Open Source is not Commoditizing Software

Perusing the title of this essay I imagine the reader bursting out, "what? Preposterous!" Or at least, that's what I imagine if I was writing this a hundred years ago, and my subject was petroleum oil. If that were the case, you'd be right, of course. The development of the oil industry did result in the reduction of its product to a near-perfect commodity. In my actual context, the first decade of the 21st century and the continuing evolution of the software industry, however, I stand by the statement: Open Source software is not turning software into a commodity.

"But," you might say, "you can download and use for free any flavor of Linux or BSD, OpenSolaris, Apache projects, Mambo, Drupal, Wikimedia, SugarCRM, Compiere, Nagios and the list goes on and on." The acceptance of the commoditizing effect of Open Source software (OSS) has gained credence through the statements of various industry luminaries, such as Johnathan Schwartz. If the new CEO of Sun says OSS is commoditizing his own products, who am I to disagree.

Except, at the risk of repeating myself, what is happening to software is not that it is turning into a commodity. Without wanting to sound too pedantic, a commodity is defined as a good that can be obtained from any supplying vendor without preference. The suppliers market interchangeable 'versions' of the commodity. No one will claim that SugarCRM and Apache HTTP server can be used interchangeably. Even restricted to a single category, choosing OpenSolaris or RedHat Linux involve very different skills requirements and deployment practices.

To draw the contrast even more sharply, we can return to oil. Software is not oil, a sea of bits to be stored in enormous steel tanks, pumped through pipes (Senator Ted Stevens notwithstanding) and purchased without preference at whatever retail outlet is most convenient to your computing needs. Network connectivity may be oil, but software most definitely is not.

So what is OSS doing to software? If not commoditization, then what? Extending our analogy to the oil industry may prove illustrative. Finding oil, extracting it from the earth, refining it and distributing it required the development of very sophisticated tools and techniques. As the technology of oil exploration, production and distribution developed, innovative applications of the newly available energy source developed in parallel. Oil-fired ships, airplanes and automobiles all drove demand for oil quicker than the industry could create the supply. Each new application had broad direct and indirect effects on our society, and created massive amounts of new economic value.

The technology of oil production also made some people very very wealthy -- the Hughes fortune, for example, came not from the ownership of oil fields but from the manufacture and sale of innovative drill bits and well equipment. However, as innovation solved the initial set of hard problems, the value slowly leaked out of the tools and equipment. They transitioned from exotic, differentiating advantages to everyday appliances, utilitarian, indispensible, but simply table stakes for an oilman.

I submit this is what the emergence of OSS represents to the software industry. The hard problems of multi-user operating systems, of relational databases, of office productivity, of systems management, and the like, have been solved. Or at least the solutions to the problems that most people want to solve most of the time have been solved. Those tools are no longer exotic. Linux doesn't drive innovation in and of itself, no more than Sugar does. They adopt and implement the hard-won lessons learned in the R&D labs, Comp Sci departments, and IT departments over the past forty years.

This is in no way meant to denigrate what these OSS projects represent -- only a fool spends time reinventing the wheel. And if these OSS projects do not innovate at the core of what they are, by standing on the shoulders of giants they free up energy and resources to innovate all around the OS or the database or the office suite. Building a new consumer electronics device? Grab an embedded Linux distro. Want to host Web Services on it? Embed Apache in it. Tie it into live, network-delivered customer service? I hear SugarCRM works quite well in a hosted mode, maybe you can leverage that and extend the customer-facing function onto your new toy.

Solving the hard problems doesn't mean the solution becomes a commodity. It just becomes commonplace. When formerly exotic tools become commonplace, it enables exploration at deeper and deeper levels, unlocking the potential for value creation in previously unimaginable ways.