EEPW论坛

Mr. Schacker spent a lot of time explaining what an Embedded System is, how it is defined, and how it compares to PCs. Besides being customizable, the EC (Embedded Computer) has three basic advantages over a PC: it is capable of doing real time applications, which the PC operating under a Microsoft Operating System is not, it is "down" scalable since its Operating System can be customized and therefore have a much smaller memory footprint, and it is reliable, i.e. it functions 100% of the time because peoples' lives could depend on it functioning (anti-lock breaks, phone service 911 calls.) Mr. Schacker thought that in the very near future, there would be a new model for designing and marketing devices that utilize these ECs in them. Today when the user chooses to upgrade, the device is completely replaced with a newer device with more advanced and faster technology and the outdated device is simply discarded. Tomorrow, to do an upgrade, the user will keep the device, simply download new software, and then reboot the device. This has several very positive implications. Vendors can make more sales by releasing newer, more powerful versions of their software more frequently. In addition, although Mr. Schacker didn't mention it, I particularly like the implicit environmental conservation in this new model, i.e. less hardware is being discarded because the devices are being "recycled". However, there is also a down side of this new model scheme. Mr. Schacker stated that there is a lot more to think about when designing software to be upgradable. What was not made clear, however, was that if the design effort for such devices was more complex, then it would take more time to do the design. Mr. Schacker did not address the fact that perhaps this more time consuming design might increase the time-to-market of such products, thereby eliminating the vendors' ability to release new versions more frequently which would obviously obliterate the economic advantage claimed above. There is a phenomenon that is quite remarkable about VxWorks, the Wind River's operating system. The embedded computer market has many diverse applications: industrial automation, digital imaging, military, aerospace, automotive, global communications, consumer electronics are some. Yet all these embedded systems need an operating system. Wind River can build just one software product for all these very different applications because there is a layer of commonality in their requirements. However, there is a limit to how far this can be taken. At the limit point, there has to be the ability to customize the Wind River OS for the specific requirements of each different application. This is accomplished in large part by VxWorks splitting up all the separate functions of an operating systems, i.e. Java, networking, real-time processing, etc., and offering each one as a separate feature of VxWorks. If the customer wants the feature, they get it, if not, they don't have to get it. Thus the operating system can be customized by using standard parts. Mr. Schacker pointed out that Sun MicroSystems did the opposite. For Sun, everything is in JAVA. It has a very large memory footprint (4-8 MB) as compared to VxWorks which is 1-2 MB. If enough users do not want what JAVA and Sun have to offer, this could be diastrous for Sun. Sun's strategy is a big risk, but if successful it is also a big win. Market and technology studies are predicting that the embedded market is growing at a very fast pace. What used to be simple, single function devices are now being designed to be "smart", i.e. to have embedded microprocessors which will enable these devices to do more complex jobs for the user. For example, everyday more and more devices are being connected to the internet, and are interactive, which means an embedded user interface( i.e. software) must be present. An example of this trend is the evolution of the television. In 1990 televisions needed only a 4-bit processor with 4KBs of RAM in order to perform the simple task of operating with the remote control. In the year 2000, this is going to change. Buying a television will be a very similar experience to buying a computer. Which microprocessor is inside the TV, its operating speed in MHz, how much RAM is present, the capacity and bandwidth of its embedded hardrive will all be marketing features. This growth in the embedded market has attracted the attention of MicroSoft and Sun Micro Systems. In desiring to get their fair share (or more) of the growing embedded market pie they have become new competitors for Wind Rivers. Sun bought a real time operating system company, and Microsoft is pushing WinCE. However, Mr. Schacker pointed out that getting into the embedded market for either company would still be extremely difficult because it would involve a complete overhaul and redesign of their current systems in order to cut the memory footprints in order to fit into an embedded device. Mr. Schacker asserted that strategic marketing is a highly dynamic environment. He talked about the four P's of marketing: Product Strategy, Pricing Strategy, Promotion Strategy, and Placement or Distribution Strategy. Wind River's Product Strategy is to emphasize to their customers the advantages gained by using VxWorks. If the Wind River's potential customer buys the Wind River's OS then it means that they do not have to build the entire OS themselves, the time-to-market for their product will be shorter and the OS will be highly reliable, both of which are very crucial in the high technology field. By customizing VxWorks for their own embedded device, the Wind River's customers can create their own unique product, thereby differentiating themselves from their competitors. Furthermore, Wind Rivers made a decision to put their R&D money into an enabling technology. They bought a graphics company so that they could leverage off of their software (thereby cutting their own time-to-market.) In addition, Wind Rivers made the graphic interface software available to everyone, including their own competitors, by keeping the graphics company a wholly owned subsidiary. This was a very resourceful thing to do since it insured Wind Rivers that their interface would become the standard in addition to giving them sole control over defining that standard. Also, they are making money from their competitors, which, to me, was a brilliant coo! Each time a competitor who had bought and integrated the Wind River's graphics software into their product sold that product, Wind Rivers would get a royalty from such a sale. So they had a finger in almost 100% of the market! Another very interesting point that was made was to think more in terms of solutions rather than just products. For example, systems are getting so complex that the customer, who used to be the system integrator, is becoming overwhelmed with such complexity. Thus, the system integration must be done by someone else in the technical chain, perhaps the vendor. This then prompts such a person to solve the system integration problem better. This in turn creates a big business opportunity for the vendor. Thus, by solving the system integration problem for their product, the vendor has magically and perhaps significantly increased the value of said product in the market, and reaps all the accompanying rewards. I asked Mr. Schacker if Wind Rivers considered doing advertising of their OS to target the customers of the Wind Rivers' customers, much like Intel advertises their Pentium family of chips not to their own customers, who are company's which buy the Intel chips to put into their own system products, but to the individual buyers of these systems. Mr. Schacker said that Wind Rivers made the very conscious decision not to do this kind of "brand" marketing for their OS because they felt it was too expensive, and it was a logistical challenge. They feared that TV makers wouldn't want their own clients to have the "Wind Rivers' OS" experience rather than the "Sony" (for example) experience. Of course this is just Wind Rivers' executives' opinion. I don't think they thoroughly did the market research necessary to really determine if this "logistical challenge" really existed. The strategy works very well for Intel, and I would imagine it could work just as well for Wind Rivers. If it was anywhere near as successful as Intel's efforts have been, the expense would justify itself with healthy profits. It seems that Wind Rivers has a head start in the real time OS for ECs market, and they should solidify and take advantage of this lead by making the end customers aware of what OS they are experiencing. I think it is a serious mistake that they are not pursuing this avenue of "brand" advertising. The only time the expense of such advertising would not be justified would be if Wind Rivers had a monopoly. Then, their sales would probably not be significantly affected by such advertising since it would be the only available embedded OS anyway! But Wind Rivers does not have a monopoly in the market and they are not taking advantage of the lead that they do now enjoy. (They could really exploit the fact that their OS was in the Pathfinder. People know what a success that project was, and if they knew that the same OS was in the device they were about to buy, most people would choose the device with the "best", i.e. the Wind Rivers', OS over another device with some other unknown, untested OS.) Along these same lines, however, Wind Rivers does advertise their "partner programs" as much as possible. Mr. Schacker called this "leveraging the industry". He claimed that the partners were interested in letting people know that they are partners with Wind Rivers. This advertises Wind Rivers for free! Works for me. Another advantage for Wind Rivers of having partners is that the partners increase their penetration into the market. Mr. Schacker called it "seeding" the market. The Wind Rivers' partners send their own customers directly to Wind Rivers for anything else those customers may need that goes beyond the partners' product scope. This, of course, is a great distribution strategy since there is an implicit recommendation for Wind Rivers with each referral. Finally, when discussing their pricing strategy, Mr. Schacker mentioned the difficulty that Wind Rivers has in convincing their potential customers that the price of the OS software should be valued well over the comparable cost of the man power to build an "equivalent" OS. It is obvious to me that it is not the actual building of the software, but the copious testing that is necessary to determine if it is correct. This seems especially true of a real time OS that has to be correct 100% of the time. On the other hand, if Wind Rivers had a large enough sales volume, it is possible that they could make a healthy profit without charging such a healthy price to each and every customer! Mr. Schacker did not discuss this point when he discussed the Wind Rivers' pricing strategy. Perhaps if Wind Rivers reconsidered their position on "brand" advertising, their volume would grow to such a level that they could actually sell their OS for less than the "man power hours" cost! Send email to: randit@cs.berkeley.edu