An Embedded Transformation Is Likely to Occur

Is Intel poised to revolutionize the embedded market?

By Geoffrey James, Contributing Editor

The embedded segment of the semiconductor business is white hot and growing. New devices are entering the market every day with an ever-increasing list of functions and capabilities. To create and support these new devices, embedded-systems designers are turning to more powerful chip architectures, says Linley Gwennap, founder of the Linley Group and co-author of A Guide to High-Speed Embedded Processors. “We’re seeing applications that really push the limits of what these embedded systems can do,” Gwennap states.

Intel has never been a dominant player in embedded devices. Yet it’s likely to be a major beneficiary of the trend toward greater complexity, according to Ryan Parker, director of marketing at Intel’s Embedded Computing Division. “We see the game changing from a custom-design model to one that leverages the standardization that’s inherent in an Intel® architecture-based design,” he explains. Is Intel poised to dominate the embedded market in the same way that it dominates the market for generalized computing?


To answer that question, it’s necessary to understand exactly how the embedded-systems market is changing. Perhaps the most important shift in the embedded space is the need for greater connectivity. In the past, many embedded devices were standalone. Or, if they were connected, they were hardwired into some kind of proprietary network. Today, however, an increasing number of embedded devices are connected to the Internet and the wider world of general-purpose computing. “By 2015, there will be 15 billion connected embedded devices and most of them will be interacting through some kind of standard wireless interface,” says Parker.

Rather than developing custom chips, Parker believes that the need for wireless connectivity will drive embedded-systems designers to use chips that are standardized. While it’s possible to include wireless IP inside custom-designed chips, such IP doesn’t always work reliably. In the worst case, Parker says that it can create a “back to the drawing board” situation that can be fatal to the marketability of an embedded device.

As embedded chips become more complex, incorporate more circuitry, and move closer to the system-on-a-chip (SoC) model, it becomes less economical to do a custom design. At the same time, it becomes more economical to use the kind of standardized chips that are Intel’s forte. According to Parker, “We believe that embedded developers are going to be drawn toward an architecture that has proven reliability—especially in the areas of computing, where they currently lack expertise.”


In addition to adopting wireless connectivity, the embeddedsystems market is moving toward the increased use of multicore architectures. The driver for this standardization is the rapid unit sales growth of small-form-factor devices like smart phones. An increasing number of these devices are being used to interact with and integrate with more traditional, general-purpose computing devices, explains Ken Cayton, research manager of enterprise platforms at the market research firm IDC. “As more functionality is crammed into smaller devices, they look more like larger devices in their compute requirements,” he says. “Just as desktop and laptop clients and servers have moved to multicore devices, you will see other devices adopt the same type of device.”

The embedded market is no stranger to multicore. But now things are different, emphasizes Cayton. “Embedded devices have long been using chips with multiple cores and multiple types of cores. In the past, they were likely to be custom chips—whereas today they are beginning to leverage the volume manufacturing of companies like Intel and the common x86 tools and code base to simplify their development process.”

This isn’t to say that the transition won’t be without its share of hassles. “Embedded systems are generally very-well-defined application environments [while] the desktop environment must accommodate a significantly more eclectic application domain,” says Neil Henderson, general manager of Mentor Graphics’ Embedded Systems Division. “Programming models for the two vary significantly due to these distinctions, [which] must be taken into consideration when designing embedded systems in general and even more so when it comes to multicore environments.”

Yet even the benefits of standardization outweigh the difficulties, says Parker. “Developing software for embedded systems is currently a very labor-intensive process involving a great deal of specialized training. Moving to a more standard set of development tools allows embedded-systems firms to draw upon the expertise of a much larger pool of programmers.”


As embedded systems become more complex and provide more functionality, they inevitably become more difficult to support. That’s where Intel has an incredible advantage over traditional custom designs, states Parker. “Our Core processors support Intel® Active Management Technology (Intel® AMT), which will greatly reduce ongoing support costs.”

Intel AMT consists of a tiny piece of hardware and firmware, which is built into the CPU chip, that’s essentially “always on.” As long as the platform can be contacted (either wirelessly or through a hardwired connection), an authorized computer can access the “guts” of the system—even if the system is on standby or powered down. A technician can remotely poke around at the firmware, check various settings, and make changes as necessary to restore the system to health. It doesn’t matter if there’s a crashed hard drive or a locked operating system.

According to Parker, Intel AMT also reduces overall power consumption. “Many users of embedded systems leave them up and running—even when they aren’t in use—simply because they’re afraid that they may need to download a patch,” he explains. “With Intel AMT, they can power the system down and know that essential maintenance tasks will still take place, which provides considerable cost savings.”

Will Intel dominate the future of the embedded market? Obviously, it’s too soon to tell. However, it’s clear that Intel takes the market very seriously and intends to do whatever it can to expand beyond its dominance of general-purpose computing. “Each time embedded systems become more complex, it drives more people to our architectures,” says Ryan. “If you believe that embedded systems will continue to dominate the future of computing, then you’ve got to believe that Intel will capture its rightful share.”

Geoffrey James is a regular contributing author for Embedded Intel® Solutions magazine. He is both an author and journalist who writes about business, technology, public policy, strategy, and sales/marketing. Geoffrey has written over a hundred feature stories for national publications.