In through the out door

Red Hat’s open-source software is an open book that it urges users to rewrite — and a glimpse into the future of business.
By G.D. Gearino

Standing on a stage at N.C. State University, Red Hat Inc.’s president and chief executive describes to students and faculty his first visit in 2007 to the company’s headquarters, which are on the university’s Centennial Campus. It was a Sunday morning, Jim Whitehurst recalls, and when he arrived at the appointed time, he found the doors locked. Whitehurst, who had been Delta Air Lines chief operating officer and was being considered for CEO at Red Hat, was flummoxed. His taxi had disappeared from sight, and he was alone outside an apparently empty building, his initial interview seemingly gone awry before it had begun.

His audience is captivated. Here’s the top guy at the world’s leading seller of open-source software telling a story every student who ever has suffered through some indignity as an intern can relate to. And Whitehurst surely reacted with the same thought any entry-level newbie would allow himself: What the hell kind of operation is this?

But his tale is only getting started. After finally being rescued from the sidewalk by Red Hat Chairman Matthew Szulik (then also its CEO) and taken to a coffee shop, he got stuck with the check after Szulik sheepishly confessed to leaving home without his wallet. Another high-ranking officer took him to lunch. Again, a money problem emerged, and again, Whitehurst paid. “At the time I was thinking, ‘Are they really interested in me or just hungry?’” On the way to the airport, the exec realized he desperately needed gas, which Whitehurst bought. After making him pony up all day, the least they could do was offer him a job.

What Whitehurst doesn’t point out to his listeners, however, is that he probably was the subject of the world’s first open-source CEO job interview. He did the traveling and the paying — the heavy lifting, in a sense — while Red Hat reaped the benefit. That’s the essence, and beauty, of the world of open source, and no company has capitalized on it better than Red Hat.

In tech circles, “open source” is a widely understood concept, a phrase as common as “showing blitz” is to a football fan. But it bears some explanation because within that phrase is not just the foundation of Red Hat’s success but also a peek into the future of business. Open-source software is that which is available to anyone for tinkering. It is more common for software, particularly when engineered to power home computers or network systems, to be closed-source — which is to say, its coding is a closely guarded and copyrighted product. The benefit is obvious: If you have exclusive custody of the coding behind a desirable piece of software, you can become very rich. This is how Windows made Microsoft Corp. fabulously profitable. But open-source software is tossed out into the world for anyone to play with, with profit motive set aside in favor of communalism. Red Hat was the first notable company to successfully bridge that divide. It is a capitalist enterprise that has managed to extract profit from a community endeavor.

Red Hat packages and sells a computer operating system available for free to anyone with the rudimentary tech skills to download it. Not only that, the company works hard to put no-cost versions into the hands of users and encourages them to tweak it. Joel Berman, Red Hat’s senior director for global field marketing, likens the process to a farmer who welds a custom-made implement to his tractor — with the difference being “what [that farmer] didn’t do was go back to the manufacturer and say, ‘See what I did.’” Red Hat, in contrast, wants so badly to know what farmers are doing that it gives them tractors and asks only that they share with everyone the results of their tinkering.

That free tractor has a name: Fedora. After being updated and improved, it becomes the Linux-based operating system Red Hat sells. (Quick history lesson: Linux is the open-source software created in 1991 by Linus Torvalds, a Finn who for all his digital genius wasn’t much of a prophet. In an e-mail that accompanied the software’s initial release, he described his effort to fellow geeks as “just a hobby, won’t be big and professional.” Today, Linux-based computer platforms are hugely professional, used worldwide by such mainstream organizations as the New York Stock Exchange and Ecuador’s internal revenue service — both of which use Red Hat’s version.)

Here’s how it works: The Fedora Project, a quasi-independent group overseen by Red Hat, supervises the work done by the Fedora “community” — the multitude of tinkerers invited to add features, improve on each other’s work and retool applications. Every so often, typically between 18 months and three years, the current version of Fedora is retrieved from the community. “We essentially take a snapshot,” project leader Paul Frields says — meaning that the software is frozen in its updated state and turned over to Red Hat’s experts to fix bugs and test security. When completed, the software is sold as the latest version of the company’s signature product, Red Hat Enterprise Linux.

There’s a small downside and a huge upside to this process. The downside is that Red Hat, to some degree, turns over control of its business to an unruly (and at best, loosely supervised) bunch of cowboy programmers. Needless to say, this is unusual in most businesses. Strict control of the product is an obsession with corporate leaders, who typically strive for uniformity in everything they do and produce. The upside can be expressed in a single number: 10 billion. That’s the number of dollars Red Hat officials estimate they would have spent so far on research and development had its product been proprietary, closed-source software. Maybe now you see the appeal of an open-source world.

Let’s pause here to pose a question (the first of two) that any thoughtful, cost-conscious executive is surely asking himself/herself right now: How can I get some of that free research and development?

Bob Young’s answer: Just ask, and you shall surely receive. It is his belief that the business world has stealthily become much more open-source than anyone realizes. Young, who co-founded Red Hat and now operates an online book-publishing company in Raleigh called Lulu.com, sees examples everywhere. In fact, he makes the case that the most open-source line of work around — the business in which competitors are not only free to help themselves to each other’s work but are actually encouraged to do so — is the legal profession. To prove his point, Young asks that you imagine a lawyer who spends months and maybe even years pursuing a case through the system and finally ends up making an oral argument before the U.S. Supreme Court, where his novel and clever interpretation of the law is embraced by the court and entered into precedent. “As soon as the words are out of his mouth, any other lawyer can use those words without his permission,” Young says.

The work behind that new Supreme Court precedent — the R&D, so to speak — would have been paid for by the lawyer’s client, who was on the hook for the billable hours that led to the decision. But neither the client nor the lawyer would be able to make any claim to the new interpretation of law. As Young points out, it would be immediately available at no cost to anyone who would like to use it for their own cases.

The main impediment to an open-source business model is, ironically, the law itself — specifically, rules that protect intellectual property. As a general principle of law, one person can’t help himself to someone else’s work without paying compensation of some kind. While this obviously benefits the creator of such properties, it inadvertently encourages the constant reinventing of the wheel. Advances in technology come much more slowly when groundbreaking work is kept proprietary but much more quickly when it’s shared. As Young points out (and give him credit for quoting both Thomas Jefferson and Isaac Newton in the same thought): “Jefferson says lighting another candle doesn’t diminish yours, and the other person is enlightened. Knowledge is like that. That’s why knowledge needs to be shared. Too much intellectual property will stymie the ability, as Newton put it, to stand on the shoulders of giants.”

In the tech world, that impediment often is sidestepped by the “general public license,” which basically requires that anyone who upgrades or improves open-source software agrees to make those improvements available to all other users at no cost. That’s how Red Hat can vacuum up all upgrades to its product — so long as it makes its own improvements to the software likewise available. What the company sells, then, is service and convenience, more so than the actual product (which, as bears repeating, is not only free elsewhere but Red Hat’s version sometimes is lifted wholesale, stripped of the corporate logo and sold by other companies). Or as Young, in describing Red Hat’s business model, puts it analogically: “I can make a car in my garage that’s the equivalent of a Toyota Camry. The reason I don’t is because Toyota does it better.”

There’s clearly a market for service and convenience. In its most recent fiscal year, Red Hat posted net income of $79 million on revenue of $653 million.

The second question: So what other businesses or industries are ripe for an open-source approach? An obvious one is the newspaper industry — which, as is painfully apparent these days, needs all the help it can get. In one sense, a newspaper is a basic manufacturer: Raw materials (newsprint and ink, mostly) go in one door and a finished product comes out the other — hardly an operation lending itself to an open-source model. But the gathering of news could be ripe for that approach, and in fact reporting already is no longer the exclusive domain of paid professionals, as bloggers have made abundantly clear. Many newspapers now recruit volunteer “reporters” for online coverage of such low-intensity topics as community news and parenting issues. It’s not hard to imagine that someday much reporting could be open-source, with professional editors performing rewriting and fact-checking functions.

But that’s theoretical, something that perhaps will unfold in the future. Red Hat’s Whitehurst, when asked the question posed above, pointed to two examples of businesses opening part of their operations to the world in the hope that tinkerers could solve a particularly knotty problem. Just a few minutes of research fleshed out the details of both. In the first, the chief executive of a Vancouver, British Columbia, company that owned an apparently tapped-out gold mine in Ontario attended a seminar 10 years ago at which Linux and open-source technology were much discussed. Rob McEwen later told Fast Company magazine that it sparked a light-bulb moment: If smart people are willing to work cooperatively on a computer operating system, maybe other smart people would be willing to help Goldcorp Inc. find more gold. A year later, he posted all of the mine’s proprietary geological information online and offered a prize to anyone who, after sifting through the data, could lead him to more ore. More than 1,400 people accepted the challenge, and suggestions began to roll in. Within a year, the mine was producing nearly 10 times the gold it had in 1996, when production had fallen to a dishearteningly low level.

In Whitehurst’s second example, Netflix Inc. is in the middle of a five-year competition — open to math geeks around the world — to improve the online movie-rental company’s method of predicting which films any specific customer might wish to see. Netflix, which asks customers to rate movies they rent so it can suggest future rentals, is making more than 100 million ratings from nearly a half-million customers available to anyone who wants the data. A researcher who comes up with an algorithm more accurate than Netflix’s system could win $1 million. The competition opened on October 2006 and is scheduled to continue through October 2011. There is a catch, of course: “You must share your method with (and non-exclusively license it to) Netflix, and you must describe to the world how you did it and why it works,” contest rules say.

For Whitehurst, 41, exposure to the open-source culture at Red Hat has been both educational and enlightening. A native of Columbus, Ga., he has a bachelor’s in computer science and economics from Rice University, studied at Friedrich-Alexander University in Germany and earned a general-course degree from London School of Economics and an MBA from Harvard. He was vice president and director of The Boston Consulting Group, working in Chicago, Atlanta, Hong Kong and Shanghai before joining Delta in 2002. Problems there were profoundly troublesome. (So much so that The New York Times, in a 2006 article about him, described conditions this way: “Poor pay, long hours and the contempt of many employees and customers — that pretty much sums up the job of running an airline these days.”) While many woes were unique to the industry, Whitehurst can see in hindsight that he was captive to what he calls “old-world rules” of business. “The value of collaboration was not really what any of us were schooled in,” he says. With his voice fading into wistfulness, he adds: “The ability we could have tapped into at Delta, with 50,000 employees … ”

The culture shock Whitehurst confronted at Red Hat is the same anyone can expect to endure during an initial exposure to the open-source world. One of the important underpinnings of economic theory, of course, is that scarcity equals value — which is to say, if you can control the supply of a desired commodity, you can increase its value in the market. But open-source thinking turns that belief on its head. Whitehurst describes it as “the economics of abundance rather than the economics of scarcity.” An open-source business model is built on the theory that wealth is most efficiently generated (not to mention the greater good served) through shared information rather than through protected, exclusive information.

In short, the world will always reward the building of a better mousetrap. But the market eventually will belong to the person who figures out that it’s smartest to give free mousetraps to the few fanatics who want to tinker with them and then sell upgraded versions to the rest of us.