Dedicated Systems Experts, the publisher of this magazine, continued its RTOS evaluation program activities by tackling Windows CE 3.0 from Microsoft Corporation. Given the market’s interest in this new product, we were eager to apply our real-time performance test suite to it and see for ourselves what it is worth. The results were as expected: Microsoft got it right this time. Windows CE 3.0 includes all the features necessary to accommodate good real-time system design. Fundamental flaws found in CE 2.0 like memory leaks, an oversimplified threading scheme and non pre-emptive interrupt handling were all taken care of. This doesn’t mean it has become the holy grail of RTOS, but neither have its competitors. One thing is for sure, Windows CE 3.0 can hold its ground.

Our market research indicates that Microsoft and WindRiver Systems will be the major players of the future, equally sharing about 45% of the market, while the other commercial RTOS (more than a hundred of them) share the remainder.

But how will these companies capitalize on their product’s lead? After all, revenues yielded by (run-time) license fees are expected to diminish in years to come. I believe the “license free” business model is the one of the future, and that the RTOS vendors will be coerced into adopting it. We’ve observed the Linux community setting this trend. QNX is currently exploring the middle ground by letting users freely download the platform for non-commercial use, while still charging license fees for production. WindRiver is also looking for alternate sources of revenue, and is now offering services to help its customers with their software development projects. But what will Microsoft do? So far they are still bucking the trend by focussing on production only, while relying heavily on third-party vendors for distribution and providing value-added services.  It is difficult, if not impossible, to predict which business model will prevail. The economic downturn we are currently undergoing might provide clarity in the near future.

A troubling observation we made during our continued market studies, is that roughly 50% of the developers using a particular RTOS are unsatisfied with it, and would be willing to switch to a competitor. Needless to say, every product has its shortcomings, but not to a degree that can explain these rather disturbing findings. In similar studies carried out on case tools, which are arguably less mature products than RTOS, the results obtained were far less dramatic: only 10% of the case tool users are unconvinced of their product’s qualities and are considering using another tool. Perhaps education is an explanation: from experience in the field we have learned that engineers embarking on a real-time project for the first time, often don’t really understand what real-time design is all about. They think the RTOS will solve all their (design) problems, much to their disappointment when in retrospect, they find out it didn’t.

Dr Martin Timmerman
Contact

Dedicated Systems Magazine in general