The “Internet of Things” or IoT for short is increasingly getting a lot of attention around the world from various stakeholders including from no less than the U.S. Congress (see, for instance, the recent Bill introduced in the Senate, i.e., Internet of Things (IoT) Cybersecurity Improvement Act of 2017).
Since there is not an internet uniquely dedicated to things, the expression “Internet of Things” has come to be used as a metaphor or paradigm to describe a broad movement, still in its early stages, which is sweeping a wide array of economic sectors. In essence, it is broadly viewed as relating to the interconnection of intelligent things, i.e., endpoints with various degrees of smartness. These things, for most of them “out of scope” until now, are progressively inserted in the communications space thanks to the timely convergence of scientific, technological and societal advances and trends.
The IoT will bring about a “pulsating world” emerging from data being constantly sent to and received from IoT devices. As the IoT unfolds, products will be manufactured, as a matter of course, with embedded intelligence capabilities, spawning new ecosystems and related business models. It is bound to transform society and the economy on a scale not experienced before. If we add to these transformational powers the growing pervasive cybersecurity vulnerabilities, it is not hard to comprehend the scrutiny that it is currently enjoying.
The purpose of this article is to reflect on key engines of the IoT growth. While there are admittedly many, three well-known so-called “laws”, which should be perhaps better defined as “empirical observation”, “trend description”, or “educated projection” stand out. They all are influencing the IoT expansion, with distinct degrees of immediacy and impact, but have not yet enabled the Internet of things to reach its critical mass.
Moore’s Law
The September 2015 report produced by the Semiconductor Industry Association and the Semiconductor Research Corporation with support from the National Science Foundation about “rebooting the IT revolution” through realizing the “full benefits of the Internet of Things and Big Data” underscores the importance of Moore’s law in “insight technologies”: “Over those fifty years, the ability to reduce the size of the individual transistors by half roughly every 18 months has led to increased performance at lower cost and greater functionality in ever smaller form factors.” (Gordon Moore, Intel’s co-founder, used 24 months in his seminal 1965 article).
While there are legitimate concerns about Moore’s law “running out of room" and the battle to keep it afloat is getting harder, technological prowess has not been deterred. In June 2017, IBM, Samsung and Global Foundries announced that they had developed an industry-first process to build silicon nanosheet transistors that will enable 5nanometer(nm)chips, paving the way for 30 billion switches on a fingernail-sized chip. This is a remarkable feat given that, as a point of comparison, the most recent study of the brain shows that on average the human brain has 86 billion neurons.
Demonstrating that “there is a lot more room to shrink our electronics,” researchers at the Lawrence Berkeley National Laboratory reported in October 2016 that they proved the possibility (proof of concept) of a working 1nm transistor. While these transistors have not yet been packed onto a chip and mass production is still on the drawing board (in and of themselves gigantic milestones for sure), we may fairly anticipate that incredible computing power dwarfing the impressive capabilities already available today could be put on a sensor the size of a quarter in the near future.
It is no wonder therefore that, in the IoT community, edge computing, i.e., data processing at the border between the physical and digital worlds, is viewed as an optimal option.
Incidentally, concomitant to the striking reduction in size of sensors, actuators and wireless transmitters necessitated by the Internet of Things, the field of nanoenergy, i.e., the study of the small-scale, highly efficient energy harvesting, storage, and applications by using nanomaterials and nanodevices, has also greatly expanded within the last decade or so. “Nanogenerators” (at the core of self-powered IoT applications), which harvest small-scale energies in the ambient environment, have been invented in 2006 by a Georgia Institute of Technology team led by Professor Zhong Lin Wang.
In sum, enormous computing power within reach at a Lilliputian scale on site thanks to the continuing unfolding of the now-struggling-but-yet-still-alive Moore’s law stands at the ready to serve the emerging IoT space.
Cooper’s Law
Martin Cooper is an iconic figure in the telecommunications world. He is perhaps best remembered as the father of the cell phone who, as a vice president of Motorola, playfully placed the first mobile call to Joel S. Engel, a rival working at the headquarters of the Bell Labs in New Jersey on April 3, 1973 (as the story goes, Joel Engel does not remember taking the call). A true visionary, innovator and entrepreneur, Martin Cooper’s impact is far reaching.
For instance, in July 2003 he wrote an article for “Scientific American about how adaptive antenna arrays can vastly improve wireless communications by connecting mobile users with virtual wires. Titled “Antennas Get Smart,” the article describes how directional wireless antennas can improve the efficiency of mobile networks and reduce the amount of radio frequency exposure people endure,” as posted on the website of Dyna, a company that he and his wife, serial entrepreneur and inventor, Arlene Harris founded in 1986.
In 2010, Martin Cooper wrote a famous position paper on “The Myth of Spectrum Scarcity” and why “Shuffling Existing Spectrum among Users Will Not Solve America’s Wireless Broadband Challenge.” In it, he argues that the best and most economic solution is to use current allocations more efficiently, and underlines again what was already known as “Cooper’s law”, i.e., “technological progress has doubled the amount of available spectrum available for telecommunications since 1897 with a concomitant reduction in the cost of information delivery.” Subsequently, a presidential advisory committee concurred with him in a report published in 2012, which concludes that the radio spectrum could be used as much as 40,000 times as efficiently as it is currently, and increase capacity a thousand-fold.
As spectral efficiency is improving, networks may be able to absorb the demands for capacity of the spectrum-hungry Internet of Things.
Thomas Hazlett in his new book on “The Political Spectrum: The Tumultuous Liberation of Wireless Technology, from Herbert Hoover to the Smartphone” (Yale University Press, 2017), concisely summarizes the strength of this performance gain; “we today enjoy one trillion times the wireless capacity of networks a century ago. This furious pace is not slackening.” (p. 2). That the pace is not abating is vividly demonstrated by the announcement in August 2017 that international researchers from Brown University (United States) and University of Lille (France) “have demonstrated the transmission of two separate video signals through a terahertz multiplexer at a data rate more than 100 times faster than today’s fastest cellular data networks.”
Metcalfe’s Law
Metcalfe’s Law is named after Robert Metcalfe, the co-inventor of Ethernet and co-founder of the networking company 3Com Corp., in Santa Clara, California. Many times restated (initially, around 1980, it was about devices), as reportedly formulated by George Gilder in 1993, it claims that the value of a network grows as the square of the number of its users — or stipulated somewhat differently, the value of a telecommunications network is proportional to the square of the number of connected users of the system. It also points to an enticing proposition that once the network gets past a critical mass of early users it can snowball into very profitable territory where the benefits expand faster and bigger than the costs.
Essentially, this rule of thumb is about the network effects familiar to any graph theorist, economist, or industrial system engineer. Trying to see the statement as a strict representation of reality may be a futile exercise (how is value defined? Is it really proportional to the square of users? – pitfalls and possibilities are an ever-present topic of research and other “laws” have been proposed to ascertain the value of a network), but viewing it as the catalyst of network success can be very helpful.
It is easy to see that with a starting point of 100 users, 4950 unique connections between the network members (i.e. = (100 x 99)/ 2) can be established and when the number of users in the network augment to 150, the possible links jump to 11,175 (i.e. = (150 x 149)/ 2). Consequently, while the network has increased by 50 percent, the number of potential relations has now grown by 126 percent, i.e., more than double, which shows that the “value” of the network, however it is defined, increases faster than its size.
Of course, such network effects (as captured in Metcalfe’s law) are inherent to the success of social media companies (SMCs).
SMCs are digital billboards that promote their fast-growing number of captive eyeballs to advertisers willing to pay good money to reach selected targets. In order to secure these eyeballs, SMCs provide free appealing services, which could be construed as “anchor magnets”, in return for keeping the advertising rights. This business model can perhaps find its roots in (or at least similarities with) advertising-supported revenue models, such as traditional TV and radio networks, and also the innovative concept JCDecaux introduced to French city mayors in 1964, which consisted in providing and maintaining bus shelters, fully financed by advertising.
A side of Metcalfe’s law that may be overlooked is the incremental value potential generated by newcomers, a key element of the SMCs’ growth that is not related to pre-existing network residents. Metcalfe’s law rests on every new user/node in the network connecting with every pre-existing user/node and creating, as a result, more than a linear increase in the overall value of the network. However, new entrants will not only connect with existing network members, but also bring with them “fresh nodes” into the network, i.e., individuals who are not already in, and at the same time, can create the conditions for a positive feedback loop regarding the usage intensity since more members are likely to generate novel (if not more of the same) ways to connect with the whole network.
As it has been highlighted many times, a telephone, fax machine, or social media account have, by themselves, absolutely no value. They need companions. Non-users of social media can attest to the nudge, sometimes pressure put on them by friends, colleagues and family to join them in the network. This is a non-trivial point. Growth is inherent to the very nature of the SMCs’ business. Their leaders don’t have to market their freebies (anchor magnets) to their captive eyeballs (the product), the expansion, by and large, takes care of itself. While we are only interested in this article in exploring the growth phase, it is clear that adverse effects, based on the same principle, can play in reverse.
In their growth phase, successful SMCs have been able to leverage the explosive force of Metcalfe’s law. By designing extremely attractive offerings (anchor magnets) they have been able to trigger a self-sustaining chain reaction, thereby allowing them to approach advertisers with an effective threefold value proposition i.e., we have a sticky user base; we know it well (multi-dimensional segments); and our network is getting bigger.
By contrast, on the IoT side, companies are still in search of scale and must find ways to bring Metcalfe’s law to bear; as a whole, the IoT space is still far from having reached critical mass beyond which it can sustain self-reinforcing growth.
As a case in point, IoT Analytics, a Hamburg, Germany-based provider of market insights for the Internet of Things (IoT), M2M, and Industry 4.0, reported in July 2017 that only 7 percent of the 450 IoT platform companies they had identified around the world generated revenues in excess of $10 million with their IoT platforms in 2016.
Conclusion
Erik Brynjolfsson and Andrew McAfee, respectively Director and Associate Director of the Center for Digital Business at the Massachusetts Institute of Technology, have noted the importance of such growth engines in their well-received book The Second Machine Age: Work, Progress, and Prosperity in a Future of Brilliant Technologies (Norton, 2014) — p. 37. For Brynjolfsson and McAfee, the fundamental characteristics of technological progress are exponential (i.e., constant rate of increase and not constant amount of increase as in the linear case); digital (i.e., product does not diminish or disappear when it is used, and is extremely cheap, i.e., close to zero, to reproduce; hence the term zero-marginal cost society that Jeremy Rifkin advances to describe an IoT-framed society); and combinatorial (i.e., recombining things that already exist; a perspective, which echoes the view on innovation – or, at a minimum, its spirit - of Professor Joseph Schumpeter, one of the greatest economists of the first half of the twentieth century, summed up in his book on Business Cycles in 1939 as “in short, any ‘doing things differently’ in the realm of economic life”).
While everything seems to be set to propel the Internet of Things to oft-announced dizzying heights, the IoT — albeit undeniably rapidly taking off — has not enjoyed thus far the same astonishing growth witnessed on the social media side (“the Internet of People”).
Moore’s and Cooper’s (exponential) laws have continued to deploy their wings with remarkable fidelity to their original concept and today provide solid foundations for the IoT expansion, but Metcalfe’s (or any version of this law centered on the critical positive non-linear repercussion of additional members on the network’s value) has still to make an impact.
This is of vital importance for the IoT space since mass induces value.
As emphasized above, exponential growth is a key feature of SMCs’ success and, more generally, technological progress. However, exponential growth functions display different behaviors depending on a range of parameters. Some grow faster than others.
We saw that the fast growth of SMCs’ platforms may be due to some extent to the extra value new entrants/social media accounts add by bringing with them potential not-yet-in-the-network users (as in the case of telephone or fax machine networks). There is some type of codependency that serves as a growth accelerant. When individual “A” joins, he/she may bring with him/her individuals “B” and “C” who in turn may add their own followers, i.e., all newcomers who may also spur network intensity usage. Growth is a built-in component of the SMC product.
There is no denying that the Internet of Things is on the rails and has, for all intents and purposes, already “left the station.” But, when IoT applications will foster by design the incorporation of new connections, pulled in at wildfire speed by existing nodes whose network membership is closely intertwined with those new connections (i.e., one connection begets another connection that begets another etc., in the manner of social media), the Internet of Things will grow beyond critical mass and reach a size that we can barely fathom now. This possibility may take some time to materialize but has certainly already moved out of the science fiction realm.
The views expressed in this article are solely the author’s and do not necessarily represent those of the Georgia Institute of Technology (“Georgia Tech”), the Georgia Tech CDAIT members, the University System of (U.S. State of) Georgia or the (U.S.) State of Georgia.
