I had the pleasure of attending and speaking at the EE Live Internet-of-Things conference in San Jose last month. Since then, I’ve been reflecting on some of the bleeding edge technical concepts that were discussed there.
It occurs to me that many of the IoT devices we are asked to design act as a standalone island. That is, there is a device, some back office service, and a mobile app, and these three items talk with each other — but they generally operate independently of anything else. I call this IoT 2014, but aside from the prevalence of Wi-Fi and the cost of components, the thinking really hasn’t evolved much since 1996. Almost every IoT device on the market fits into this category today: Belkin WeMo, smahttp://www.emersonclimate.com/en-us/products/thermostats/sensi/Pages/Sensi.aspxrt power strips, and numerous Wi-Fi thermostats. Some of these devices have minimal ability to interconnect to other services but that functionality, if present at all, is usually an obvious afterthought.
One theme from the conference is the idea of forming ecosystems of devices that work together to provide much more transformative functionality. As ecosystems start to take shape it becomes more cost effective to look at short-range wireless technologies such as Bluetooth, Zigbee, or Z-Wave, along with some level of autonomous operation when not connected to the Internet server. Such technologies need gateways to access the Internet, but they also come equipped with point-to-point, low-power, ad-hoc networking that generally operates on a store-and-forward basis. As a result, these devices can form impromptu local networks. Such networks allow nodes to come and go from their fabric building in great robustness.
So let’s run with the idea. As devices with short-range wireless technologies become increasingly common, it might become possible to build ad-hoc networks over longer ranges, via node hopping. When density is insufficient, you can bridge to the Internet to route over gaps.
Over time, as the node density increases, network gaps would decrease, and therefore Internet reliance decreases. Likewise as node density increases, so do communication paths, which yields greater effective bandwidth. Sound familiar? In a general sense, we’re starting to think in terms of the BitTorrent protocol.
Such a network evolution suggests a reduction in the importance of the centralized Internet as we understand it today. We arrive at a crowd-sourced network paid for and operated by consumers in the form of the devices they purchase and use.
Given recent advances in brain-inspired processor design, one wonders if science fiction finally meets science fact and we create an intelligence higher than our own while we’re at it. OK, that thought is really out there!
Whether we birth a new life form or not, certainly the opportunities to build ecosystems of IoT devices that enhance and improve everyday life are there. Furthermore, your personal ecosystem need not even be hosted on your own hardware. For instance, when you walk into a hotel room, your personal preferences for lighting color, room temperature, and mood music could automatically follow you everywhere you go. This merges neatly with the concept of ubiquitous computing, and takes technology to a whole new and personal level.
I’m pretty excited about this future possibility, but there is also no small amount of fear surrounding the consequences of a global communication and computing platform. (In particular, this well-reasoned piece by Tim Geer should give pause to all IoT engineers.)
Time will tell if this future will unfold, or I’ve just echoed another hopefully entertaining science fiction story.