Lots of reasons ... look up "barriers to entry" in any MBA textbook.
Without making a case for specific barriers to entry that apply in the
case of MySQL's business model, you're not actually making a point.
Throwing economics terms around without any context for applicability
does not constitute evidence for an argument. For instance, "barriers to
entry" applies to anyone entering any market, including small to medium
firms like MySQL, individual developers trying to sell individual pieces
of software like TextMate, open source software support firms, huge
hardware companies that are more well known for the software side of
things, and so on. It doesn't matter what your business model is, how
big your company is, or what the legal situation is -- there will always
be barriers to entry. The point where all three of those things come
into play is in determining *what* barriers to entry apply, and how
insurmountable they are.
There are plenty of business models. There are many fewer *profitable*
business models, and almost all of them require that a product or
service *exist* and outrank its competitors on important *non-price*
differentiators of quality as perceived by customers.
All attempts to enter a market require some initial investment (barring
pathological exceptions). It's actually easier (read: "cheaper" in terms
of invested resources) to build recognition and desirability for a piece
of software when its source is open than when it's closed. In the open
source case, you can just toss it out in the wild and see if it picks up
enough steam to provide you with some opportunity for generating revenue
(whether it involves direct support, feature commissions like RMS
provides for GNU Emacs, secondary advertising revenue, or something
else); in the closed source case you have to find some way to convince
people to give it a try without opening the source to them, which
significantly limits your potential audience even if you initially give
it away, but leaves you opportunity to develop a per-unit retail revenue
stream.
People seem to have some strange impression that open source development
means no revenue streams -- but what it really means is that *different*
revenue streams are available to you. You can either follow in the
footsteps of people who have already developed some of those revenue
models, or you can pioneer a new one if you're clever enough and
potentially reap huge rewards. The differing development models result
in differing opportunities. Thinking there aren't any opportunities
demonstrates a lack of imagination, not a lack of opportunities.
Red Hat is another good example of a thriving open source business
model. However, in the case of RHEL, I would argue that it *is*
technologically superior to its competition, with the possible exception
of Solaris. There just isn't another server OS out there that runs on
commodity hardware with that level of security, reliability and performance.
Well . . . there is, but I don't think you're counting other possible
examples in your comparison (like the BSD Unix variants). I'll just
assume you're lumping all Linux distributions together when you say RHEL,
and not excluding other distributions like Debian and Slackware.
) seem much more