What's IoT All About Revisited
I often open with a quick, tongue-in-check definition of the Internet of Things (IoT)
"You have this thing called the Internet. You have all these other things that you use everyday, in every aspect of your life. When those things, start using that first thing, you have the Internet of Things."
Of course, the details of what “use” means is where things get interesting.
The first time that I wrote an explanation of the Internet of Things (IoT) was in 2012. Since then, I went a bit deeper and, over the next few years, we provided a discussion as an introduction to the IoT to university students and also to corporate R&D labs and to government agencies and advisors. It is time for an update. The Internet of Things brings a compendium of concepts, processes – both business processes and best practices operating procedures – and people skills that are generally applicable across every aspect of our lives: government, every industry, healthcare, entertainment, vacations, and home activities. We go into the steps required to begin and explore the IoT in our 5Cs Maturity Model. There have been many predictions about the scale of the IoT. All of these have been based upon the number of devices connected. The truly important number though is that there are now over a trillion connected sensors, and trillions of new data streams. All of this data is becoming available to augment humans and to collaborate with other things, in making better decisions.
Unfortunately the IoT has been developing in silos, and still ignoring the potential to be gained by properly managing, sharing and analyzing IoT data throughout ecosystems. We are, however, starting to see some calls from others to bring together IoT data and sensor analytics from various parts of organizational units, and from various members of an organization's ecosystem. This is the minimum that is needed. For a true IoT to exist, responsible and ethical data sharing will need to be accomplished.
An excellent example of how the concepts associated with IoT transcend industries is that of digital twins or avatars. The concept of computer simulations of a system is not new, but the wealth of sensor data and computational ability and machine learning algorithmic advances have allowed these simulations to provide real-time updates as to the condition of the physical system. In addition, feedback between the sensor data for the physical system and the digital twin allows great improvements of our understanding of the physical system, the environment around it, the duty cycle, and the science governing these interactions, allowing updates to the machine learning algorithms to improve their predictive ability. What makes this a great example of IoT concepts being useful across different industries is that digital twins are useful in smart cities, manufacturing, autonomous systems (such as self-driving cars, drones or robots), agriculture and healthcare. The same digital twin concepts for a manufacturing plant can create an avatar of a unique human being, or of a field in a farm, or the interactions of an augmented city with its citizens. Predictions can be made for maintenance of the production equipment, or irrigation/fertilization of the field, or preventing that heart attack, or location in a city related to attitude and health of a citizen or experience of a tourist; predictions can be made for product quality from any area. I think that the IoT concept of digital twins shows the importance of cross-fertilization of ideas across industries and across disciplines…the importance of system thinking. This type of thinking, with the sharing of both ideas and data, among disciplines, industries, organizations, business units and departments, is required for the IoT to meet its promise. But economic factors, such as industry dynamics and market forces, as well as cultural precepts, often come together to prevent this, both in that there is no one, or no group, providing IoT system and ecosystem oversight to an organization, and in sharing of data. Privacy, security and transparency also come into question for sharing of data; convenience of using the IoT service or product also impacts how easily the data can be properly collected and shared.
The IoT holds the promise of vastly improving the efficiencies of the supply chain, as one example. But if rather than one IoT, we see an Industrial IoT and a consumer IoT seen as two separate concepts, two separate implementations, each with their own multitude of sub-genres and silos, than the idea that an efficient supply chain based upon environmental and duty-cycle data from the consumers, as well as plant maintenance, warehousing and transportation data from the industrial side, would never come together. If the trucking company transporting the product can't share their monitoring data (such as temperature of the crates being carried) than important information that can impact quality of the product, is lost. More future looking ideas, such as inexpensively providing customized one-off versions of products that exactly meet the customers needs would certainly never become real. But the technology exists today where all of these ideas are, indeed, achievable.
While the term, IoT, only dates back to 1999, when Kevin Ashton coined the term, the concepts go back much further. The earliest may be Nikolai Tesla working on his remote control boat in 1898. Other examples that may be more pertinent to connected sensors transmitting data are the home automation attempts and NASA biotelemetry in the 1960s. Today, IoT concepts exist everywhere:
- connected pens uploading the words you write to the cloud and providing local text analytics to translate handwriting to search-capable printing
- fitness bands becoming health monitors
- predictive maintenance for every product
- integrated architecture controlling irrigation and fertilization of fields through artificial narrow intelligence (AnI)
- augmented cities responding to the challenges of traffic and parking, and the increasing urbanization forming megacities
Some of the areas that have led to the IoT being so popular today include
- decreasing cost and size of sensors, micro controllers and microprocessors
- advances in data management and analytics including machine learning (ML), deep learning (DL) and artificial intelligence (AI)
- elastic infrastructures in the Cloud and at the Edge decreasing the cost of computing
- the rise of social media, smart phones, tablets and voice AnI assistants
That last one may be surprising, but it is the acceptance of powerful computers in our hands, the idea of always being connected to humans around the world, the novelty of chatbots, the usefulness (and frustration) of voice assistants, that have led us to expect all of our things to be connected and even to anticipate our needs. We are also seeing the downside of this as social media platforms continue to make us better targets for advertising and politics. Elastic compute infrastructure has led to dropping computing and data storage cost, thus allowing the collection of the extreme amounts of IoT data and the compute power for sensor analytics, both streaming and combining with historic data from IoT and other systems. And while the cost and size of sensors, microcontrollers, mircroporcessors and all types of electronics has been decreasing, their capabilities have been increasing dramatically.
The IoT is one, very large complex system…a system of systems…an ecosystem. For literally everything to come together as a true Internet of Things, we need overlapping sensor analytics ecosystems (SensAE). The Internet was transformative as it made everyone a producer and a consumer of knowledge, opinions, thoughts, literature, nonsense, photos, videos, art, and products. The IoT must come together as overlapping SensAE so that each and every thing, animal, plant, human… become a producer and a consumer of data, and the insights and predictions that may be drawn from those data sets. Even if IoT doesn't come together, the various silos do produce value now. We focus is on how evolving IoT implementations come together with advances in data management and analytics, especially in data preparation, data science and data anthropology. The most valuable result of this conjoining of disciplines is in predictive algorithms that continually learn to provide ever better predictions combining streaming and persisted data sets. There are several really good examples of where this is happening today.
- predictive maintenance of everything from ATMs to hot tubs to factory equipment
- healthcare becoming personalized, predictive and preventative (P3Health) as well as looking more at population health effects
- traffic and parking in campuses, cities and megacities
- railways benefiting through fuel economy and preventing derailments
- airlines benefitting through fuel economy and reduced aircraft downtime
- elder care improvements through telemedicine, biotelemetry and robotics
- field service and warranty management
- telemedicine, mobile health (mHealth), robotics and AnI extending healthcare to the poor and remote areas of the world
But with all this, the true reason to be excited by the IoT, is the potential to make things better globally, though there is also potential for great harm as well. While I don't believe that neither a utopian nor a dystopian future will be the ultimate outcome, I do believe that we are inching, incrementally, towards better and better lives for more and more people. The convergence of the Internet of Things with evolving data management and analytics technologies and processes can bring together various solution spaces into a sensor analytics ecosystem to solve global challenges brought about by climate change, population increases, the aging of the human population and concentration into megacities, tracking endangered species, fisheries and forests, and the vast changes in agriculture and supply chain/logistics.
One purpose of this site is to delve deeper into those concepts, technologies, processes and skills that are shared among all current industry silos that are called "IoT". As I mentioned above, a starting point can be found in our IoT 5Cs Maturity Model. Another important area can be found in an older blog post on data quality. You can listen to more on a podcast [on iTunes] [or listen to this podcast on libsyn] where I was interviewed on my high level views on IoT.