development of the product keeps it buggy (my word, not his) and out in front of his developers, such that support remains relevant. Vendors can also offer services like the JBoss Operations Network that make maintenance of the software easier.
A combination of both is optimal, but Dirk is right that it's a bit depressing, this prospect of the customer leaving just when you've made them the happiest.
- There is no clear leader in commercial open source anymore.
It used to be Red Hat, hands down. But Red Hat has been hands off for so long that the commercial open-source vendors have learned to fend for themselves. Microsoft and Sun have been far more interested in engaging with these vendors than Red Hat has (for various motivations). Red Hat has made some middling efforts to engage with its commercial peers, but it has not shown much desire to extend its market leadership into becoming the hub of the open-source ecosystem.
There have been good reasons for this absenteeism. Red Hat has been 100 percent focused on building its business. But Red Hat Exchange was an acknowledgement that its business, in part, also depends on the success of other open-source companies. A market of one is not very interesting.
My hope is that Jim Whitehurst will spend the time to engage with other open-source companies. We need Red Hat's leadership. We need its channels. But we also have much to offer. No open-source company can afford to bean island unto itself.
- Competition within the commercial open-source community is heating up.
This revealed itself in the open-source database panel (with deeper analysis at The 451 Group's blog), but it's also evident in the market wrangling between Novell and Red Hat, Pentaho and JasperSoft, Zimbra and Open Xchange, etc. While I think it's fair to say that none of these companies spend a lot of time angling specifically against their open-source peers, it's also true that they're ready to do so when necessary. This is good. This is healthy. Let competition flourish.
- With Microsoft and patents, the spirit may be willing but the flesh is weak.
If there were an easy answer to how to keep its patent power against the HP's, Oracle's, IBM's, etc. of the world and still engage the open-source community, I think Microsoft probably would have done it. It certainly will never be able to make much money directly from its patent licensing.
Oracle, Red Hat, IBM, Novell, and others have contributed patents to the Open Invention Network. That's how you fix the downstream problem. Microsoft is trying to have its cake (patent club at the ready) and eat it, too (engage open source). But it can't have it both ways. It really is an either/or. For it to be relevant in 21st-century software battles, it's outdated patent strategy won't win it any friends...or markets.
- As Jim Whitehurst pointed out in his keynote, we really need to do a better job of bringing enterprises/customers into the open-source development community.
Stuart Cohen's Collaborative Software Initiative is an interesting way to get there, but every vendor needs to be focused on this. We need to find efficient, protected ways for enterprises to engage with vendors, communities, and their peers. They are a massive development pool that still largely develops in isolation. That needs to change.
- We are still in commercial open source's infancy.
I think it very likely that business models and deployment models around open source will change over the years. We are nowhere close to optimizing how we sell, develop, and deploy open-source software. Not even remotely. The answers as to how will come from customers, not vendors. That's what makes it such an exciting market to be in.
We clearly have a lot of work to do in 2008 to advance open source. Where there's work, there's money. Good times for open source.