One thing that makes analysis of business strategies in open source difficult (even for professionals) is a confusion of means and ends.
Take my earlier post on IBM. I asked about IBM's open source strategy in isolation, which is a mistake. That's because for IBM, open source is a means, not an end.
For IBM, open source is a means to unify product lines, to sell hardware and services. It'w a way to gain contracts for managing big systems and solving big problems. IBM is not primarily a software company. It's a systems and solutions company.
For companies like Microsoft and Oracle, on the other hand, as well as for Red Hat and 1,000 others, open source is the end. They are software companies. They can sell software support, software updates, and software services, but the end product remains software.
Open source, by its nature, drives down pure software revenues. You can put some of those revenues into other bins, you can call them support, or contract programming, or services, but over time recurring product revenues wither.
If software is just a means to a larger end, this is not a problem. It's an opportunity, because for you software is a cost. This is true for customers, and for companies like IBM that make software part of a larger sale.
But for dedicated software companies, for those who depend on software as the end of their business, open source is indeed a threat. It threatens their basic business model. It means change.
But it's business evolution in action.