One area I’m paying a lot of attention to these days is what I’m calling “embedded intelligent networks,” or networks that integrate seamlessly into physical reality, enabling us to monitor and ultimately manipulate the real world across the net.
EINs include sensor networks, concatenations of devices that monitor, track and control the environment. Sensor networks are often wireless, and have a wide range of applications: Tracking environmental factors like temperature and humidity; performing industrial applications (monitoring pressure in pipelines, for instance); and even mapping traffic.
But EINs are more than sensor networks. They also include networks that don’t just monitor reality, but also manage it. The hot new example these days: so-called “smart grids”, which delivers electricity from suppliers to consumers using embedded intelligence to save energy (and thus costs) and increase reliability. Folks use the term for both transmission and distribution grids, and smart grid functions refer to anything from meter monitoring to automation of bulk power transmission. The notion is that by tracking and managing energy more efficiently, utilities will be able to reduce costs; by providing early warning of potential outages, they’ll increase reliability.
There are also a host of other EIN applications: security cameras, for instance. Data center monitoring networks are another example. You could even think of the network of consumer camera phones as comprising an EIN of sorts.
Then there are the way-out applications: Remote-controlled robots, like the kind the folks over at the Readybot project are up to. And of course, telemedicine. There’s also so-called “M2M” — machine-to-machine — communications, though I’m not in love with the moniker (I understand it used to have entirely different connotations over at Craigslist, before they shut down that particular service offering).
Most of these applications have been around for a while — so what’s changed? Two things. First is that many of these networks are now transitioning over to IP. Second is the increasing capability and ubiquity of wireless — yesterday’s low-bandwidth links are on the verge of being replaced by broadband.
In both cases, there are benefits and challenges. Start with IP: the upside is that devices and systems can communicate using well-known, well-understood protocols. That makes systems both potentially less expensive and infinitely more flexible. However, in previous columns I’ve highlighted some of the challenges with IP scalability — issues that are likely to get worse before they get better. Similarly with wireless: on the bright side, there will be plenty of bandwidth (at least where it’s available). On the downside, scalability is in question (there are physical limits to transmission).
Notwithstanding the above, the rise of embedded intelligent networks will be significant, changing our world in much the same way that the rise of microprocessor did. Those of you who can remember back to a time before every watch, lock and greeting card had embedded intelligence know what I mean.
The upshot? Networks are about to get very, very interesting again