Real-Time Operating Systems Market Review

As mentioned in my editorial of the RTOS II issue, you will find here some detailed results of our RTOS market review carried out from June till September 1995.


By Dr. Martin Timmerman,
Chief-Editor,
Real-Time Magazine.


Introduction
First we should remember that this market review was carried out to poll for the interest of our readers in our RTOS Benchmark Project. Simultaneously we asked them which RTOS they are actually using, how satisfied they were about it, and which one they want to be benchmarked in the first, second and third place.
We observed by looking to all the details that a lot of RTOS manufacturers or distributors returned the questionnaire. The preliminary figures I gave in the above mentioned editorial took into account these forms. In the figures I give you here we excluded the RTOS manufacturers and distributors. Actually only real" users are listed.
The review is about all possible RTOS applications, from small embedded ones like GSM telephones, towards large VMEbus based systems.

Figure 1. RTOS Used (1st, 2nd, 3rd place) in %.
Figure 2. RTOS Used (Total) in %.

RTOS Use
Some companies are using more than one RTOS depending on the type and size of applications. Figure 1 gives the percentages for each RTOS they use in the 1st, 2nd and 3rd place. In this case OSE is the winner with 16.3 %. This is less than the announced 21 % in the previous issue due to the fact that the OSE distributors seem to be very active to promote their own stuff. What is amazing to see is that if people use OSE, they put it in the first place. OSE is not well indicated for use in the second and third place. VxWorks seems not to have a well defined privilege position as 1st, 2nd or 3rd RTOS neither does the distributors promote the product by answering the questionnaire. As a consequence, VxWorks is the winner if we make no difference between the 3 votes. They have then a market share of 25% (see Figure 2).
We also asked how satisfied the users were about their RTOS and if they wanted to change it in the future. For all RTOS together we found that 50% wanted to change in the future, 39% will stick to the one they use for the moment and 11% was undecided. This clearly shows that the RTOS market is not yet stabilised. There are different reasons why they want to change but the one we saw the most is due to the lack of one or another processor support. This means that people still decide first on the processor they want to use in an application and then on the RTOS.
The "want to change", "no change", "don't know" rate of 50,39, 11 is fairly the same for all RTOS mentioned in the figures 1 and 2 except for VRTX were there seems to be a generalised dissatisfaction for reasons as

  • too expensive
  • poor performance and debug tools
  • need for better UNIX support
  • they only keep it because it takes too much time to change the application program

We also observed that people totally disagree on what a good or bad RTOS is. For example for the same product such as OSE, 85% consider it as excellent or good and 15% as mediocre.

RTOS to Benchmark
We then asked people which RTOS they wanted us to benchmark in the first, second and third place. Figure 3 gives the results. 2 0Ss are asked for here which were not in the "RTOS USE" list: Solaris and AIX. This could mean that some people are now considering the more "heavier" RTOS for some type of applications. Again OSE is more asked for in the first place compared to the second and third place. Again VxWorks is the winner with 20% (Figure 4) if we make no "place" difference.
One product (PSOS) is perceived differently in the RTOS USE and RTOS to Benchmark questions. This could indicate a possible increase in interest in the PSOS product.

The Benchrnark Project
Finally we were interested how our Benchmark project was perceived. There we have bad news. Most people were satisfied with the executive summaries in the magazines and did not show interest in the full reports, mainly for price reasons.

Figure 3. RTOS to benchmark (1st, 2nd, 3rd place) in %.
Figure 4. RTOS to benchmark (Total) in %.

The answers tell us that they are only interested in such reports the day they want to change their RTOS. They are not interested in a subscription with continues updates.
Therefore we decided to go along with the project, but in a less extensive way and to produce a yearly report on the most popular RTOS which we consider to be OSE, LynxOS, VxWorks, 0S9, VRTX, PSOS, Solaris and AIX.

Some Personal Conclusions
I went through all the reply forms myself and was rather upset. There are too much products on the market with low interest so a lot of engineering effort is wasted. No user seem to be happy, they all complain about something. We hope that the Benchmark Project in its new form will help the end user in better defining what a good RTOS is. We also hope that the manufacturers will use the Benchmark Project information to produce better quality products.

Technologies, Methods, Tools, Products and Services for       Embedded Systems To Be      www.es2.be
© 2007 Dedicated Systems All Rights Reserved   Privacy statement.