The business of arranging millions or billions of zeros and ones in exactly the right order—also known as the business of software—has undergone many significant shifts over the history of computing, and we might be about to experience yet another.
In the early days of commercial computing, software was seen as a constituent of the hardware platform. You bought a computer from a manufacturer such as IBM, and the computer came fully installed with all the software you would need.
Independent software vendors—ISVs—existed during the mainframe era, but it was not until the minicomputer and client server era that ISVs became truly significant. Minicomputers manufactured by companies such as Digital and Data General ran their own operating system, but by the early 1980s were typically running application software created by an independent company such as Oracle. When IBM created the IBM PC in 1981, it outsourced the operating system to a small company called Microsoft. By the end of the century, organizations such as Oracle and Microsoft had revenues that dwarfed that of the hardware manufacturers and they had become some of the most valuable companies of all time.
As a business, software had some incredible advantages. The incremental cost of producing software is negligible: Once you have perfected a piece of software you can replicate it at virtually no cost. Software also has a fairly low barrier to entry: You don’t need a massive manufacturing plant to create a piece of software; a single programmer can create a billion-dollar software product on a $500 computer.
However, over the last 10 years the software business has become a lot tougher. There are many reasons, but in particular open source software has created a significant and persistent disruption for the ISV. The concept of open source software emerged in the early 1980s from Richard Stallman’s GNU initiative, which aimed to produce a free open source software version of UNIX (GNU is a recursive acronym: GNU’s Not UNIX). Linux emerged in the early 1990s under a GNU license, and by the mid-2000s Linux had become the most significant server-side operating system. Red Hat created a billion-dollar company providing services and distributions of Linux.
Many enterprises were reluctant to build their infrastructure around open source software, preferring the security of a commercial software vendor such as Oracle or Microsoft. However, the economic advantages of Open Source proved decisive and by now, Open source has become an acceptable—and sometimes even mandatory—ingredient in the modern enterprise application architecture.
Ironically, just as the acceptance of open source in the enterprise has finally become established, the enthusiasm for open source in the software industry and venture capital community is starting to wane.
The image of open source as developed by hip hackers in garages is far from accurate: Almost all open source is developed inside commercial software companies. For example, the open source databases MySQL, MongoDB and Cassandra are predominantly supported by the Oracle, MongoDB, and DataStax corporations. These companies invest in OSS in order to gain market share or competitive advantage. Often, the investment is backed by venture capital firms hoping to find another Red Hat—i.e., another billion-dollar OSS company.
Unfortunately, the billions of dollars invested in open source has not created another Red Hat. If there is no return on the open source investment, then eventually the investors will invest somewhere else. There is already significant scepticism around open source within the Venture Capital community, and a sense of fatigue within large companies struggling to establish Open Source business models.
Open source has powered an enormous amount of innovation and has literally changed our world—mostly for the better. However, programmers need to eat, and corporations need revenue. Unless a truly viable and repeatable open source business model emerges, I think we will see a decreasing investment in open source software going forward.