acm-header
Sign In

Communications of the ACM

Communications of the ACM

The Business of Open Source


In his discourse comparing various economic systems, Schumpeter [5] declares it is new products, new markets, and new forms of production and distribution that impel the creative destruction engine of free enterprise. Entrepreneurs strategically weave an organizational design of customer value, product offering, and production and distribution technologies that enables them to compete with, and often displace, existing organizations. Customers decide whether to accept the new firm's offerings based on their perceptions of value.

Open source software (OSS) appears to be creative destruction in action: all three of the components that fuel the destructive fire are evident. There is an abundance of new and innovative products emerging from the OSS community. The zero-cost licensing structure of most open source projects has opened up the acceptance of these products into a number of previously untapped markets. The Internet has created an environment in which software distribution costs are approaching zero. Products freely and rapidly flow across borders. There is no packaging—no shelf-space requirements. OSS is not confined to one economic system. It overcomes both the tyranny of distance [1] and oppression of borders. OSS also espouses new methods of software production that utilize the public as the production mechanism and allows for coordinated, location-agnostic access to the raw materials for these new products.

To assess if OSS has the potential to revolutionize the development and distribution of software, we must first understand how the entities involved in the development of such software are organized. Because different business models are not equal in their capacity to create value, we must analyze each particular blend of customer, product, and production and distribution mechanism. After presenting five models underlying the software development business, we focus our attention on the one that has perhaps the most disruptive potential.

Back to Top

The Business Models

We distinguish five models of software production or distribution: proprietary, open community, corporate distribution, sponsored OSS, and second-generation OSS. Whereas the first two constitute the extremes of the closed-open continuum, the other three are hybrids of closed and open models.

Proprietary and Open Communities. Proprietary and open communities both have their origins in the early days of computing, when some people freely exchanged code while others recognized there were customers for their programs and accordingly sold executable versions of their products while carefully securing the source.

The proprietary model has dominated the marketplace for decades. Firms employ programmers to develop software and customers purchase it. The code is considered a major intellectual resource, and traditional software firms protect their code from outside eyes by erecting physical and legal firewalls between their code and the outside world. Proprietary firms rely heavily on both copyright law—to ensure that "leaked" source code cannot legally be used in a competing product—and patent law—to protect their intellectual property from duplication. While the code is most often sold for license fees, proprietary firms can and do distribute their products as freeware. It is not the price that distinguishes proprietary software but rather the public's inability to view and modify the source code.

At the other extreme is the open community model, which involves the development and support of software by volunteers with limited or no commercial interest. This model dominates the OSS movement in terms of number of projects. Many of them can be located through large, Internet-based project management and source code repositories such as SourceForge, which hosts over 170,000 software projects.1 While the majority of these projects involve only one or two developers and have a small number of users, many have a vast base of both developers and users and have produced products that provide unique functionality or offer compelling alternatives to commercial products.

Corporate Distribution. Based on the high level of adoption for many OSS products, it seems apparent that quality products are being produced through the open community model. However, some entrepreneurs recognize that identifying appropriate products, interacting with open communities for support, and developing the required support skills can be challenging for many potential OSS customers. As a result, firms, such as RedHat, SpikeSource, and OpenOSX, have emerged to create value (and generate revenue) by identifying best-of-breed OSS projects, improving distribution methods for these products, and providing complementary services in order to make these OSS products more accessible to a broader market.

Sponsored Open Source. Corporations and foundations sponsor some OSS projects. For example, the Apache Software Foundation fosters the development of the Apache server and over 50 other OSS projects. Some corporate sponsors directly contribute development resources to OSS projects. IBM is a high-profile example of a corporation contributing developers to Apache's Web server. In some cases, sponsored OSS projects have been initiated by corporations releasing previously closed code and encouraging their employees to continue to work on the now open project. Eclipse, an integrated software development environment, was released as OSS by IBM, whose developers are still primary contributors to the project.

Second-Generation Open Source. Second-generation open source (OSSg2)—also known as professional open source—firms are essentially a hybrid between corporate distribution and sponsored OSS. As with the corporate-distribution model, OSSg2 companies typically generate the bulk of their revenues by providing complementary services around their products [3]. Like sponsored projects, OSSg2 firms provide the majority of the development resources required to create and maintain their products. However, unlike most corporate-distribution companies, OSSg2 firms generally do not sell licenses for their products,2 and unlike most sponsored projects, OSSg2 firms typically own or tightly control the software code and can exploit their intimate knowledge of the code to provide higher-quality service than could potential competing service providers. As the leading OSSg2 firms (including JBoss,3 MySQL, Trolltech, and Sleepycat4) are privately held or have been acquired, we do not have data on their profitability. However, based on interviews with the CEOs of the four firms mentioned here, it appears they are cash-flow positive while growing rapidly.

We contend that OSSg2 firms have a very promising business model that could emerge as a dominant model for OSS development in the coming years. Here, we examine four leading OSSg2 companies, highlighting three important benefits of their business models.

Back to Top

Evaluating the OSSg2 Business Model

We studied four companies that are among the OSSg2 leaders. In order of business longevity, these are Trolltech, MySQL, Sleepycat, and JBoss. Three important characteristics of the OSSg2 model exhibited by these firms that lead to specific benefits are: accountability (and the benefit of reduced liability problems); talent base (and the associated benefits for code quality and support); and ecosystem (and the associated benefits of trialability and quality assurance). These benefits should improve the value proposition to customers for OSSg2 products and are central to our contention that OSSg2 is a threat to traditional software firms. To quote Marc Fleury, former CEO of JBoss: "We (OSSg2 companies) are proving that professional open source can do it better and faster and cheaper than our traditional competitors."5 How the OSSg2 model addresses three specific strategic risks is discussed later in this article.

OSSg2 Leaders. Trolltech was founded in Norway in 1994 and currently has more then 4,400 customers. It manages two software products: Qt, a cross-platform application development library, and Qtopia, an application platform built for embedded Linux. A second OSSg2 leader, MySQL, was founded in Sweden in 1995. In mid-2007, MySQL's OSS relational database had 11 million active installations. MySQL is an attractive alternative to higher-cost relational systems from commercial vendors. A third leader is Sleepycat Software, a U.S. company founded in 1996. Its flagship product, Berkeley DB, is an OSS-developer database that boasts over 200 million deployments. Finally, JBoss Inc., a U.S. company founded in 2001, provides middleware through its JEMS (JBoss Enterprise Middleware System) portfolio of products. Three of these products are currently market leaders: JBoss AS (a J2EE-compliant application server), Hibernate (an object-relational mapping solution), and Tomcat (a Java Servlet container). JBoss's former CEO, Fleury, who coined and trademarked the "Professional Open Source" label, has greatly influenced the OSSg2 business model [6].


Because different business models are not equal in their capacity to create value, we must analyze each particular blend of customer, product, and production and distribution mechanism.


There are differences among these OSSg2 companies. First, Trolltech, MySQL, and Sleepycat are based on a dual-license strategy offering both commercial and OSS licensing options. Customers may use a product without paying a license fee; however, if they augment the original source code and do not wish to release the modifications under an OSS license, they must buy a commercial license. Mike Olson, former CEO of Sleepycat, acknowledges that the dual licensing strategy is a "great judo trick for competing with proprietary vendors." JBoss is based on a Lesser General Public License (LGPL) license and only receives revenues from services including software support, training, and consulting. A second difference is that, whereas Trolltech, MySQL, and Sleepycat own the source code underlying their products (which allows them to offer a dual-licensing scheme), JBoss does not own the code of the software products it services. Nevertheless, as an OSSg2 firm, it controls the code more tightly than companies based on the other OSS business models. For example, the extent of its contribution to its three leading products is 85% for JBoss AS, 95% for Hibernate, and 60% for Tomcat [4].

OSSg2 Characteristics Improve Customer Value. Three key traits of OSSg2 firms provide specific benefits to their customers and can thus improve their value proposition.

Accountability. All four OSSg2 companies we studied indemnify their paying customers from any legal liability associated with their products (potential patent or copyright infringements). The indemnity provision provides a necessary level of security for potential adopters of OSSg2 products who are still apprehensive about OSS. This accountability may do more than just bring OSSg2 companies to parity with proprietary software vendors. Sleepycat's former CEO, Mike Olson, asserts the risks of patent infringements and copyright problems are lower for an OSSg2 company: "Anyone that wants to can look at my software. If there was a claim pending, if I had stolen something, it is overwhelming likely that it would have been seen by now. [...] No proprietary vendor's customer has that degree of assurance. If there has been intellectual property misappropriated in that product, it is a secret and it may still be lurking" [7].

From Olson's perspective, OSSg2 companies offer greater peace of mind than proprietary firms, because patent or copyright infringements should be uncovered relatively early—reducing the impact the copyright infringements may have on adopting customers' systems.

Talent base. Each OSSg2 company retains talented coders, wherever they are located, to maintain and support its software products. Indeed, all four OSSg2 leaders insist they recruit from among the world's best and most productive programmers, drawing not only from the immense pool of first-generation OSS contributors, but also from a growing collection of emerging talent. MySQL and JBoss allow their programmers to live wherever they wish. In addition, recruitment is different. Typically, those employed by OSSg2 companies have an established record of contributing code and identifying bugs as volunteers prior to their hiring. They have demonstrated their understanding of the code base and their ability to fit within the OSS development culture. This is an important competitive advantage because it means OSSg2 companies reduce hiring risks without significant up-front recruiting and training costs.


Open source programs have moved beyond the desktops of code hackers and are now in production in a growing number of corporate IS departments.


Trolltech's employees (approximately 230) come from more than 20 different countries; they were recruited almost exclusively through the OSS community. Trolltech has learned that great developers want to work with each other. When asked about the criteria for hiring, CEO Havaard Nord emphasized that what really counts is "code, code, and code...merits...formal education is less important." Trolltech's employees are the company's most valued assets. The founders and the employees—owning two-thirds of the shares—control Trolltech. Outside investors have majority ownership and control of many software companies. Only in rare cases does Trolltech get outside contributions for its products. When this happens, either the submitting contributor is hired (if coding quality and knowledge of the product have been demonstrated) or the firm gets ownership of the code. As a result, Nord claims Trolltech's customer base is "extremely happy with the code," thanks to the company's careful approach to recruiting and retaining its high-skill employee base. Fleury insists that employees are JBoss's greatest asset as well: "I treat my elite developers like royalty. I overpay them. I cover my lead developers in stock. Many of them walk around with an executive package, which is rarely the case, if ever, in traditional software companies."

This focus on hiring the best programmers results in a quality of code that is at least commensurate with that of proprietary development while maintaining the benefit of "mass innovation" [4] shared by all OSS products. Similarly, the support and education offered by OSSg2 companies meet industry expectations because of the quality of the personnel.

Ecosystem. There is an encompassing ecosystem that evolves around OSSg2 companies that typically includes all the entities that gain from the OSSg2 companies' presence in the market (support services, authors, educators, publishers, partners, user communities, and so forth). This translates into multiple Web sites, email lists, newsgroups, conferences, and published materials providing up-to-date information about OSSg2 products and their applications. OSSg2 companies typically benefit greatly from their ecosystem without much strain on their resources. Mårten Mickos, CEO of MySQL, emphasizes that MySQL tries, with minimum involvement, to make its ecosystem thrive: "We try to be open about our intentions so it's easy for others to plan their business and their life around us. We try to move the obstacles of getting our product, distributing our product, using our product. [...] We just make sure the friction is as low as possible."

Potential OSSg2 customers can download and test a complete software product extensively before making an adoption decision. Because the ecosystem provides an effective pre-sales support apparatus, potential customers receive a significant advantage in the form of trialability that is limited neither in time nor functionality.

The ecosystem can also provide for an efficient, external quality-assurance mechanism above and beyond what may be carried out in-house, as Mickos points out: "When we release a new version, within 24 hours 35,000 people have downloaded and tested it. That's fantastic. Not even Microsoft has 35,000 QA engineers. [...] Just based on statistics, we know that there are enough people out there who certainly test all relevant features, without our specific instruction" [7].

The OSSg2 model thus has a significant advantage in leveraging an important ecosystem that is willing to work on its behalf.

How OSSg2 Addresses Risks. OSSg2 has an adroit answer to dealing with the major risks facing all software firms. Every firm faces three strategic risks: demand, efficiency, and innovation [2].

Demand risk and pricing strategy. Wal-Mart and Dell have altered the structure of the retailing and computer industries through their low-cost strategies. Similarly, OSSg2 firms push the cost of software acquisition to the lower limit. Assuming requirements are met by an OSSg2 product, cost-driven IS departments will be attracted by zero acquisition costs. Extensive trialability, discussed earlier, also contributes to mitigate demand risk. For OSSg2 firms, as with both Wal-Mart and Dell, revenue losses from low-cost strategies are largely offset by increased operational efficiencies.

Efficiency risk and the Internet. OSSg2 firms gain from efficiencies associated with their Internet-based infrastructures. Many employees work remotely, software is downloaded rather than packaged and distributed physically, and high trialability obviates many traditional marketing costs. Consequently, OSSg2 firms tend to have a lower cost structure than traditional firms and thus enjoy efficiency differentials over proprietary software competitors.

Innovation risk and open source. When code is open, many coders can inspect it, and faults often will be detected more rapidly than when only a handful review it. Furthermore, those who can see the code can suggest improvements and submit code changes. As with all OSS communities, the developers and supporting community members for OSSg2 projects are drawn from all areas of the world, an immense talent pool from which OSSg2 community members can be recruited on the basis of talent and contribution, unfettered by physical location. This ready supply of programmers ensures innovative ideas can be contributed to the OSSg2 community from both traditional sources and sources previously untapped by traditional software firms. This phenomenon directly attacks innovation risk.

Back to Top

Conclusion

The open source movement is challenging the status quo in the software marketplace. Open source programs have moved beyond the desktops of code hackers and are now in production in a growing number of corporate IS departments. OSS is amplifying the demand, efficiency, and innovation risks that traditional software organizations face and is driving a period of creative destruction that has the potential to permanently alter the competitive landscape within the software industry. OSSg2 firms offer a significant customer value proposition and have effective strategies that should aid their prospects for long-term survivability. However, during this period of creative destruction, we also recognize that the market is constantly changing, traditional firms are experimenting with adjustments to their strategies to address the stresses that OSSg2 is placing on their business models, and new models are emerging that will blur the lines between the categories we have outlined in this article. This makes the business of open source both extremely fascinating and highly consequential.

Back to Top

References

1. Cairncross, F. The Death of Distance: How the Communications Revolution Will Change Our Lives. Orion, London, 1997.

2. Child, J. Information technology, organizations, and the response to strategic challenges. California Management Review 30, 1 (1987), 33–50.

3. Fitzgerald, B. The transformation of open source software. MIS Quarterly 30, 3 (2006), 587–598.

4. Goetz, T. Open source everywhere. Wired 11, 11 (Nov. 2003).

5. Schumpeter, J. Capitalism, Socialism, and Democracy. Fakenham and Reading, London, 1943.

6. Watson, R.T., Wynn, D., and Boudreau, M.-C. JBoss: The evolution of professional open source software. MIS Quarterly Executive 4, 3 (2005), 329–341.

7. Wittig, C. and Inkinen, S. MySQL open source database in 2004. Stanford Graduate School of Business, 2004.

Back to Top

Authors

Richard T. Watson ([email protected]) is the J. Rex Fuqua Distinguished Chair for Internet Strategy in the Terry College of Business at the University of Georgia, Athens.

Marie-Claude Boudreau ([email protected]) is an associate professor in the Terry College of Business at the University of Georgia, Athens, GA.

Paul T. York ([email protected]) is a Ph.D. student in the Terry College of Business at the University of Georgia, Athens, GA.

Martina E. Greiner ([email protected]) is a Ph.D. student in the Terry College of Business at the University of Georgia, Athens, GA.

Donald Wynn, Jr. ([email protected]) is an assistant professor in the Department of Management Information Systems, Decision Sciences and Operations Management at the University of Dayton, OH.

Back to Top

Footnotes

DOI: http://doi.acm.org/10.1145/1330311.1330321

1As of late February 2008, SourceForge hostsed 170,539 projects.

2As discussed here, some OSSg2 firms offer a proprietary license for organizations not wishing to comply with the terms of the OSS licensing model.

3JBoss was acquired by RedHat, Inc. in April 2006.

4Sleepycat was acquired by Oracle, Inc. in February 2006.

5All uncited quotes are based on personal interviews.


©2008 ACM  0001-0782/08/0400  $5.00

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.

The Digital Library is published by the Association for Computing Machinery. Copyright © 2008 ACM, Inc.


 

No entries found