Internet of Things: The King has no clothes

Dr Peter Harrop
Internet of Things: The King has no clothes
Some use the term Internet of Things (IoT) to rename a large number of existing topics including the burgeoning world of cellphones, tablets and wearable connected devices, better referred to as the Internet of People (IoP). Cisco and now others, usefully refer to the Internet of Everything as encompassing 'data', the 'Internet of Things' (in a more narrow sense of "thing to thing") and the 'Internet of People'.
IoT focusses mainly on systems with Internet Protocol (IP) addressed, sensing tags containing mocroprocessors. "Billions of intelligent connections that will encompass every aspect of our lives and make our world smarter, greener and safer" (Freescale Semiconductor). Like Freescale, we believe that the biggest opportunities within this IoT will be in the transformational shift from the computing nexus to highly intelligent nodes - when intelligence massively scales, and the nodes have the power to learn, adapt and communicate, triggering action without human input. Cisco talks of connecting new places to the internet - such as manufacturing floors, energy grids, healthcare facilities and transportation systems. When an object can represent itself digitally, it can be controlled from anywhere. This connectivity means more data, gathered from more places and of better quality because human input, the source of most data on the internet today, is not required.
Contrast between IoT and IoP
The contrast of this embryonic business with the maturing, huge IoP could not be greater. Sensing by nodes and collaboration between nodes are key functions of the IoT where human interfaces are simple or non-existent, whereas ever more sophisticated and varied human interfaces, voice and written communication and data access by humans are defining features of IoP where ubiquity is the big priority.
A dream whose time has come?
Many of the things preventing the IoT dream of "things that think" from 1999 and the earlier "embedded computing vanishing into the fabric of society" that were not in place then have recently fallen into place. They include:
• Low-priced appropriate microcontrollers and sensors.
• Internet IP addresses increased to an effectively infinite number by IPv6.
• Internet access far more ubiquitous particularly via cellular, Bluetooth and WiFi.
• 900MHz WiFi 80211ah could provide WiFi with dedicated bands for ultra-reliable, always-on machine-to-machine connections.
• Getting hungry - Large suppliers to the cellphone and allied industries see fast growth ending and seek to invest huge funds in something that has at least as much potential and leverages their skills.
• Willingness to invest - The major suppliers involved are sometimes declaring that they will now invest heavily in the IoT. For example, in February 2014, GE announced it would invest $1.5 billion "industrial internet" research in this area from 2012 to 2015. Intel and Qualcomm are investing in prizes and start-ups and Google bought the tiny Nest smart thermostat business for $3.2 billion.
• Large government investment in smart cities particularly in China where things communicate and act autonomously.
• Large potential IoT saving identified by suppliers in aerospace, industrial sectors etc.
It is likely that the Internet of Things will eventually become much larger than the Internet of People, if only because there are far more appropriate 'things' than 'people' as candidates. All that is the good news and it has led to a rather extreme form of euphoria where analysts and aspiring suppliers outbid each other to predict ever increasing numbers of nodes in the tens to hundreds of billions deployed in 2020, from a tiny number today. However, in researching the new IDTechEx report, Internet of Things: the Business Opportunities 2015-2025 IDTechEx has found another side to the story that strongly indicates that IoT will have a slow start, mainly based on small companies doing small projects. A large number of impediments to mass rollout are identified. There is little attention to dealing with some of them.
Most crucially, there are often other ways of doing the same thing and the owners of such systems often have neither the funds nor the inclination to change to IP enabled sensor nodes. Large potential users thinking of committing the huge funds that the large suppliers hope to attract are difficult to find. Meetings are engineering and supplier driven. Other problems include inter-operability, standards, getting mind time of potential users to configure these smart devices, lack of experienced engineers, proven paybacks, scalability, security, safety (hackers make object weapons of attack?), the lack of suitable energy harvesting for nodes deployed in huge numbers without access and the many system, software and processing issues. Will most data go to the center (IBM, Oracle) or be processed in the nodes (Cisco, Intel)?
The report gives the results of many new interviews and discussions at recent events. These often reveal a more cautious appraisal than that seen in company publicity and analysts' publications selling the dream. IDTechEx carefully appraises how IoT fits into the Internet of Everything, avoiding the common practice of renaming many other things to boost the figures. We consider the reasons why the technology and market first called Internet of Things failed to live up to more than one thousandth of its promise after 15 years. This was the EPC Global system, based on dumb RFID tags. This is important because errors are to some extent being repeated with today's IoT (over promising, tackling the wrong problems first, proposing unwanted applications) partly because different players are involved.
We then assess the realistic potential for today's IoT. We examine and, in many cases, forecast for ten years ahead the allied markets such as RFID. More on that is in our report, RFID Forecasts, Players and Opportunities 2014-2024. Other allied markets analysed include traditional sensing systems, body area networks, mesh networks that are not currently IoT such as ZigBee networks, many of these options connecting to the internet only via backhaul. There are two reasons for this. These other technologies are often good ways of fulfilling certain needs without recourse to IoT and some are targets for conversion to IoT.
Many potential applicational areas are appraised, recognising that it is too early to be certain which will be largest and most profitable and when, though we give some pointers. For example, we give ten year forecasts for wearable electronics which substantially leverages the world of cellphones today but will become substantial for IoT. More on that is in our new report, {Wearable Technology 2014-2024: Technologies, Markets , Forecasts. We consider the considerable number of systems issues such as the newer incompatible-but-improved short-range radio communications systems offered. New collaborations and standards that seek to orchestrate the systems based on the 400 or so lower level standards for incompatible networks and their parts.
There is an excellent investment climate including large companies sometimes overpaying for small ones. The report ends with those revealing interviews we have carried out and meetings we have attended across the world from 2014 onwards, though our earlier experience is, of course, also brought to bear in our analysis. The key conclusion is that IoT will become a large business but later than most people think because so much work needs to be done.