Internet of Things, Internet of Everything

(Last Updated On: September 5, 2013)

Cyborg man MC900326134 200pxScience Fact: Internet of Things, Internet of Everything. It’s hard to miss those image ads from Cisco and the breathless hype from many sides: from technologists; from self-styled futurists; from investment advisors looking for the next big stock killing; from professional societies starting up new conferences and journals. I’m obliged to my friend Joe Behrens for nudging me to look beyond the headlines to the reality that may lie here.

It’s undeniable that we are increasingly interconnected to things as well as each other, and “Everything” sometimes seems like not too much of an exaggeration. Consider a few facts:
– We already have chips in buildings, vehicles, appliances and animals, and many of these are in continual communication with the Internet
– We are becoming increasingly bionic — our bodies used to use only passive prosthetics like hip joints and blood vessels, but now have electronically active, interconnected attachments; some are implanted like bionic skin, others are worn like Google Glass and smart watches, but even the implant-versus-wear distinction is disappearing
– The types of system architecture, the way things talk to one another, are proliferating without any apparent limit: cloud-based networks, ad-hoc nearest-neighbor messaging, network designs that may be thing-centric, data-centric or service-centric.

Almost any parameter you care to plot — numbers of devices, data bandwidth, network complexity — is continually increasing with no obvious ceiling.

Science Speculation: We have to wonder where it’s all going, this proliferation of sensors and Internet connections in every object around us, and in lots of living beings as well.  And the remarkable durability of Moore’s Law through 40 years — and many generations of technology — might tempt us to simply draw a line and extrapolate. But if we try to predict the future, we are speculating, not strictly analyzing facts and data.

Linear extrapolation is not the only natural law: cycles, like the swing of a pendulum, are also extremely important.

Here’s an example that sure to annoy you, whatever your politics: The activities of man contribute to global warming — but how much? The world has cycled between warming and cooling many times in geologic time. Like many things in science, the answer is neither black or white. Cycles also occur in human society: for example, business cycles and swings in the stock market; and the periodic swing between intervention and isolationism in global politics.

It’s possible that the proliferation of interconnected “things” is one of those swinging pendula (yep, that’s a possible plural) that will swing back, to a much less interconnected world. And when the pendulum swings, it usually goes PAST the center point to another extreme.

Am I being a curmudgeon? A Luddite? Perhaps, but we need to pose the question. I can see social trends that could make us all demand significant controls on “connectedness”:
– The ubiquity of viruses and hacks that threaten our data, our money and our privacy: what would we pay, in money or in limitations on the Internet, to be free of these risks?
– The flood of spam and advertising that blunt our Internet experience
– The economics of pay-per-view Internet which might take over from subscriptions and advertising
– The lack of security and accountability that comes from when we can’t positively identify the person at the other end of our communications

Are you convinced that the Internet form of Moore’s Law will continue till all present readers have gone to their reward? Or could there be a swing back to the middle and even beyond?

What do you think about Internet of Things and Everything? Are we headed there, or could there be a turn in the road?

Additional Sources:


Internet of Things, Internet of Everything — 4 Comments

  1. Moore’s Law is sooooo 2010! Intel and such cannot admit in public that it is at best flattening out — its proponents have simply been redefining what it means as they search for the next tech to replace the CMOS IC based on photolithography. But–as I wrote a while back in article “Tunnel at the End of the Light” in Issues of Science and Technology, Spring 2012, the S curve is clearly bending with potentially major geo-economic and even geo-political consequences. While fundamentally new types of ICs or IC-alternatives may emerge, they are not yet here and it is highly uncertain what they will be or who will produce them. On the other hand the “internet-of-things” may well be driven by other tech, such as MEMS and optoelectronics–so the dynamics of more info everywhere may still be in the offing, whether “we” want it or not.

  2. It does feel like society at large is currently buying into every new technology or gadget with a Mr. Magoo-like optimistic blindness. Who would have predicted 20 years ago that people will compulsively upload pictures of every dessert they eat or tweet every thought they have? Maybe we will become more cautious with technology as the negative ramifications are discovered, but we’re also now so dependent on so much technology that a retreat into the pre-internet days doesn’t seem likely.

  3. In trying to understand time-based physical or social phenomena, I believe there are several useful mathematical models. At the minimum those would include linear extrapolation, cycles, and dispersion around a mean via a probability (Gaussian) function. Linear extrapolation is most often used during a rapid rise or fall when no one really understands the drivers of the situation, and because it is easy to calculate … but nothing truly behaves like that in the real world if you adopt a long enough timeframe.

    A more realistic model than linear extrapolation is the S-curve:

    … but I tend to believe that an S-curve in the real world is just the front end of a long-term Gaussian form. All things come to an end with a long enough timeframe. The problem with an S-curve and most of the non-linear mathematical models is that the exact parameters which apply are hard to determine, and so the usefulness of the concept is more intuitive in the sense of an analogy than directly amenable to calculation for the purpose of prediction.

    For that reason, most people attempting to make use of time-based data readily adopt linear models because they are easy to apply and understand. An example would be Microsoft’s stock price if you examine it over its history using a semi-log plot on the price … from about 1990 through 2000 it shows a very nice linear ramp which is tempting to believe will go on forever, but then the stock price rather sharply flattens in 2000 to be the low-return stock of today. Cisco shows a similar pattern. A linear model for those stock prices would have been very useful for an investor if adopted in 1992, somewhat useful in 1997, but useless or disastrous in 1999. If you are going to apply a linear extrapolation to any real world phenomena, it’s useful to regard it the way you would log-rolling … it’s important to know when to “get off” the current log (model) and switch to another one.

    So, what does this say about the everything-interconnected world we seem to be headed for? That it very much depends on your timeframe as to “which tool” best models what is happening, and what might happen in the future. Yes, Moore’s Law still governs … a remarkable model. And it is very likely to continue in the short term (years), still possible in the medium term (decades), but unlikely to continue over a long term (centuries). The same is true for any current social phenomena you wish to discuss though the short-to-long time frames will differ. So, today’s most popular sports, or the choice of conservative vs “liberal” sentiments in this country, or how the unrestricted “Wild West” nature of the Internet will all change over time — these can all later be described by social historians (the masters of hindsight) using one of the mathematical models I’ve described above, or something similar.

    The real issue is not whether you can fit a mathematical model to any observed phenomena (that’s always possible) but instead how successfully you can use that model to predict the future in some reasonable projected timeframe and with some reasonable accuracy. If you get it right you’re regarded as a genius, if you get it wrong you’re a dud. The truth is that projecting the future is a dangerous business, and too-often based more on luck and timing than anything else.

    Isaac Asimov, one of the grandmasters of science fiction, wrote an article for the New York Times in 1964 attempting to predict what life might be like for us 50 years later. Like opening a time capsule, it makes interesting reading, and though some of his thoughts were spot-on, others are wildly divergent from our present reality:

    or if that URL “breaks” because it’s too long for this forum, this URL is equivalent to it:

    • Thoughtful comment, thanks very much Charles!
      Cisco’s promotions plot a sequence of S curves, each taking over from the other — but of course they don’t want to forecast an eventual falloff, or even a leveling out of the family of curves.
      Thanks also for the Asimov reference – a well-beloved author.