|
|
Embedded Linux and the end of amateur hour
2003-09-10 Article Rating:  / 0
The Embedded Market Forecasters (EMF) report seems to make Microsoft look better than it is, but the report is not far wrong about the bloated costs of unmanaged embedded Linux projects. Embedded Linux is an outstanding value for well-managed projects. The software is flexible, powerful, reliable and resource-frugal.
System developers new to Linux find it hard to believe that they need so little memory and get so much functionality. Even more impressive, embedded Linux runs an enormous range of applications from Oracle to Apache to FSMLabs' own hard real-time RTLinux to Java. And users of embedded Linux don't need to worry that the vendor will insist on a new licensing policy and a new, incompatible, base operating system that needs new and more expensive hardware next year. Our own company delivers embedded Linux as part of a software package to customers who do everything from building jet engines (Pratt and Whitney) to making dairy test equipment (Motion Design). EMF's conclusions directly contradict our own and our customers' experiences. So, where did EMF go wrong? Part of the problem comes from tactics such as considering both XP and CE .NET in the development case studies but making the costly XP disappear when EMF considers royalties. EMF's table of costs for FSMLabs' RTLinux distribution -- and possibly those of other vendors -- does not provide comparable volume prices. The table implies that CE .NET is a real-time kernel, which, to put it charitably, is dubious. The table quotes the higher-cost version of our per-seat price. This makes me wonder about the validity of the report's other price comparisons. However, I believe the greatest problem with the report is in something that the data reflects correctly. Embedded Linux can be a disaster for companies that want a free magical substitute for expertise or have not correctly analyzed their value-add. The report considers this to be a technical defect of embedded Linux, but CE customers who decide to employ a neighbor's prodigal teenage son because B-Squared is so expensive and the kid did something like this in high school will see exactly the same problem. The "defect" in Embedded Linux is that because it can be downloaded for free, there is a perception that it can be integrated for free. Embedded operating systems are complex. If you throw in a requirement for hard real-time, complexity spirals up. Downloading embedded Linux cannot cure the huge costs and slow time-to-market that plague "home-made" embedded operating system solutions. In-house engineers who have other critical tasks and non-specialist consultants simply cannot compete with a solid, quality-assured, tested solution. A recent book on embedded Linux advises programmers to " . . . start with the most recent stable versions of each package and replace them one-by-one with older ones if they fail to build" to get a working set of development tools. Some managers might say, "I have plenty of programmers sitting around drawing salaries with nothing better to do than to screw around trying combinations of tools until something seems to work, and we are in no hurry for this stuff, so why should we buy a tested and pre-configured solution from FSMLabs or another vendor?" These same managers will probably decide that in-house staff or some local contractor -- who has all the qualifications needed to download software developed by someone else -- can configure their embedded operating systems. And these same managers will have their projects show up in reports about the enormous expense of embedded Linux. When used properly as part of a process of making the embedded operating system a commodity component, embedded Linux can improve system reliability and capability at an exceptionally low cost. On the other hand, a long time ago, economist Ronald Coase explained that firms need to purchase those products that they cannot manufacture with added value. Embedded Linux illustrates why Coase's work won him a Nobel Prize in Economics. Embedded Linux is a great value, but the availability of free downloads presents a risky temptation to cut corners and invest engineering without adding value. That's the reality that emerges from EMF's data. This column, which originally appeared in the August 2003 edition of RTC Magazine, is copyright © 2003, The RTC Group. Reproduced by LinuxDevices.com with permission of The RTC Group.  About the author: Dr. Victor Yodaiken, CEO and Co-Founder of FSMlabs, began his career in 1983 as one of the chief developers of Auragen's distributed fault-tolerant UNIX and he had an active consulting business before starting FSMLabs. He has also worked in academia, as a professor and department chair at New Mexico Tech, and as a research professor and port-doctoral fellow at the University of Massachussetts in Amherst. Currently he is an adjunct faculty member at the University of New Mexico. Dr. Yodaiken is a technical advisor to EMBLIX Japan and he is on the board of the Embedded Linux Consortium.Analyzing the EMF "Total Cost of Development" report These three editorials originally appeared in the Aug. 2003 issue of RTC Magazine:
Do you have comments or questions on this story, or on the EMF report? But please, before you jump into this discussion, read the full report! You can download it from the EMF website (requires free registration). Related Stories:
|
news feed
