Gould: Oracle to Red Hat: It's Not Your Father's Linux Market Anymore

> The only difference between bojan’s account and mine is that he fails to recognize that this is precisely what a proprietary differentiation strategy is. In short, it’s about earning your living by hustling to make a better mousetrap than the other guy.

Jeff, maybe you are under impression that I think public availability of code means that any idiot can run a successful business around it. Absolutely not. There are so many other aspects that a great business make.

The other day I had a discussion with Florian here on LWN about Red Hat v. MySQL AB. He claimed that if you wanted to get Linux (distribution), you could get it anywhere, while if you wanted to get MySQL, you had to get it from MySQL AB. That claim is, of course, misleading. Because there is no such thing as Linux (distribution). You have to _make_ one (give it a try here: http://www.linuxfromscratch.org/ - it's not that easy). So, I countered that if you wanted to get Red Hat Linux, you had to get it from Red Hat.

And this is what you call proprietary, because it comes from a vendor. If Oracle, for instance, put much better support and development behind their RHEL clone, they will then be making better Red Hat Linux (or RHEL) than Red Hat and will eventually win over more customers. Of course, if then someone else does the same thing to them, the "ownership" will change again. All of this is not easy to do. Fedora community didn't spring to life yesterday. Red Hat employees didn't accumulate their knowledge overnight.

This is not what we (or anyone I know) call proprietary in the world of software. That better mousetrap here is neither patented nor protected by copyright in such a way that would prevent others from copying it. Just about the only thing you cannot copy is Red Hat's expertise and trademarks. Now, that is hard to do - building you own brand and expertise.

This just goes to show that giving companies monopolies is not always necessary for them to take a good portion of the market. As a well known anti IP economist noted: ideas floating around are not worth much. Only the ones inside someone's head are.

Now, let me give you a practical example. There is simply no way I can convince my employers to put Debian or Fedora boxes in place of RHEL boxes. And it has nothing to do with particular software RHEL contains (in fact, I'm pretty sure all of it runs on Fedora, without modification - that's what runs on my workstation). It has to do with the fact that RHEL comes with an iron clad contract for support. The company I provide services for is a global company, with over 100,000 employees worldwide. They simply cannot afford to give SLAs to their customers without having professional level support for the OS.

Now, if Oracle were to provide an honest effort of actually supporting a Linux distribution, by having an open source development community and by employing a wide range of open source folks, that would be an entirely different matter. I'm sure management could be persuaded to switch, if the price/service was right.

Software is complicated. Good software is even more complicated. Even if everything is open source, it takes a great deal of effort to mount a formidable challenge against an established player. Even when using their own parts. None of this make the parts or the player proprietary. It just makes it hard, complicated and risky.