Managing the growing confusion of opportunities described by the all encompassing use of the term "Internet of Things" requires clear thinking and categorisation, as well as fewer marketing managers tying their "things" to an inappropriate bandwagon.
The Internet of Things—indeed many aspects of electronics and embedded systems design—needs a standards group for naming things, much like the IETF, to define what things are and what they mean. And its membership should exclude vice presidents of marketing, product marketing managers and anyone who deals with directly selling or marketing or promoting a product.
Thank the universe that there were not marketing managers around in the age of Volta, Gauss, Ampere, and others in the seventeenth and eighteenth centuries when they defined and investigated the basics of the technology we all use. I shudder when I think about the mess we would be in now. Indeed, I doubt we would be where we are now. More probably we would be just now pulling out of another Dark Age like the one that started in or around 1013.
Earlier in this decade, when academics and researchers were pioneering the underpinning of what we now call the Internet of Things, it was clear what IoT and even Machine to Machine meant to them—interactions between things over the network, no humans involed.
Anything that directly or indirectly, involves humans interacting, or causing to interact can no longer be called an Internet of Things. Maybe an Internet of Humans and Things (IHAT?) or an Internet-Defined Interconnected Overlay of Things (IDIOT) for humans, but not IoT.
When you have clear definitions and classifications you can define technical characteristics. In an Internet of Things, without humans directly involved, for example, the definitions of real time and deterministic are in the millisecond and microseconds and determinism is very tightly defined.
If the industry insists on the use of IoT we should completely pluralise the terminology. Not Internet (singular) of Things (plural) but Internets (plural) of Things (plural). Human-involved conceptions of the IoT aside, there are many Internets of Things: for building automation , for lightweight sensor networks, for home automation, for the Weightless whitespace. And each will operate by different kinds of requirements.
Before the "Internet" was there was the Internetwork, a matrix of multiple vastly different networks connecting servers and minicomputer and mainframes and PCs. The Internet Protocol was a "Lingua Franca" by which these proprietary networks exchanged data at the interfaces between them. Eventually, most specialised network protocols, such as in industry, or for mainframes, minis, and so on, disappeared and the interface became the common language.
Now the reverse is at work. For a time, there will be an Industrial IoT, a mobile IoT, a building and automation IoT—each of which will only partially be able to talk to each other. And then, as in the past, some common interface protocol will emerge that as in the beginning to allow all of the multiple definitions of IoTs to interact. But we aren't at that point yet.
In the not too distant future, when IPv6 becomes truly ubiquitously used, the sheer numbers of URLs that are available will be thousands of times larger than the total number of humans on the planet, all their pets, all their mobile devices, and every household appliance.
In that environment the main users of the IPv6 will be things, billions and billions of things, not humans, and it is we, not they, who will have to do make the accommodations.
As Rick Merritt reports in "Needed: A glue for the Internets of Things," various market and industry segment groups, such as in the industrial control, are forming groups and special interest groups to define what is meant by the IoT phenomenon. Similar efforts are underway in building automation, and within already existing and new markets, such as Zigbee and Weightless white space spectrum use.
But what it will take will be someone outside, someone from academia, possibly, someone in the footsteps of Vint Cerf, who among others came up with the Ethernet protocols upon which TCP/IP is based. And it will with certainty NOT come from within any of the various groups currently defining their conceptions of IoTs. Why? Because at the same time they are trying to define what IoT means in the context of their particular segments, they are also trying to ride the wave of enthusiasm that surrounds all things IoT.
But from whatever direction it emerges, please, for the sake of my sanity at least, ban all vice presidents of marketing, marketing managers from involvement. They need not apply.
But hurry. I get dozens of product pitches each month, half of them person-to-person and there is always some aspect of the Internet of Things thrown in. It often takes me at least a quarter of the interview to pin down what they mean by the Internet of Things. It is really getting tiresome.
用户3720650 2013-8-24 11:11
I have to respectfully disagree with your definition. Machine to machine interaction is just one aspect of extending worldwide networked communication of sensed information to non-human actors.
Perhaps Cisco's "Internet of Everything" is the most descriptive term. Whatever is used, there is one concept of a network of data generators, communication channels (bidirectional), and data analyzers, for which the mammalian nervous system is a good model. (One needs actuators along with the sensors in order to keep this analogy.) Sometimes humans are involved; sometimes not. I don't see that as a defining characteristic of this theme at all.
To be sure, different applications may exclude humans, and some of those (not all) will function on time scales which our slow neurons can't handle. That can be taken care of by any number of IEC standards.