How Much to Build a Linux Distro? Try $10.8 Billion [4]
The Linux Foundation today announced it is publishing a new report written by Amanda McPherson, Brian Proffitt, and Ron Hale-Evans on the value of Linux development. The paper, "Estimating the Total Development Cost of a Linux Distribution [5]," finds that it would take approximately $10.8 billion to build the Linux community distribution Fedora 9 in today's dollars with today's software development costs. It would take $1.4 billion to develop the Linux kernel alone.
This report is an update of a 2002 study done by David A. Wheeler that examined the Software Lines of Code (SLOC) present in a typical Linux distribution (Red Hat Linux 7.1). At that time, Wheeler found that it would cost over $1.2 billion to develop a Linux distribution by conventional proprietary means in the U.S.
The authors examined the Fedora 9 distribution using Wheeler's tools and methods, specifically the SLOCCount tool that estimates value and effort of software development based on the COnstructive COst MOdel (COCOMO). The report goes into detail on the methods used, how they specifically apply to the Fedora distribution and the Linux kernel, and what an estimate of Linux's value really means.
Highlights of the paper include:
- How Much Does a Full Distribution Cost?
Using 2008 salary figures, the tests published in the paper revealed that if developed today, the full set of Fedora 9 distribution packages would cost $10.8 billion. The Fedora 9 distribution contains 204.5 million lines of code in 5547 application packages. The development effort estimate comes close to 60,000 Person-Years. - How Much Does the Linux Kernel Cost?
Applying this test to the Linux kernel included in Fedora 9 found the value to be 6.8 million lines of code worth $1.4 billion. The development effort estimate for the kernel alone exceeds 7500 Person-Years. - How Does This Really Measure the Value of Linux?
This study reveals that collaborative development creates enormous economic value. In the past two years alone, over 3,200 developers from 200 companies have contributed to the kernel. An even larger number has contributed to full Linux distributions. Measuring the economic effort involved is imperfect, but this report clarifies why the methodology is the best approach and some of the limitations.
"This year has seen an incredible proliferation of Linux-powered devices outside of traditional Linux strongholds: devices powered by the Moblin [6] platform, netbooks like the eeePC, mobile phones like Android and the Gphone, and consumer devices like the Amazon Kindle. Would these products be possible without Linux?" said McPherson. "I think this points to the power of the collaborative development model. Monopolistic software companies used to be able to fund heavy R&D budgets, keeping out competition. Given the cost associated with building an OS like Linux one wonders if proprietary companies will ever go it alone again."
The Compelling Economics of Linux: What It Means for the Future of Computing [7]
The Economics
Today the Linux Foundation issued a report looking at the value of the Linux platform in terms of code. This was an update of a 2002 study that estimated the value then at $1.2 Billion. Today’s value: $10.8 Billion. The study focused on the Fedora project, which has been a core part of Linux success in the server and desktop market place. Although it wasn’t specifically covered in this paper it is also worth applying the economics of Linux to one of the fastest growing segments of technology; mobile devices, consumer electronics and low cost netbooks. This is the future of Linux and the smart bets are leveraging a $10.8 billion investment to the hilt.
Linux is Everywhere
I am constantly amazed by how rare it is to work with any consumer electronics (CE) device that does *not* run on Linux. Other then two big markets — laptops and mobile phones, nearly every new consumer electronics device runs Linux. Sony televisions, Amazon Kindle, Dash automotive GPS, and nearly every other device you can imagine.
A CE company can either try to roll their own operating system, license a proprietary one like Windows or VXWorks, or use Linux. The reasons they use Linux are simple. It is easiest to hire people familiar with it. It supports more devices than any operating system in the history of the world. It is completely open, so if something doesn’t work, you can fix it yourself or pay someone to do it. There is amazingly great support available from mailing lists, or commercial support available at any service and price point. You can brand the device however you want. And it gives you a real Internet experience, with the capability to do any level of networking and application support.
One to Watch: Moblin
The final two frontiers for Linux in consumer electronics are mobile phones and laptops. I’d like to congratulate Google on shipping their first Linux-based phone this week. This is a great accomplishment, and Android should prove to be a major competitor in building a mobile phone ecosystem.
Another consumer electronics project I’m excited about is Moblin. Though initially focused on NetBooks (i.e., small laptops), I see Moblin as creating the ideal platform for a large universe of devices from MIDs to in car entertainment and more. Unlike Android, which uses Linux at the base but rewrote most of the upper level software, Moblin leverages the enormously valuable work of the entire Linux ecosystem (that $10.8 billion). But they do this while working to fix the small bugs and incompatibilities that can still cause frustrations in desktop Linux. And by working within the Linux ecosystem, the improvements they are making to a whole array of different packages and libraries will be passed back to the upstream authors, so that all Linux users can take advantage of them and adding even more value to that multi-billion dollar pie.
In a couple years, I expect Moblin to be playing the role of a standard platform for netbooks, MIDs, consumer electronics, and more. Already there is an incredible ecosystem around the platform with hundreds of ISV’s, dozens of hardware OEMs, and many Linux operating system vendors on board. Given the compelling economics of their approach I think it will be harder and harder to find devices that don’t use it in the future.
George Gilder wrote an enormously influential article in 1993 titled Into the Fibersphere. He stated:
” As industry guru Andrew Rappaport has pointed out, electronic designers now treat transistors as virtually free. Indeed, on memory chips, they cost some 400 millionths of a cent. To waste time or battery power or radio frequencies may be culpable acts, but to waste transistors is the essence of thrift. Today you use millions of them slightly to enhance your TV picture or to play a game of solitaire or to fax Doonsbury to Grandma. If you do not use transistors in your cars, your offices, your telephone systems, your design centers, your factories, your farm gear, or your missiles, you go out of business. If you don’t waste transistors, your cost structure will cripple you. Your product will be either too expensive, too slow, too late, or too low in quality.”
The same is becoming true with Linux, and for one of the fastest growing segments of computing, the project to watch is Moblin.