15 billion devices, one embedded network - Embedded.com

15 billion devices, one embedded network


November 2028 is the 40th anniversary of ESD. Click here read other 2028 lookbacks.

Near the beginning of this millennium, Intel's CEO Andy Grove predicted a future in which one billion computers would be connected to the Internet. At the time, the number seemed astounding. But here we are, having far surpassed that prediction, just as if it's always been that way and connectivity was never not there.

In 2008, Intel predicted a future of 15 billion connected devices. We turned out to be right. Now in 2028, these devices are not only connected but have massively simultaneous wireless communications among them. Computing is truly ubiquitous, where people are connected to people and machines to machines–always, anywhere, and without fail. Technology is now so finely woven into the fabric of our lives that we aren't even able to see it any more.

We've arrived at fourth-generation Internet, what Intel calls the Embedded Internet, where computing is ever-present, portable, and abundant. Incredibly small yet fully featured devices don't need wires or “plugged-in” power supplies to get connected and stay that way. And the supporting infrastructure is so all-encompassing that connectivity is seamless, without gaps or holes.

As an industry, we still have much work to be do to maintain our future. We must update and follow the Universal Standards, because they're the only assurance that everything continues to works together, seamlessly and pervasively. The hard-won standards ensure interoperability and commonality so that every distinct device and idea can connect and communicate with every other one.

First and foremost, everything to do with computing and communications must continue its move to an IP-based infrastructure. That's because some of the more advanced solutions today–and those of the future–combinate distinctly separate technologies. And to ensure that the different technologies and ideas developed in different places in the world can come together quickly to make complete solutions, is for them to be developed on the same protocol platform–an IP-based infrastructure platform.

In that same vein, we must continue improving the wireless technology that enables us to seamlessly connect everything together and remain connected wherever we go. There's so much great broadband wireless technology available today; we can download applications to our cell phones and do some really cool things on some of the latest devices. But we still have challenges, just as in 2008, when I, for instance, could not maintain a cell phone connection when driving from the Intel campus in Chandler, Arizona to the Phoenix airport. The call would get dropped every time, and I could tell you exactly which mile marker it was. The Embedded Internet requires us to become even more unified and standardized in our wireless infrastructure so that computing and communications could be always-on, always-connected, without fail.

As an industry and especially here at Intel, we continue to move toward the goal of truly low power. We already broke the sub-one watt platform barrier, which made devices more portable and power efficient. We also had to get away from our dependence on batteries and power supplies because we couldn't make batteries fast enough for 15 billion devices. So we developed proximity power sources like solar energy and parasitic vibration-powered devices, with the ability to recharge anywhere we go.

Energy efficiency, too, must scale a few more orders of magnitude from where we are today. This isn't just a factor of low power on the platform, but it includes everything from carbon emissions to cooling and fans and the total energy equation. The entire computing grid needs to get more power efficient. In 2008, we had started to do things like time-slice power consumption so that devices were responsive when we needed them to be and slept deeply when we didn't. And we continue to look at things like solid-state memories instead of hard drives, because they don't put out as much heat. As an industry, we couldn't scale linearly in energy efficiency or we would never had made it to 15 billion connected devices.

Today interaction between man and machine is vastly different from our experiences in 2008. In 2028, we're already at the point where I'm even willing to predict that the time is not far off when we won't have to type anymore. The keyboard as we know it is giving way to visualization, precise speech recognition, and even cerebral interaction between people and their devices, where you can control a device simply by thinking about it or gesturing toward it or looking at it. This means we'll have to scale our processing capabilities accordingly, including a whole new dimension of granularity in terms of what kinds of data and distinctions can be made by machines, and then analyzed and processed in near real time. Such fine-grained sensory ability in the platform does not fully exist today.

Twenty years ago, I firmly believed all of these things would happen by 2028, and many of them sooner. Truely, today many of these technologies are so comfortably ensconced in our lives that we barely realize how or when they arrived.

In 2028, we look back at our embedded journey with a combination of awe and amusement. Awe–because the accomplishments we celebrate are shocking in their magnitude. And amusing as we recall the awkward way we did things back then. How did we ever functioned that way?

As vice president of the Digital Enterprise Group and general manager of the Embedded and Communications Group, Doug Davis is responsible for Intel's business in the communications infrastructure and embedded market segments.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.