Linux in Government: Providing a Successful Model for OSS Enterprise Users and Linux Companies
Not too many years ago, I had an encounter with a subordinate who reported to me and another manager. I considered the event unpleasant, worthy of dismissal, so I met with the other boss to discuss it. During that discussion ,the question of coachability came up. We reached a consensus that perhaps we lacked the ability to coach our subordinate. We later met with the young man and asked him what it would take for us to coach him. Knowing his job sat on the bubble, the young man at first professed an eagerness for coaching but then finally admitted no one could coach him. He said that he was not coachable. At that moment, I knew the opposite was true, and over time he became an excellent manager himself.
Allowing oneself to become coachable requires what I sometimes call an existential moment of courage. It requires an ability to allow someone else to reach us with our defenses down and to admit we do not know something. It also requires a high degree of self confidence and knowledge that we need to have the same courage and vulnerability each time our need for coaching surfaces.
Bob Dylan once wrote, "he not busy being born/Is busy dying". Most people interpret that to mean the when we stop growing and learning, we start the final journey of our lives.
People who reach a certain level of competency in any field often forget that learning and growing continue to be requirements for excelling. For that reason, we all need to find strong models of success to emulate. The most successful man I ever met told me that he spent every moment he could studying the lives of great men. That man once served as the Chief of Staff to Dwight D. Eisenhower.
In the emerging industry we call open source, finding people to emulate does not come easily. One company in particular, however, has demonstrated a successful commercial model, about which few Linux people know and fewer understand. When I mention the name JBoss, I'm surprised by how few Linux advocates and government CIOs have much awareness of the project or the company.
On Freshmeat, you can find this description of JBoss:
JBoss is an Open Source, standards-compliant, Enterprise JavaBeans application server implemented in pure Java. JBoss provides JBossServer, the basic EJB container and JMX infrastructure, JBossMQ for JMS messaging, JBossMail for mail, JBossTX for JTA/JTS transactions, JBossSX for JAAS based security, JBossCX for JCA connectivity, and JBossCMP for CMP persistence. It integrates with Tomcat Servlet/JSP container and Jetty Web server/servlet container, and enables you to mix and match these components through JMX by replacing any component you wish with a JMX-compliant implementation for the same APIs. The goal is to provide a full J2EE stack in the Free/Open Source software world.
This description tells us little, though, about a company that now owns the market in which IBM's Websphere and BEA's WebLogic once fought each other for control of the J2EE application server space. How did an open-source company that gives away its product with an LGPL license and that allows developers to bundle, include and redistribute freely its software in their own products become the industry bell-weather? In answering that question, you basically answer what everyone has been asking open-source software: How do you make money with a straight Free Software license?
Figure 1 provides a visual depiction of JBoss's growth. Last summer, I conducted an interview with the founder of JBoss, Marc Fleury. I learned that as with most open-source advocates, he started the company in a garage, set up a project, attracted like-minded developers and strove toward building and releasing working software.
Unlike other open-source projects, however, Fleury began generating revenue by offering training and personal support. He worked to train developers to use his product and answered their questions when they ran into trouble. While providing an income for himself, he also established a community of interest that ultimately blossomed into a robust Open Source community.
Fleury's initial activities as a small consultancy may provide the point of delineation between JBoss and other similar ventures. In fact, my own success in business occurred when I began offering training, seminars and a lot of content about my products, waiting patiently for the business to develop.
Fleury solved the problem that we often hear about open-source companies and projects: He provided people with clear information on how to use the software and always was ready to help. Contrast that with projects in which you're expected to find information from mailing lists or forums--the same forums in which you are blasted for asking questions already answered in a previous thread.
The vast majority of open-source projects have significant challenges that have caused commercial vendors to shy away from them. In speaking with open-source leads at Apple Computer, HP and IBM, each sited the inability to find someone at open-source projects with whom to interface as the biggest business problem. The simple fact is, business people need other business people with whom they can interface.
And, that's only the start. Distributions such as Red Hat, SUSE and Mandrake have similar concerns when it comes to working with open-source projects. That's why some famous projects do not make it into these distributions. A cross section of concerns people have mentioned to me include a lack of:
adequate support and maintenance
visibility/certainty around product road maps
functionality/ease of use for IT managers, particularly across enterprise size environments
stable business models to fund new development and expand into new product areas
structured and scalable partner ecosystems devoted to enabling customer success
Additional concerns include projects becoming stale and misunderstandings and worries over intellectual property risks.
If you use the model that has evolved around JBoss, you should find that it contains normal open-source aspects, such as free licenses (under LGPL), source code, enterprise-quality software, observable quality assurance processes and a robust community. You also can find extended business offerings, including training, professional sales and support (24x7x365), indemnification, product roadmaps and management and quality documentation. Therefore, JBoss has become a staple at places such as Apple Computer and HP.
In Bernard Golden's book, Succeeding with Open Source, he talks about something he calls an "Open Source Maturity Model". In his model he provides criteria for scoring mature organizations; they are software, support, documentation, training, integration and professional services. According to this model, those companies that provide adequate functionality in each of these areas can be scored according to the model. In this article, Golden rates JBoss and provides an overview of the Open Source Maturity Model. He calls his model a product-independent methodology that architects/developers can use to assess the suitability of a wide range of Open Source offerings.
I applied Golden's criteria to several successful open-source projects, including the OpenOffice.org project, GNOME and Mozilla, and found them lacking in the six criteria. I also compared them to some Linux desktop distributions and found them equally lacking. Considering this, it does not surprise me, for example, that the Linux desktop has had some difficulty breaching the enterprise desktop.
I have one major disagreement with Golden's six criteria, however. I believe it should include the quality and quantity of an open-source project's community. Without community support, no open-source project can succeed.
From an enterprise perspective, Golden makes a lot of sense. JBoss does stand out as a strong business, one on which one might want to model her or her own business. One even would recommend it to venture capitalists, who made so many mistakes with open-source investments.
And although JBoss embodies the quintessential open-source model, we cannot ignore Red Hat, which also adequately covers the six criteria presented in Golden's model. One difference does exist, here, though. If Red Hat does not feel it can do an adequate job at something, the company refrains from offering it as a supported product.
In addition, we might want to use some caution when providing Golden an honorary knighthood. Some IT managers, enterprises and others may consider Doc Searl's Do-It-Yourself IT (DIY IT) theory to be a more inviting concept. In Doc's model, the consumer becomes the provider, and in many organizations that's adequate. In fact, for many businesses, one could not provide the resources needed to build these businesses if he or she solely followed the Open Source Maturity Model.
I have a high regard for JBoss and wish more open-source companies would follow the JBoss business plan. I also have the highest regard for new and unfunded projects that start off like Fleury's EJB-OSS did, back in 1999. We can applaud the people who joined Fleury and made JBoss wildly successful. We also should applaud everyone who reaches out in an attempt to realize his or her own dreams and aspirations. Along the way, I hope these people become coachable and look for successful models to emulate.
Tom Adelstein lives in Dallas, Texas, with his wife, Yvonne, and works as a Linux and open-source software consultant with Hiser+Adelstein, headquartered in New York City. He's the co-author of the book Exploring the JDS Linux Desktop and author of an upcoming book on Linux system administration, to be published by O'Reilly and Associates. Tom has been writing articles and books about Linux since early 1999.