Don’t know what to cook today? Ask your Fridge! About the Internet of Things and its drivers.
Disclaimer: This article is written for non-IT people who are interested in this topic and it may contain generalization for better understanding.
It is hard to avoid the Internet of Things (IoT) whenever you read any economic newspaper or publication these days. For starters, let us borrow the definition from Wikipedia to establish a common understanding of the term:
The Internet of Things (IoT) is the network of physical objects — devices, vehicles, buildings and other items—embedded with electronics, software, sensors, and network connectivity that enables these objects to collect and exchange data.
In this article, the focus shall be on the advancements in the four technological fields which led to the IoT as we currently know it, namely electronics, software, sensors and network connectivity. There is also a lot of information out there which deals with the IoT in a less technological way. I can recommend looking at this infographic from Goldman Sachs to gain an overview on general trends and market size.
How we made your toaster intelligent
Let us take a look at what drove the development of the IoT in the first place. Surprise, your toaster didn’t get smart overnight. It was rather a gradual development. The first step toward the IoT were embedded systems. Think of them like mini-computers in all kinds of devices. Your DVD player, your TV and also your toaster – all these devices are embedded systems of varying complexity. The thing with these systems is that they had access to data all along – your toaster could theoretically collect data on how long you toast to get that fabulous crunch. But all that data was useless since the toaster could neither use the data (remember: you handle the toasting time yourself via the knob) nor could anyone access this data. But as the technology advanced, the idea to gather all this info from your Toaster seemed appealing and most of all viable.
The electronics (fruit) salad: Apple, Blackberry and Raspberry Pi
All these fruits are also companies in the electronics sector which commoditized extremely powerful yet potable hardware. Since the days of the first computers in the 60s, hardware got consistently smaller. This is called miniaturization. On the other hand, the computation power doubles roughly every two years. This was first observed by Gordon Moore, who coined the term Moore’s Law to describe this phenomenon. But how could the processing power be increasing while simultaneously decreasing the chips size? Well, the power lies in the increasing transistor density. A processor chip consists of transistors, which are essentially little switches inside the chip (head over to Wikipedia if you want to know more on this matter!). These little switches are what enables the processor to work. Transistors got smaller over time and are now measured in nanometres (pretty small). Today, billions of transistors fit on processors roughly the size of a finger nail. This small size of hardware allows the new iPhone 6S to be as powerful as a Laptop. But today’s smartphones by Apple and Blackberry are just the tip of the iceberg in electronics.
The exciting stuff is happening with Single-Board Computer like the Raspberry Pi. These Single-Board Computers (or SBCs short) contain all the components of a full-sized computer, including USB ports and network connectivity modules for WLAN and Bluetooth. And they sell for roughly 35€ apiece. SBCs can be used as gadgets at home, but it really gets interesting if you look at it in an industrial setting: Industrial plants and their owners possess very expensive and specialized machines. Because these machines tie much money, they cannot be substituted every two years (economists call this asset-heavy). A big part of the industrial machinery is not even digitalized today. SBCs offer the opportunity to digitalize these pre-existing machines very cheap. They allow for very powerful and standardized embedded systems, able to run modern software environments and are therefore easy to integrate into new and existing data networks.
The secret war: Software dominance in the IoT
The increasing computation power has a very positive side effect: It allows even power-hungry programming languages to be used in the IoT. In the past, Java was one of the most commonly used languages for embedded systems. Today, other, web-oriented languages like Python and JavaScript gain increasing traction in the IoT, because these languages are often used in the later stages of data computation and therefore offer advantages in terms of reliability and data structure.
But although more web-oriented and standardized languages gain popularity, one important question remains: How do we organize the communication between devices? We do need standardized platforms to let an infrared-sensor send its information to a vehicle break system. To allow effective machine to machine communication (M2M), we need a platform with a good application programming interface (API) to handle that. And this is where the strategic moves are made: Controlling the platform with its API means controlling access to the IoT and, more importantly, be the interface of consumer to the devices. Remember when Google bought NEST, a smart home thermostat production company? Well, the thermostat was not the interesting part, the platform for the smart home was. For a platform to be accepted by customers, it needs to be widely in use. It needs to support a lot of third-party applications, which is why the open source community is increasingly relevant and frameworks like ThingSpeak are so popular. The direct reaction to this can be witnessed when looking at Microsoft and Apple: Both place increasing emphasis on remaining the relevant interface to the customer more than being a restricted, locked system. Both are taking steps toward the open-source community to secure the support of more relevant third-party developers and integrate them into their system. The big markets of Smart Home, Industry 4.0 and the connected car are still a battleground for companies who try to establish standards. Keep an eye on these secret battlefields!
Yo mama so fat: How sensors and their development weight in
Let us take a moment to think about what a sensor actually is: A sensor is an object whose purpose is to detect events or changes in its environment and then provide a corresponding output (thanks wiki). When taking a look at this definition, what is the oldest sensor that comes to your mind? You may say a weight or a thermometer. Those been around pretty long. In the first half of the 20th century, ground-breaking technologies like infrared and motion sensors came in. After that, the developments accelerated. Light sensors checked if the surroundings are illuminated, physical and chemical sensors got increasingly capable (think of current weather stations), fibre optic sensors allow for an incredible high speed of information transport and even touch screens are sensors. And we barely scrape the surface here.
But let us get to the hot topics, the current sensor trends. We can divide them into four big categories, which we will cover one by one. Optic and sound sensors, communication sensors and biosensors. Optic sensors are currently a very important topic. Photosensitive sensors like cameras predate the 20th century, but these sensors only provided uninterpreted images. Currently, the automotive sector is working at full throttle to allow optic sensors to spatially interpret these images to enable autonomous driving. Although the sensors are not perfect, as the recent Tesla autopilot crash shows, the advancements do look very promising. To spatially interpret the images, algorithmic three-dimensional orientation of surroundings through computing the inputs of several sensors or the use of the Doppler effect are the main research areas. The latter is also used for creating a model through sounds, so called beamforming. Noises provide information about changes in the environment and these can be located by sound sensors. Amazons Echo for example is capable of that. It is kind of like a bats ultrasonic system, but it doesn’t emit waves, just listens to them. The next category, communication sensors, are basically sensors that check the surroundings for devices, mostly with integrated emitters and communication protocols. Near Field Communication (NFC) in the electromagnetic spectrum or Bluetooth and RFID in the radiofrequency are examples here. Especially the cheap RFID and the newly glorified low-energy Bluetooth (BLE) gain increasing traction. These sensor technologies not only allow machine to machine communication, but also give exact spatial information about their counterpart. For example in a library, a RFID-Scanner at the exit checks for the location of RFID-chipped books which were not checked out (aka stolen). In another case, an iBeacon system may provide indoor navigation at an airport directly to your smartphone. Our last category are the biosensors. Think about any smart wearable like FitBit, Apple Watch or even tech tattoos like invisibles. They scan blood pressure, oxygen level, temperature and many other vital stats of your body. These sensors got a lot cheaper throughout the last years and are currently powering the digital consumer health sector which all major pharmaceutical companies invest heavily these days. Other pioneers are for example the SCiO nutrition scanners, which measure nutrients in food optically. Potentially, blood analysis may be handled through a wearable device and analysis labs will get obsolete. Just imagine all your favourite TV detectives running around with these things.
But taking a brief glimpse into the future, organic sensors may be the long term research goal. All the sensors we talked about are anorganic, but since every living organism has several highly-effective sensory systems which interact perfectly, a future with organic sensors is possible. Maybe a dogs’ nose or a bats hearing can be reconstructed. Imagine a living eye as a smartphone camera. All those possibilities…
Home is where the WiFi is: Internet and network connectivity
The introduction of the internet in 1993 was probably one of the most important milestones in the history in mankind. A quarter century later, it seems fair to make this statement. The internet of things is not thinkable without the internet itself. But of course the internet changed during all this time. Today, one of the biggest issues with the IoT is the increasing number of devices on the internet, all of which need a unique address. Imagine it as a postal address in a city with an increasing number of houses, but where there are no further street names and numbers available. That is a problem if you try to send a letter to one of these houses. Therefore, the old addressing system, the IP Version 4 is about to be substituted by the much larger IP version 6 (IPv6) system to make enough addresses available for the future. However, the deployment is taking longer than expected, considering it roots in 1996.
But not only the address system changes, the access to the internet lost its’ cable connection: Wireless connections (WLAN) at home and mobile internet allowed increasingly more devices to gain uncomplicated access to the internet. The current standard WLAN is the 802.11n with a transfer speed with up to 600 Mbit/s, the newer standard 802.11ac will offer up to 6.9 Gbit/s, transferring effectively ten times more data per second. These high transfer rates are especially important for transferring big amounts of data through in real time, for example to allow stream analytics. The mobile Internet has developed from the lousy 2G with a speed of 9.6 kbit/s to the currently halfway decently covered 4G with a speed of up to 300 Mbit/s. But the next step towards 5G is not focused on speed. Although it is said to increase to up to 10 Gbit/s, the main point for 5G is to enable more devices to be connected simultaneously and be more efficient in doing that. Also the coverage of more areas, especially more rural ones, is one of the goals. Next to 5G, the Bluetooth Version 5 protocol is planned to be released in the next time, also focussing on an increasing number of devices, double the volume for data and four times the range of the current Bluetooth standards.
But there are also completely new concepts out there, for example firechat. This instant messenger follows the completely decentralized idea of a network of clients without the need for an internet connection. Messages are handed through the network of all smartphones which have firechat installed, one connected to the other via WLAN or Bluetooth. Of course this idea is more fitting for urban areas with a high density of potentially connected devices and may have problems to send messages bridging large, only minor populated areas. But the idea is very intriguing.
The Future of the IoT
As you can see, an incredible amount of research and development was what enabled the Internet of Things in the first place. But where do we go from here? What overarching visions do we have for the IoT? Let us try to envision the transportation system of a future city: Imagine a world, where traffic in would be fully autonomous. You would simply call an autonomous Uber whenever you needed transportation. A car would share out of the ever flowing traffic and pick you up, just to get back into the ever-flowing stream after dropping you off. Traffic lights would not be necessary any more. The grid would be filled by demand. There would be no defect cars, as they predict their own failures through complex sensory systems and would get to repair before defects occur. Transporting goods would no longer need separate vehicle like trucks today, but would load goods onto every car in the system, providing a more flexible logistics system.
Just an idea. Let us see what will happen.
Back to the overview on all Big Data topics
Thanks to 8icons for the icons!
