Mark, you keep mentioning RSYS now QNX.
RSYS is not a direct competitor to WIND; it is mainly in the hardware business. If it deserves investor attention, start up a thread on the company. But this thread is the WIND embedded RTOS system thread, so please get off the RSYS kick. Besides, you have a consistent pattern of completely misrepresenting facts about RSYS. For example, in Post #1849 you said,
>"Radisys was founded by former Intel employees who designed
> the i960 processor. They bought Intel's embedded systems
> business and Intel received 17% of the company. Radisys
>received the iRMX operating system as part of the deal.
> I2O is based on the i960RD processor, a derivative of the
>i960 that adds PCI bus capability. They could've combined
> their i960 and iRMX expertise to get the I2O contract, but
> they decided there were better opportunities elsewhere.
>Radisys has the privilege as an Intel company of picking
> the low hanging fruit from the Intel tree and Wind River
>must be content with unripened fruit."
This statement contravenes every fact known about I2O. Intel and WIND announced the I2O deal in January/February 1996. At the time of the announcement it was evident WIND had been working on specifying and developing IxWorks for probably at least the preceding year, i.e. all of 1995. Intel didn't sell the division containing iRMX to Radisys until mid-1996. How was it possible that Radisys could even be thinking about employing "their" iRMX to the I2O Contract when, at the time the deal went down, iRMX wasn't even in their vision?
I suspect that Intel gave consideration to using iRMX (when they still owned it), but quickly decided against it because they must have realized that I2O needed an accompanying toolset that is the most capable and most popular and on the most processors. (Intel has always known that I2O must spread to other processors if it is to become a truly open standard.) No doubt that led them to Tornado and VxWorks as a starting point. The other possible explanation for Intel's action is that they realized that the I2O OS needed some heavy lifting, both for specifications and development, so they simply went to the best. Radisys never had the privilege of picking I2O off the Intel tree.
It seems plausible that after Intel worked successfully with WIND on I2O, and after noting what RTOS are most popular on their i960 processors generally, they decided to dump the embedded systems division to Radisys. You might call it low-hanging fruit, but I call it rotting remains of fallen fruit.
Lest there be any confusion. I have no opinion about RSYS. It may be the best buy in the stock market. But whether it is or is not, has nothing much to do with WIND.
QNX is at least relevant to the thread, and I appreciate your referenced link to the article interviewing Dan Dodge, a QNX co-founder. (Isn't it a coincidence that Dan Dodge, from DEC, was a fairly frequent contributor to this thread. How many Dan Dodge's are there?). The problem is that you never put QNX into context.
QNX is a private company that has been around for years, as a DOS cum Windows alternative on the PC. QNX, like Lynx, is Unix-like but real time. QNX began life by targeting networked PCs as a lean, cheap alternative to DOS and Unix, then as an alternative to Windows NT. As a full-fledge replacement for a desktop OS, QNX traditionally was not directly competitive with WIND. QNX apparently made some headway in process control, but otherwise has seen muted growth. Recently, QNX apparently has been making hay by embedding itself on x86-based boards, and attracting a wider audience. Essentially, Radisys and QNX are targeting the same applications, but from a hardware versus strictly software perspective. (Nevertheless, they compete since Radisys emphasizes Windows NT/ with a kludge real-time add-on using iRMX).
I think you should apply your x86 wisdom to convincing Radisys to acquire QNX.
Based on the QNX article you seem to take Dan Dodge at face value when he says that signed deals will translate into a $100 million business within two years, enabling QNX to overtake WIND. First of all, signed deals often do not translate into anything, because only products that meet with market success generate hoped for royalties. That's why WIND never boasts about future royalties - because the outcome is not in their control. Second, if QNX has signed deals theoretically worth $100 million in two years, how many signed deals potentially worth how many $million of dollars do you think WIND has? Hint: WIND has over 13,000 users today trying to get new products to market. I would bet any amount of money that if you added up all the "signed deals" WINDs customers are working on, they would add up to many $billions of dollars of royalties revenues. Sadly, many of those products will not get to market, just as with QNX, and so those $billions of dollars of potential royalties will be reduced to a mere many hundreds of $millions.
QNXs Dan Dodge was able to make those projections because QNX is a private company, so he can say whatever he wants. The projections may even be achieved. But even if they are, they will not bring him any closer to passing the market leader, WIND.
Allen |
