The commoditization argument for open source.

Executive summary: Computer hardware commoditized in the 1980's, with the rise of PC clones replacing integrated proprietary systems with interchangeable parts available from multiple sources. Data communications bandwidth commoditized in the 1990's, with bundled proprietary online services giving way to the commodity internet, sweeping away CompuServe/Delphi/Genie but bringing great profit to those like AOL who chose to participate in the new commodity market instead of fight it.

Today, open source is commoditizing the software market. Participants in the software industry need to figure out if they want to embrace it or fight it, and follow the model of Tandy or Compaq, CompuServe or AOL, Windows or Linux.

Commoditization happens as markets mature.

Time goes by rapidly in the computer industry. Technology generations rise and fall like fruit flies. Moore's Law states that integrated circuit technology depreciates 50% every 18 months, and using this half life to define the length of a generation means that in the 30 years since the launch of mass-produced microcomputers, the industry has seen 20 generations of technology.

As any market matures, it commoditizes. Once-proprietary advances become generally understood, are improved upon, and become obsolete. The rapid pace of information technology has accelerated this cycle as well.

Once upon a time data transmission used proprietary networks, from the GE time sharing service through CompuServe, Prodigy, and AOL. These days, senders and receivers exchange packets through the internet, which has taken over and made data transfer a bulk commodity, generally at a flat rate per month. Suppliers insisted they were hurt by this, and lobbied for protection of their established market niches, but what actually happened?

The proprietary suppliers who refused to participate in the commodity market saw their high-margin customers defect, but this this didn't lock out those who chose to participate: AOL became the largest ISP. The commoditization of data transfer bandwidth lowered the margins of bandwidth suppliers, but it also increased the total dollar value of the market due to greatly increased demand for bandwidth, a much bigger market to supply. And of course this has been such a great benefit to bandwidth consumers that it's given rise to entire new industries.

The mainframe, minicomputer, and microcomputer markets all began with proprietary hardware, from the Univac to the Apple II. But these days, commodity PC hardware consists of interchangeable parts available from multiple sources, all of which connect through standard interfaces. The commodity PC has been the clearly dominant system for 20 years now. While some recently introduced PC hardware niches (such as 3D video) still retain proprietary components, these components plug into commodity systems through standard bus interfaces, and no PC requires any hardware component that does not have an alternative implementation available from another manufacturer. Most niches once dominated by a single supplier (sound blaster audio, 3COM ethernet) have since been so thoroughly commoditized it's difficult to name a single dominant vendor.

Commoditization does not mean that innovation ceases in that niche. On the contrary, innovation is greatly helped by the modular nature of a commodity marketplace: it's easier to upgrade to a better mousetrap when your mousetrap didn't come bundled with the house. Commoditization merely means that consumers have a general idea what they want and multiple suppliers know how to give it to them. A mature market has a usable baseline and a frame of reference. If it was static and boring, the commodities market wouldn't be so active.

Commoditization of a given market niche can take a while. The microprocessor was invented in 1969 and the first commodity microcomputer hardware (the S100 bus, the precursor to the PC's ISA bus) appeared in 1975, but the dominance of the PC wasn't cemented before about 1990. The internet was also invented in 1969, and switched to a scalable protocol (TCP/IP) in 1980, but wasn't plugged into the majority of all homes and offices in the united states much before 2000. (And free wireless in coffee shops became commonplace in 2003.)

The software market is maturing.

PC software is commoditizing now, with the gradual switch to open source. Until now, software simply hasn't been a mature market. Although hardware capacity and software data rate both doubled with each new generation, software slowed and bloated to compensate. The result was that (for example) a word processor in 2005 took slightly longer to respond to each keystroke than a word processor in 1985. Consumers were willing to upgrade to the newer word processors to get multitasking, wysiwyg printing, interactive spellcheck, higher resultion fonts, and so on. The market was not mature.

But the new proprietary OS, Vista, brings no significant new features despite being in development for 6 years. The last truly revolutionary OS in the Windows line was Windows 95. (In a larger sence this was simply Microsoft's 11-year-late answer to the original Macintosh, slightly disguised by moving the task bar from the top of the screen to the bottom, and renaming the Apple menu "Start".) Since then the rationale for new operating system versions (Windows 98, Windows ME, Windows 2000, Windows XP, and now Vista) has boiled down either to support for new hardware or pervasive bugs (especially security flaws) in the previous version. The software market is now mature enough to commoditize.Open source software is commodity software. It's a level playing field in which anyone can participate. Open source systems consist of interchangeable parts available from multiple sources. Getting a Linux distribution from Red Hat, Ubuntu, or Gentoo is no different from getting PC hardware from Dell, IBM, or Gateway, or buying network service from Cable, DSL, or through your cell phone. It's the same principle: this is a commodity. A customer may prefer one supplier over another, but is not tied to them. Customers can comparison shop, and can mix and match.

Viewing open source as commodity software.

The common thread between all the opponents of open source is that they are opposed to commoditization. Generally this means they're in a dominant proprietary position, are using it to extract higher profit margins, and are sometimes even directly controlling aspects of the market (such as the DVD Copy Control Authority). Although most of them know that commoditization is good for the market as a whole, they don't expect it to be good for them personally.

They do not expect to benefit from the growth of the industry due to lower margins and loss of market share. They don't want a competitive industry because they don't believe they can compete.

Cornering the market

Markets seldom de-commoditize naturally. Disruptive technologies are often a strong commoditizing influence (allowing a new player or players access to a stagnant proprietary market), but in the "better mousetrap" category the commodity status of a technology is itself an advantage.

This doesn't mean suppliers aren't motivated to make a concerted effort to swim upstream. Attempts to resist commoditization, or to de-commoditize existing markets, are as old as capitalism. It's called "cornering the market". The general theory is to buy up all the food and then sell the hoarde at a higher price during the ensuing famine. It allows a small number of people to get rich at the expense of a large number of people, and is the reason things like antitrust legislation exist. The main defense against it is consumers recognizing that it's massively against their interests, and refusing to go along with it.

Microsoft spoke about its' attempt to "de-commoditize [internet] procols" in the first Halloween memo. It has made a number of attempts to de-commoditize the hardawre market, from Palladium to the DRM in Vista. Its' entire bundling strategy is an attempt to deny the "interchangeable parts" aspect to head off the "available from multiple sources" part.

The "network neutrality" debate was a response to AT&T's attempt to decommoditize data transfer by adding toll lanes to the "information superhighway". The "broadcast flag" is an attempt to decommoditize audio and video data formats (hijacking an incremental upgrade to lock off the new format under proprietary control; common tactic). The "sue your own customers" flailing of the MPAA/RIAA/BSA/SCO is an attempt to intimidate the market into doing what the supplier wants. Suppliers also appeal for sympathy ("don't copy that floppy"), threaten ("what we did will stop and nothing will replace it, a new dark age!"), and sometimes just flat-out lie.

Attempts at decommoditization won't stop. The "genetically modified foods" debate is a reaction to Monsanto's attempt to decommoditize grain production, a market niche that literally predates recorded history. To prevent niches we care about from being de-commoditized, we need to realize that what we're defending is a commodity marketplace. We must be able to explain how it works and articulate its benefits, but first and foremost we have to be able to name it, and say that it's what we're fighting for.


Bundling is a common method of fighting against commoditization, by tying a commodity niche to a proprietary one. Often, only part of a business model is commoditized, such as music publishers which historically both produced and distributed music.

The internet has completely commoditized music distribution, but the editorial function of publishers still adds value (wading through the slush pile of demo tapes to find talent, and recording/mixing/producing the result). By tying production of music to distribution and charging only for distribution, publishers cannot avoid coming into direct conflict with the nature of the internet: all bits are the same and sending them is too cheap to meter. Charging for copies on the internet makes no sense.

The long term solution for publishers is to unbundle the editorial and distribution functions, charging directly for the creation of content. (For an example of this in action, see Diane Duane's "the big meow".) These problems are not unsolvable. But working out the details of a new business model, switching to it, and making it scale isn't necessarily easy, either. There will be winners on the internet (there already are), but it's a new game which must be won all over again. It's easy to forget that when it's time for the next round, digging a foxhole in the old winner's circle won't stop challengers from running the race.

Unfortunately, when faced with a suppler who refuses to unbundle, often the only course of action is to abandon the supplier and replace all components of the bundle. New suppliers will emerge that produce an unbundled component that interfaces with the commodity, and use their lack of bundling as a competitive advantage. (Musicians posting their music to their myspace page, video produced for upload straight to YouTube, etc.) They will capture new business, and time will take care of the rest.

This is why open source software for Windows (FireFox, OpenOffice, Thunderbird, the Gimp, etc) is important. Although Windows is the center of Microsoft's proprietary software monopoly, and switching the market to a commodity alternative for that component (Linux) is central to commoditizing the software market, it's just as important to unbundle. Interchangeable parts available from multiple sources is the nature of a commodity marketplace, and unbundling the proprietary market creates fault lines. Divide and conquer.


The strongest anti-commoditization weapon of entrenched proprietary interests attempting to prevent the software market from maturing into a commodity market is an increasingly dysfunctional patent system. This is widely recognized to be against the interests of the software industry as a whole.

Most players have no choice but to take a "mutually assurred destruction" approach to accumulating patent portfolios they can cross-license as a defensive measure against everyone else's patents. Unfortunately, this is not compatible with open source. Bundling a patent portfolio on to each software producer or software project interferes with the unbundling required to commoditize a mature market.

Patents as applied to software have two fundamental problems:

Ordinarily, as a market matures the interesting and fundamental patents expire. But in software, techniques in use for more than 20 years have fresh patents approved on them every day, simply because the technique has never been patented before and the patent examiners don't have the extensive expertise required to find or recognize prior art that only exists outside the patent system.

Several efforts are underway to provide prior art for patent examiners, and to streamline the process of challenging overbroad and overly obvious patents. This is an important issue, but beyond the scope of this paper.

In the pharmaceutical industry, 20 years is barely enough time to get some drugs through the FDA approvals process and into production, but in the computer industry a 20 year patent lasts 13 generations of technology. This means that patents concerning 8-bit computers that pluged into your television hadn't expired yet when 64-bit PC hardware came to market. Windows 3.1 hasn't been commercially viable for over a decade, but patents on it are still in force.

The computer industry is not the same as the drug industry, and until the patent system recognizes this fact and gives different durations to patents on different categories of invention, the system will fail to serve the needs of the computer industry.

A more appropriate duration for comptuer patents would be seven years, which is approximately the length of time between the release of Windows 95 and Windows XP, or between Windows 2000 and Vista. It makes no sense for patents on computer technology to remain valid after support is discontinued for the software versions that introduced them.

An interesting historical note is that this 7-year value was the original duration of patents when they were first introduced into US law. What the computer industry really needs isn't a radical new development but a return to historical norms.


Occasionally, arguments have been made that laws using the phrase "commercially significant" do not apply to "free software". Arguing that open source equals commodity software can clarify this confusion, commodities are of vital commercial significance.