Right now the majority of the tech world is focusing on the carefully-scripted events taking place at CES in Las Vegas, but personally I’ve found that event to be typically a bit underwhelming. The glitz is fun to watch, but in general the stuff that is shown there is stuff that I have already seen the basic technology behind or even, in some cases, been involved in the development. That sounds elitist, but it’s actually true for anybody that follows the underlying technologies upon which these devices are built.

Given that electronic toys depend on the capability of the underlying electronics, someone who really wants to see what the future is capable of holding should check out the latest version of the International Technology Roadmap for Semiconductors (ITRS) report from the Semiconductor Industry Association. This report is a bit dryer than the typical press release, but it does lay out the expected and potential paths for semiconductors over the next decade and a half. Is it accurate? I don’t know. Personally, I never prognosticate further than five years into the future, because by then the random events tend to overwhelm the discernable trends, but the SIA is to be commended for collecting this amount of information and making it readily available. Warning: It’s not light reading. The Executive Summary alone is a hundred pages. What I have seen of it so far seems worth reading, though. This is highly recommended for those interested in the future of the industry.

Larry Mittag

It seems like someone by now would have figured out how to go about doing very large projects. The historic way of doing them is to break it into a few steps (analysis, design, implementation, and rollout), throw a heap of bodies at it, and try to manage the process. The weaknesses in this so-called waterfall model are well-known, but that doesn’t prevent it from being a popular alternative for many systems. This is particularly true in embedded systems, where conservative management practices have been reluctant to try any of the newfangled stuff coming out of the mainstream IT world.

That may be changing. A recent article on the Embedded Systems Programming site trumpeted the advantages of Extreme Programming (XP), which was quite the buzzword a year or two ago. The authors show the same unabashed enthusiasm as any True Converts to a new religion, but I’m not so sure their excitement is warranted.

Granted, there are some good ideas in XP. The concepts of iterative development and increased customer involvement can be very helpful. Unfortunately, many of the XP projects I have seen have used the techniques as an excuse for undisciplined development. XP done correctly is not free-for-all hacking. It requires detailed work to ensure that the iterations are converging rather than diverging. This discipline requires training and management oversight, which is too-often missing.

The sad thing is that when a project fails due to these factors XP usually catches the blame. There are good things and bad things about the techniques, but they can’t be fairly evaluated unless the techniques are actually followed.

Larry Mittag

It is quite common to read white papers from vendors that purport to provide unbiased information. It’s much rarer to find one that actually provides that information without attaching a commercial message to it. I ran across one on the WindowsforDevices.com site that does just that. This paper from Datalight provides a clear, concise summary of the differences between NOR and NAND flash memory that provides useful information to embedded systems designers.

As I said, this type of information is rare. I recently reviewed abstracts submitted to the Embedded Systems Conference, so an example of a non-marketing message is a welcome relief. It’s amazing how many companies think that that conference has it in for them because their papers get rejected time after time. The truth is that most of these are such bald advertorials that it’s almost funny. Granted, many technical conferences accept such papers readily, but it’s nice to be associated with a group of people that still thinks providing good information instead of a thinly-veiled commercial is important.

I have to wonder if people realize this. When you attend conferences and see marketing slides instead of hard information does it bother you?

Larry Mittag

PC programmers are beginning to catch on to the changes that are taking place in CPU architectures. A recent talk given at In-Stat/MDR's Fall Processor Forum was reported on by Computerworld. The speaker was Herb Sutter, a Software Architect from Microsoft who felt that CPU designers did not understand the plight of software engineers that were ill-prepared for multicore CPUs that are coming down the road. The tone of the report seemed to be that chip designers simply needed to understand that they were not serving the needs of programmers by following this path.

What the speaker apparently failed to understand is the level of frustration that has been building in the semiconductor design community. The sacred PC architecture has been the bedrock of software progress, but it also has been a millstone around the neck of hardware engineers that want to be more creative. I have spoken with more than one designer that is much more interested in FPGA-based application designs than creating yet another incarnation of a 20-year old platform to feed the needs of programmers. More than a few of them are talking directly to customers and building solutions that are breaking away from the traditional generic hardware / specific software model that has been the norm for the last few decades. If they can deliver cost-effective value in embedded applications by circumventing software, then all the better for them.

In the PC world it is a little more complex. CPUs simply cannot make progress the way they have in the past, so the alternative is to multiply the CPU resource. This will challenge software engineers, separating the ones that can adapt from those that can’t. The ones that can most effectively create software using the new rules will be able to demand a premium, whether they are in Indiana or India. This is not a good or a bad thing, simply a fact of life.

Larry Mittag

Robotics is in the news lately with the latest instantiation of the DARPA Grand Challenge, as pictured on C/Net. This is a 150-mile race that will take place somewhere in the southwest US – the course will be announced two hours before the race begins – and that consists entirely of autonomous robotic vehicles. Last year was something of a bust, given that no car managed to finish the grueling cross-country course, but the whole idea of this is to give the technology a chance to improve. Maybe this year someone will actually make it to the finish line.

Meanwhile, on the other side of the world they are doing their own versions of robotic races. It seems the sport of camel racing was suffering under the newly-enforced child labor laws in the United Arab Emirates. The best jockeys for these camels are children that weigh under 60 pounds or so, and “recruitment�? of these jockeys sometimes consisted of kidnapping them from poorer areas of Asia.

The solution (as reported by the BBC) was the creation of a robotic remote-controlled jockey that fits the weight requirements, doesn’t spook the camels, and can work in the heat. It’s an interesting application, but this is one of those stories that really makes you wonder whether some people just simply have too much money.

Larry Mittag

There are a lot of things happening in CPU design these days. The CPU selections for the next generation of game machines are old news, as is Apple’s conversion to the Intel camp. Now it appears that Intel is finally getting around to trying to kill off their XScale line as they bring out x86 models that finally don’t double as toaster ovens. A report on The Register talks about their plans to converge their architectures in future generations of low-power SoC’s.

This makes sense for Intel if they can really pull it off. They acquired the XScale line when they and Compaq split up DEC, and from what I heard they literally had no idea they were getting it or what they had. While they were trying to figure out what to do with it the design team got lifted out from under them, so there was real question whether they would just drop it. Instead they ran with it, and for quite a while that part of Intel was where the interesting work was being done. All of that has to be irritating Intel’s management, especially when they write royalty checks to ARM.

The questions are, first of all, whether they can pull the technology off. ARM is not just a low-power architecture, it is also fully static and is enjoying an unprecedented amount of support in embedded systems. Even the Single-Board Computer (SBC) vendors are forsaking their legacy PC-based designs and coming out with ARM (especially XScale)-based boards. One of them told me they are trying to make hay before this architecture also becomes commoditized.

The second question is whether designers will accept an x86 in embedded designs, even a low-power one. The major problem besides power has always been that x86 CPUs were too expensive. The only way Intel may be able to garner support in embedded designs may be to cut their profit margins, which may make the entire exercise problematic for them.

Larry Mittag

Embedded systems are one thing, but it takes on a whole new meaning when the systems are embedded in us. An EE Times story discusses the advent of the so-called Bionic Man, where electronics are directly interfaced with human nervous systems to either replace or enhance the natural abilities.

This is a familiar staple of science fiction, but there is quite a bit of current science fact as well. I am fairly familiar with the cochlear implants mentioned in the story, since a number of my wife’s relatives are deaf. One interesting point not brought out in the EE Times piece is the amount of resistance to this technology among the deaf community. Their argument is that being deaf is not a disease, and therefore does not need to be cured. In essence they are defining deafness as a natural state of being, one that creates a valid subculture.

This is a fascinating twist on the Frankenstein mythology. Is it a crime against nature to revert someone to the norm? If we have the technology to “cure�? this condition, is it rather a crime not to do so if the person desires it? If we accept this modification, then is it acceptable to use the same technology to enhance beyond the norm?

These are the kind of knotty questions that tend to come up when we contemplate modifications to the devices that are ourselves. In one respect we are simply meat machines that are as hackable as anything else. From another point of view such ideas are blasphemy. Personally, I would love to jump to the end result of a neural interface and be able to jack into cyberspace. If nothing else, maybe I would get to wear one of those cool black leather coats.

Larry Mittag

Intel is taking yet another shot at becoming the de facto home media device with their recently-announced Viiv campaign. This has gained new importance for them with all of the attention on the Xbox 360 / Playstation 3 contest to define the next home game machine, which is widely seen as the new generation of the home PC. The fact that both of these contenders are based on PowerPC architectures is surely not lost on Intel.

The press releases are predictably upbeat, but there is a fairly insightful analysis on the Register. They bring up Intel’s dependence on Microsoft’s Media Center Edition, which personally I haven’t been that impressed with. This leaves Intel dependent on a partner that is also betting on another hand – a weak position at best.

It is possible that the Viiv architecture will allow the creation of a series of creative PC-based embedded devices. The emphasis on lower power will certainly make it more attractive than it currently is, but the question is whether the profit margins will be at the levels that Intel demands. The x86 architecture is not as special as it once was. Is Intel thrashing here or are they really defining the future?

Larry Mittag

The news on the breathless wait for the new Xbox 360 and Playstation 3 hasn’t been all that interesting for those of us that don’t partake much of game machines, but we can finally get some solid information on the new multicore Cell processor as promised from IBM. The release consists of a ton of information on the operation of the processor, as well as an instruction-level simulator and a compiler.

I have just started to dig into this, but it shows a lot of promise. There are some C++ extensions, although parallelization of code is going to be a manual process in this release of the tools. I’ve found myself wondering if the tools will be applicable to the multicore PowerPC architecture of the upcoming Xbox 360 as well, although I suspect they will not. They may create single-threaded code for that CPU, but the unique architecture of the Cell is more than likely the only target that can really take advantage of these extensions.

The degree of customization that can be supported with variants of the Cell could make this a very interesting architecture for relatively high-volume applications. The question is whether the pricing will be attractive for lower-price applications like consumer devices. Given that the first application is a home game machine, it might be possible, but that could be an anomaly.

Larry Mittag

The Hot Chips conference is in full swing, and as expected the emphasis is on multiple-core processors. A report in EE Times details how, as I expected, the puny multiprocessing currently available from Intel and AMD is just the tip of the iceberg. The dam that has held back hardware design is finally starting to burst and it will have serious ramifications regarding the way we create software.

Just today I got an email from a software guy wanting a characterization of the impact of not having L2 cache on a system on the CPU speed. Given that the answer depended on software that had not yet been written and was a delicate balance of cache, pipelining, and the nature of the software itself I had to explain that the answer was indeterminate. How am I going to explain it when they ask me about the impact of parallel hardware architectures that are completely outside of his (and for that matter, my) experience base?

Nick Tredennick was quoted in the article as saying that the microprocessor has held back innovation in hardware design. He said something very similar to me in a panel half a dozen years ago. One way or another, it appears that the box is being opened and that we are about to be treated to a number of novel architectures.

Larry Mittag

It’s always fun when a project you worked on gets unveiled to the world. A few years ago I consulted to Wynn Resorts regarding the new resort they were building. The CIO for Wynn’s operation was one of the sharpest I had (or have) ever met, and the infrastructure she was envisioning was extremely impressive. We worked with them on the wireless infrastructure, mostly in terms of fine-tuning access point placement and some brainstorming on where technology was going. In the process we got a fascinating look into the operations of a Vegas casino.

Now the rest of the world can see what they have built, and it is just as impressive in reality now as it was in concept then. A report in Wired provides more detail on the technical infrastructure than most stories will and is a good read. Even years ago CIO Karen Bozich knew that VOIP was going to be a winner, that HDTV was real, and that packet switched Ethernet was the backbone delivery vehicle to bet on.

Congratulations, Karen. I hope Mr. Wynn realizes how well you did your job.

Larry Mittag

The CPUs and other high-end components in embedded systems get most of the attention, but sometimes the story is in the simpler components. One of the simplest is the humble LED, but this versatile element of most systems is moving to a more prominent position in a number of applications.

A story on the website of Technology Review details the advances in LED and OLED technology that are in the process of finally burying the incandescent bulb. Anyone who drives a car has probably notices either the new traffic light bulbs or the stoplights on the back end of newer high-end cars. Before long LEDs will also show up on the front end of those same cars in the form of headlights. The technology will also provide much more efficient lighting in our homes in the form of lighting panels that are as cool as fluorescents but not as harsh in terms of light. In fact, the white panels will include small embedded systems that will account for the uneven wear of the colored components that sum to white light, keeping a consistent color to the light.

This change will not get the press that Intel got this week with their new CPU designs, but it will almost certainly save more energy and impact more people. Sometimes it’s the small changes that make the most difference.

Larry Mittag

The latest round of spectrum allocations from the FCC seems to have an air of unreality about it. For example, what is the point of allocating space at 60 GHz for wireless LAN applications? Isn’t it obvious that the prime real estate is much closer to the 2.4 and 5 GHz bands that are currently in use?

Well, maybe not. A recent advance from IBM brings RF chips into the realm of commercial possibility that can handle those frequencies. These are not high-end, expensive chips, either. These are chips that can be produced in commercial quantities at consumer price points. Rather than silicon, these are based around silicon germanium and can handle frequencies well above 100 GHz, according to the EE Times article.

Granted, these are support chips, not CPUs. Maybe you could build a 100 GHz CPU out of SiGe, but there is no way you could build a memory subsystem to keep up with it. These chips will enable radar systems for automobiles and build networks with short enough wavelengths to communicate with nanotechnological devices that were described in another EE Times article as one of five deep R&D projects that must succeed. Other projects in the list will tie these devices even tighter into a global computer network that will handle projects we can only dream about now.

Maybe innovation and basic R&D isn’t in quite as bad a shape as we thought…

Larry Mittag

The DEFCON conference has created a number of stories regarding security over the last week or two. The much-publicized Cisco router weakness has been pretty much played to death, but RFID also took its hits regarding the range for the RF communications, a Gartner report revealed that many ATMs don’t take advantage of security codes to make card access more secure, and even our cars may be susceptible to Bluetooth attacks (although they didn’t manage to cause any damage in that case). Are we on the verge of a security-related meltdown of network-connected devices?

We went through a similar round of hysteria with the Y2K problem. I gave a speech in 1999 where I listed all of the things that could happen if embedded systems were disabled by Y2K problems. Once I had their attention I explained why it was not going to happen. Once the press gets hold of this type of story there is a real tendency for the signal-to-noise ratio to get quite low. There is some evidence that this is happening here, although the gain is not set nearly as high as it was in 1999.

Consider the “virus attacks your car through Bluetooth�? scenario. The inference is that a hacker can gain access to the entire vehicular network through this network interface. In reality, with current generations the most they might be able to do is listen in on your cell calls or talk to you in your car. The current level of Bluetooth integration is simply speakerphone implementation. Control networks are strictly segregated from anything externally accessible in cars.

Yes, there are security vulnerabilities in connected embedded systems, but generally these devices are much less complex (and therefore less vulnerable) than general-purpose computers. The antivirus companies have tried mightily to drum up a threat to cell phones or PDAs, but so far they haven’t come up with much. Maybe they need to get more educational material out there to the virus writers so that they can generate them a bigger market.

Larry Mittag

The idea of a sentient computer is a staple of science fiction. Some of the more famous ones such as HAL have been engineered single computers with intentionally programmed self awareness, but in Robert Heinlein’s “The Moon is a Harsh Mistress�? and more recently in the Terminator movies the concept was that of intelligence as emergent behavior from a sufficiently-complex computer network.

I hadn’t thought about that concept for a while, until I read an article in the latest edition of Wired magazine. The article talks about the Internet as a turning point in history, the evolution of mankind as an organism rather than as a species. The end of the article speculates on the emergence of intelligent behavior as a result of the directed activity and stored information on an evolving Internet.

Does this mean that I am expecting the next Governor of California to show up in a sizzling ball of light and start trying to change the future to ensure the domination of humanity? Hardly. But there is the fascinating possibility that emergent behavior of some kind may begin to occur as networks of sensors and devices respond to coordinated stimuli. In a very real way we may be creating the next generation of something.

Larry Mittag

I was playing with my 4-year-old nephew the other day and began thinking about how different his life was going to be that mine. He will grow up never having known a world without the Internet, which teenagers today are already integrating completely into their lives according to CNN. They will also know a world that has video surveillance that makes 1984 (the book, not the year) look positively amateurish. Computers will integrate much more fully into the environment rather than being something that sits alone on a desk.

I must admit I am somewhat jealous. My nephew will probably see a moon landing that means more than just a stunt, possibly even a landing on Mars. From the looks of things computer architectures will be very different as well, especially by the time he gets old enough to care about such things.

The world does not completely belong to the young quite yet, though. I once spoke with a sixteen-year old kid at a wireless SIG that was quite sure he had everything figured out. He was a consultant to Motorola when he was not going to high school, and he patiently explained to the old guy (that would be me) how current cell phones were inherently superior because they were digital instead of analog. While he was talking to me he was also text-messaging his friends, since I was obviously not worthy of his complete attention. I then explained to him the analog underpinnings of his digital phone. After a few minutes he put his phone away and gave me his full attention. Half an hour later he realized that he was really learning something.

Someday things will be changed so much that we will no longer be relevant, but for the moment at least we can still teach them a thing or two�

Larry Mittag

The Hot Chip conference this year sounds more interesting than it has been for quite a while. This EE Times report lists a number of interesting technologies revolving around multiprocessor architectures. The multicore offerings from Intel and AMD are pretty much old news here, where the buzz is about massively multiprocessor on-chip systems arranged in a computing fabric. These systems are, as the article says, "... capable of enormous aggregate bandwidths and computing rates but are dependent on the ability of programmers to find sufficiently parallelizable algorithms, map them onto processing elements and then map them onto the processor fabric". Sounds like a good weekend project for someone...

Programmers have made a lot of progress over the last few years thanks to the stability of the PC and other architectures which allowed them to concentrate on improving OO techniques while the hardware engineers were snoozing. Those hardware engineers are awake now, and I see a parade of frustratingly innovative hardware in the offing. Software engineers are going to have to figure out how to take advantage of these new options, and more than likely embedded engineers will be among the first ones in line.

Larry Mittag

I never thought much of System-on-Chip (SoC) when it was first being discussed a few years ago. It seemed to me to be just another level of integration for hardware, just another way to make systems hard to debug by hiding the signals deeper in silicon. That opinion is changing, though, as I read articles like a recent one from EE Times on design abstraction for SoC systems.

At one level, this reminds me of the old joke about the MIT grad student who was assigned an application program to write by his professor. After a few weeks the prof asked how it was going. "Great!", came the reply. "I am almost done designing the language to write it in!".

It is certainly true that a lot of time can be spent reinventing the wheel with design options like this, but it also gives us a way to experiment with new ideas for design. As I have mentioned before, I think we need to break out of the current design box and find new ways to design systems rather than hardware or software.

Larry Mittag

When everyone accepts something to be true it's usually time to take another look at it. This is certainly true when it comes to CPU architecture. Ron Wilson's latest column reexamines the great CISC vs. RISC architecture debate, especially regarding the CPU�s relationship to memory.

We have been in a stagnant period regarding advances in computer architecture for a decade or more now. A lot has gotten done in that period due in part to the relative stability of that architecture, but the cracks are getting too big to ignore. Reexamining the CPU architecture is a good start, but the advent of SoC and FPGA-based macroarchitectures are an excellent opportunity to experiment a bit with more radical departures. The PC is capable of doing a lot of good things, but it is boring. Is there a better way?

Larry Mittag

Embedded systems are not just about processors and software. Recent advances in the more prosaic technologies of fuel cells and OLEDs have the potential to affect our system designs just as much as the newest hot CPU.

Power is always a problem, or even a series of problems. Too much of it creates excess heat, and not enough creates battery-driven devices that run down too fast. This is why fuel cells are so important for portable devices. The first generations will be fussy, but this is the best hope right now for a power source that is dense and controllable enough for widespread use.

Meanwhile, LEDs in one form or another are showing up all over the place. More and more of them are staring us in the face as we commute to work in the morning in the form of stop and turn lights on cars and in traffic lights. There has been significant talk about using them in room lighting as well. The OLED displays described in the article quoted above will more than likely drive the next generation of display devices for any number of embedded systems.

These are the advances that really drive our designs.

Larry Mittag

It could be that the venerable serial port is finally accepting it's retirement gracefully. The DB-25 and DB-9 connectors for this interface have eaten up large amounts of space on the I/O panels of innumerable systems over the decades, but it looks like it is finally being replaced by USB. In fact, the integration is to the point where it is being built into integrated architectures.

As comfortable as we all are with the venerable serial port, it's time to move on. Setups like On-the-Go give us much more flexible capabilities and support for external hubs are much more elegant than kludges like multidrop RS-422.

It's a relatively small step in systems design, but it is a welcome one.

Larry Mittag

In many ways embedded systems have been defined by what they don't have. They are just like "real" computers except that they don't have keyboards and video displays. Or perhaps what they don't have is megabytes of RAM to spare. Other popular items on the list of things they haven't had have been fast 32-bit processors, full operating systems, and disk storage.

One by one these differences have been falling by the wayside. Game machines are embedded systems that utilize specialized keyboards adapted to gameplay, and they and any number of other embedded devices have the option of no longer being "headless" and supporting video output. Memory is certainly much more abundant than it used to be. An early system I worked on was set up with 32 kBytes (the 'k' stands for 'kilo', meaning 1024 for those of you too young to count in anything less than mega levels) of RAM, and the hardware engineer at the time commented that that was more than enough for any reasonable software. I seriously doubt that you could buy that small an amount of memory anymore, but even if you could it wouldn't hold the RTOS and operating libraries that are part of modern systems software.

The last bastion of these limits has been disk storage, but even that is falling by the wayside. Companies like Hitachi are building very small, power-efficient disk storage that can stand up to the rigorous environment of mobile computing. Other options like flash drives are becoming reasonable from companies like Samsung. This differentiator is getting thinner and thinner as well.

Many systems are still built without these options. In fact, embedded systems can really be defined better as computers built with only the components necessary for a particular task. It's nice to know that they are available, though.

Larry Mittag

After a long drought in the design of new systems it seems like any number of variants is being discussed. Game machines are exploring non-PC (this can be either Personal Computer or Politically Correct - choose your acronym) architectures that are better suited to their application, as are any number of applications.

The latest option these days is the possibility of nanotechnology-enhanced systems. This option would merge the electronic and biological and open up the possibility of engineered devices for medical applications on a new, much smaller physically but larger conceptually scale.

The question remains whether this is a breakthrough following traditional sci-fi themes or is it yet another remake of Frankenstein? Is it even really possible?

Even if the high-concept applications for nanotechnology as a systems design option don't work out it seems that there are possibilities for more mundane applications of the technology. A recent EE Times Article detailed work that is being done on utilizing carbon nanotubes and other options for chip interconnects.

This illustrates the fact that nanotechnology provides a new way of thinking about systems design, but it also shows how broad the field has become. Fabrication technologies built loosely around nanotechnology concepts have been responsible for most of the product enhancements that have been credited to the field, but nanotechnology purists have not been impressed with them so far.

The engineering credo is that if it is useful it will be used, whether it is buzzworthy or not. By this standard is nanotechnology on the radar screen yet?

When you see a headline that talks about new system design technology it never fails to quicken the pulse a bit. I started reading an EE Times article with the provocative title “HP uses coding theory to make nano chips” with great anticipation. Coding theory? Nano chips? This has got to be cutting-edge stuff.

Imagine my disappointment when I actually read the article. They are creating a hot new technology called crossbar switches. They seem to have forgotten that this technology exists in every cheap keypad on the planet already. The fact that they have put it into a chip substrate doesn't seem to me to be a major technological advance.

Am I wrong here? Is there something I'm missing in this article, or something else to this technology?

I sat on a panel back in 1998 that was discussing fat versus thin clients for mobile devices. Thin was the vogue at the time, with WAP getting ready to take over the world. I was pretty much alone in pointing out that the weak point was the communications link, which at the time was (to put it politely) not ready for prime time.

Now I see the argument being raised yet again in Craig Mathias' latest column. His take is that thin clients will reduce client complexity (true) and take advantage of unlimited connectivity (huh?). The oddest point he made was that Web Services would be used simply to personalize the communications. This misses the point that Web Services is the backend technology that makes full applications on embedded devices a winning proposition.

Web Services is simply a set of conventions that allow data to be placed in context. This allows applications to parse data based on tags that are meaningful rather than descriptive of the presentation appearance for the data. This allows information to flow between applications in a structured way. The result is mobile applications that can use the information on the Internet.

Do embedded designers understand this? Are embedded systems being designed to use Web Services effectively?

Larry Mittag

The battle lines are getting clearer these days on the embedded device fronts. At one time there were approximately as many choices for an RTOS as there were systems that needed one. There were constant calls for a shakeout, but it never quite seemed to happen. Slowly over the last few years things have gotten a lot clearer.

Microsoft has gone from a laughingstock in the embedded space to a serious player. Even their cellphones are becoming interesting, probably because the hardware has grown into their requirements. A recent EE Times article gave them a chance to tout their new version of the smartphone edition of Windows CE.

The other side of the line has rallied around Linux. The support from Sony and IBM is well-known, and old-guard embedded companies like Wind River are joining this side as well. Some companies are even straddling the fence, such as Bsquare, which has been true-blue loyal to Microsoft in the past but is now supporting both.

The argument has always nbeen that embedded systems is too big and diverse for any meaningful consolidation. Could it be that hardware has grown to the point where a general-purpose operating system can start really hitting critical mass?

Larry Mittag

Much is made of Moore's Law, which deals with advances in semiconductor density, but we seem to be moving into a more complex world now where the relevance of it is diminished. No one cares that much how fast you are running if all you are doing is laps around an isolated island. Increasingly computing is defined more by cooperative systems that distribute information than by gate counts.

There is evidence of this all around us. Broadband access is finally becoming the default for Internet access in the home, and that access is driving new applications like VoIP and IPTV. Corporate computing is being redefined by virtualization, which uses networks to distribute computing power more effectively. These and other applications depend heavily on fast networks, which are getting faster with every iteration. This year alone should see the introduction of 802.11n and 10 Gbit Ethernet.

In a very real way the network is indeed the computer, as Sun's PR department has reminded us over and over. this represents a real opportunity for embedded systems to play a larger role in the computing infrastructure, since this world works more on machine-to-machine communications than computers that have a face (i.e. a user interface). Embedded systems can become the sensors at the edges of this new computer, as well as the distributed resources of the device itself. Could this be the dawning of the all-dominant supercomputer that has been the staple of science fiction?

Larry Mittag

Change is in the air regarding system design. Embedded systems hardware design has never been as boring as PC design, where progress is marked by monotonically increasing clock speeds and CPU selection can be done pretty much from a single vendor, except for those daredevils who opt for AMD processors.

In fact, the multiplicity of choices has been a problem for embedded systems with tool vendors being forced to support at least half a dozen major architectures and innumerable variations on those themes. It is only in the last few years that ARM has dominated power-sensitive designs and Motorola/IBM has done the same for the rest. This has settled down to the point where we could finally get some work done, so (of course) it's time to shift some paradigms!

The most radical change coming down the pike is quantum computing. A recent article in MIT's Technology Review shows how this little gem should be on us within a decade or so. My bet is that it will show up a little ahead of that schedule, and I am quite sure that I don't know how to program these beasts.

Less radical but perhaps more immediate is an application of nanotechnology to memory. There are scattered references to these technologies, but this Washington Times article gives a good overview. This breakthrough is a real blast from the past, in that it closely resembles the core memory mentioned in a post here recently.

Larry Mittag