Open source and the commoditization of software [LWN.net]

Ian Murdock has posted his chapter from the upcoming O'Reilly book Open Sources 2.0; it is called "Open source and the commoditization of software. "If Red Hat's business model is wrong, then what is the right business model for Linux distribution vendors? In my view, the Dell model can be taken a step further than any of the Linux distributors have thought to take it. After all, what are open-source technologies but commodity software components, and what are Linux distributions but assemblers of those components into products the end customer finds useful?"

to post comments

Open source and the commoditization of software

Posted May 27, 2005 18:47 UTC (Fri) by JoeBuck (subscriber, #2330) [Link]

I don't think Ian is describing the Red Hat situation quite fairly.

It is certainly true that many proprietary software developers are specifying RHEL, because treating one flavor of Linux as "standard" reduces their costs. (Actually, many of those application developers support Novell/SuSE as well, particularly those that have a significant presence in Europe).

But Red Hat is not simply a repackager as Ian implies. They use their RHEL revenue to pay a very large number of talented people to hack free software full-time (the Linux kernel, GCC, glibc, Gnome, and lots more), and everything those people put out is usable by all of the other distros. To a buyer of electronic design automation software (to give just one example), Red Hat's fee is in the noise compared to their other costs, and that money is going toward the development of free software for all to use.

In a similar way, TrollTech uses part of its revenue from commercial licenses to pay people to hack KDE.

I don't think that Ian has really presented an alternative that can generate as much free software as Red Hat has managed to generate. Certainly there are aspects to Red Hat's agreements that I don't like, but they make the source code to everything they do available, and there are the "white box" people who will build it for you. You don't get Red Hat support if you don't pay, but that's fair.

Two serious flaws

Posted May 27, 2005 18:59 UTC (Fri) by hazelsct (guest, #3659) [Link] (2 responses)

I see two serious flaws in the arguments presented here.

First, RedHat's model is not at all about proprietary extensions, as was the problem with Unix. It's about branding, and service. The Oracle, SAP, and Dell certifications to play nicely RedHat are underwritten by RedHat, and they sell that certification as the RedHat brand. I don't see how this is against the spirit of free software, since everything remains 100% free.

Second, Dell's mix-and-match component assembly strategy worked because they mixed and matched scarce goods and sold the whole to the customer. Need a $2000 machine? No problem, they'll assemble the parts. Need a $600 machine? Fine, but you get inferior parts with lower capacity.

This approach will inevitably fail with free software because all of the best parts are available at one price: zero. And even ten year old 5 GiB hard drives can store more than enough software for the typical desktop user. (Data storage is another thing, but the point is that *anyone* can afford the storage needed to run any Linux distribution.) Want a full blown KDE-GNOME-Samba-LAMP server-desktop? No problem, pay $0, just apt-get the relevant packages. Want a minimal XFce desktop, or an even simpler router? No problem, again $0 and a bit of time and bandwidth to apt-get.

Progeny Componentized Linux has its advantages for end users in terms of shorter "release cycles" if they can be called that (though it's not clear the market will support any shorter than K/Ubuntu's six months). But the economics bear almost no resemblance to the Dell model. And RedHat has a whole lot more going for it, in terms of both market position and ecosystem relationships, than the Unix companies of the 1980s.

Sorry Ian, I respect a lot of your contributions over the years, but this isn't your best.

Two serious flaws

Posted May 28, 2005 18:17 UTC (Sat) by tzafrir (subscriber, #11501) [Link] (1 responses)

Maintinance does cost money. A major part of what RH sells ia a maintinance contract for RHEL for a period of 5 years.

Two serious flaws

Posted Jun 8, 2005 18:53 UTC (Wed) by hazelsct (guest, #3659) [Link]

Indeed, that's what I meant by "service" in "It's about branding, and service".

Class EnterpriseLinuxVendor extends OpenSourceVendor implements UNIXVendor

Posted May 27, 2005 19:45 UTC (Fri) by dmarti (subscriber, #11625) [Link] (3 responses)

Red Hat is doing a drop-in replacement for UNIX, to fill the hole left by the collapse of UNIX. In order to make that work, they have to provide the business interfaces that all the UNIX ISVs and customers are used to -- such as offering a limited set of "golden" versions that ISVs can do their own qualification and support dance with, and offering a pricing plan that customers understand.

My printer has a parallel port and a USB port. When I first got it I used the parallel port. Now I use the USB port. "Enterprise" Linux products are a way for Linux vendors to plug into businesses that don't yet support the new standard.

Class EnterpriseLinuxVendor extends OpenSourceVendor implements UNIXVendor

Posted May 29, 2005 13:07 UTC (Sun) by hansl (subscriber, #5086) [Link] (2 responses)

> they have to provide the business interfaces that all the UNIX
> ISVs and customers are used to

Yes, but what position will Redhat be in when commodotization
moves higher up the stack? When those same Unix ISVs (such as
Oracle) are being "drop-in" replaced? Will Redhat push adoption
of open source Oracle alternatives as hard as other distributions
will?

-Hans

Class EnterpriseLinuxVendor extends OpenSourceVendor implements UNIXVendor

Posted May 29, 2005 16:51 UTC (Sun) by khim (subscriber, #9252) [Link] (1 responses)

It'll never happen. Only dumb providers will try to push "same old, same old" technologies (like SCO) and will be burned. Free Software will move ever higher on the stack, true, but there are will always be something to be sold even higher - or the whole idustry will be unneeded and in such a case RedHat will not be needed too...

Care and Feeding?

Posted May 29, 2005 23:54 UTC (Sun) by man_ls (guest, #15091) [Link]

The essay you link (The Care and Feeding of FOSS) is so full of factual errors and so devoid of logic that it hardly supports your point. (The use of Apache and IIS as the main examples of how software starts commercial and then goes free is particularly unfortunate.)

Apart from that, you forget why people want to use Oracle on an Oracle-certified platform in the first place. It's a matter of confidence; corporate users want to have an assurance that things will work as expected. If the tables turned, and (say) postgreSQL was Red-Hat-certified, those users would have an equivalent assurance; Red Hat would thrive, and Oracle would become irrelevant. Or, said in a different way: Free software wins, the corporate provider of free software is still alive, and in a better shape than before.