More on open source servers
For many enterprises, the server OS presents a quagmire: They don't want to pay too much for the server OS on which they rely, but at the same time, they don't want their server OS makers going out of business. The big question is whether there's enough money in open source software to build strong and stable enterprise OS makers.
If you run your business using Microsoft's Windows server OSes, then you really don't have to worry. The Redmond giant is rolling in cash thanks in no small part to the high prices it charges for its desktop and server OSes and the client access licenses it requires to connect one to the other.
But open source companies are different. They don't sell their open source software per se, and therefore they don't make a lot of money. Peter Wayner over at InfoWorld wrote recently about two highly valued open source companies: MySQL, which Sun bought for $1 billion, and Red Hat, currently valued by the market at around $6 billion. This was what he had to say:
Yet for all of this wealth, there's a feeling that these businesses succeeded by being hybrids. They use the open source vision to attract users, but their business success comes by pushing proprietary options. Thus, the dark secret is that their open source version are merely a form of marketing.
That's perhaps a tad harsh. What attracts people to open source software is -- the open source software! Companies like Red Hat then make their money by offering something extra that people are willing to pay for, to go with the open source software. Open source software is like a free hamburger, and essentially Red Hat is saying is "Would you like fries with that?" And, of course, the fries cost extra.
Back in the real world, those fries turn out to be support contracts, documentation, training, professional services and, in some cases, proprietary software as well. The real rub is that companies like Red Hat don't actually make much money from these activities, certainly nothing like the wonga Microsoft generates from selling its software licenses. In the first quarter of fiscal 2011, which ended May 31, the $6 billion Red Hat had a net profit of just $24 million. Microsoft, in its latest quarter, made $4 billion.
Now don't get me wrong: These were good figures for Red Hat. Admittedly, the company was coming out of some difficult trading conditions, but software subscription, training and services sales were up more than 20 percent compared to the same period last year, and profits were up 30 percent. That's impressive stuff. And Red Hat has a billion dollar cash pile so there is no immediate reason whatsoever to be concerned.
But $24 million isn't that much, is it? Not when you consider that this is the company that stands behind the leading enterprise Linux, which is used and relied on by some very big-time businesses. Heck, Microsoft makes that before lunch every day.
A more reliable way to make money from open source software might be to sell hardware that uses the software -- no one minds paying for something that they plug into the wall, even if they would never, ever consider paying for software by itself. Software wants to be free, but hardware has no such aspirations, after all. Indeed, Oracle may end up making heaps of money from Solaris by selling highly optimized database appliances.
If Red Hat wants to make serious money from Red Hat Enterprise Linux (RHEL) then the only way to do it may be to use it as the foundation of some sort of appliance. But somehow one can't see Red Hat doing that in its current form, and the company's $6 billion market cap is enough to put off all but a very rich few from buying the company to do the same thing. And in any case, what would a Red Hat appliance actually do?
This leaves an awful lot of very big companies relying on RHEL, an enterprise OS from a leading open source company that really doesn't make a lot of money. That can't be healthy, and it makes one wonder if there isn't something not quite right with the open source software business model.
Paul Rubens is a journalist based in Marlow on Thames, England. He has been programming, tinkering and generally sitting in front of computer screens since his first encounter with a DEC PDP-11 in 1979.