The concept of the internet of things (IoT) is already widely known, and has even been covered on mainstream media like the television programme Carte Blanche. However, the mechanics behind the millions of sensors driving it remains a mystery to some.
In order to fully understand IoT we need to understand that it is in essence similar to machine-to-machine (M2M) communication. These machines are then observed and analysed by humans and the data that is accrued from this interaction (between machines) is ultimately used to help direct business, infrastructure and even resource decisions.
The mechanics of IoT/M2M
The premise of all networking is the assignment of an internet protocol (IP) address to a computer – in whatever form (server, mainframe, desktop, virtual machine etc.). Today in IoT/M2M computing this has been extended to multiple devices, such as cameras, access control systems, instrumentation, supervisory control and data acquisition (SCADA) solutions and the like – all of which become part of the connected network based on the fact that they too have an IP address. The web of devices grows every day and the potential for applications in manufacturing, mining and even retail are enormous. It has been stated that the bulk of future internet traffic will be driven by things and not people.
In order to assign an IP to devices, they have had to become a little smarter, and today you will see the addition of a simple microprocessor to all sorts of “things”. These include your car, a fridge, a credit card, CCTV cameras, machinery in factories and even forklifts (to name a few). The addition of a microprocessor has given these devices the power to communicate back to the system they are connected to. The result is a data rich environment that you can then utilise to develop applications and services, adjust temperature controls for optimal use in foundries, or even to help determine how much power is being used in specific areas and by specific machinery. The application of IoT is endless. Behind this, however, is smart software – typically cloud storage, business analysis tools and dashboards. Hardware is after all still just smart hardware – but software is where the real beauty comes in. As we are starting to embrace the notion of software-defined-everything, we can still put a microprocessor or communication module on any device, but drive the command centre from the data centre itself.
Patterns of usage
This speaks to the concept that computers are no longer the data, but computers are in the data. So the sensors themselves don’t need to be that “smart”, they simply need to be connected. If we unpack the actual data patterns in IoT/ M2M, this is vastly different to that of the patterns created by human Internet interactions. Why? Humans are data and application hungry, whereas sensors and devices in the chain that is the IoT use limited bandwidth to they upload traffic that is biased towards exactly what you want and need. They also use very little power, or can be self-powered – drawing from solar energy as an example. Solar-powered connected traffic lights are a good example of this. While the sensors in IoT/M2M are physical, their application can be as broad as your imagination. In agriculture you may want to attach one to a cow to observe grazing patterns, or place them in soil to measure moisture. In an office it could be to monitor access to data centres, to regulate air-conditioning or to simply advise when new stationary orders need to be submitted. These are all nodes within the IoT and you ask of them exactly what you need from them, whether that be information on location, altitude, velocity, temperature, illumination, motion, power, humidity, blood sugar, air quality, soil moisture… you name it.
Sensors are not computers
A common misconception is that sensors are computers. While some of the more advanced sensors in an IoT environment can be made up of a processor, memory, storage, inputs and outputs, operating system (OS) and software – the majority are “dumb devices” whose sole purpose is to sample, measure, quantify and relay communication via their assigned IP address. The language by which these sensors communicate are not all the same. While the majority rely on wireless communications such as Sigfox, Lora, RFID, NFC, WiFi, Bluetooth (including Bluetooth Low Energy), XBee, Zigbee, Z-Wave and Wireless M-Bus, there are some fixed-things that can leverage Ethernet. Conversely, and this is particularly relevant for African companies where connectivity is not as globally pervasive, many leverage mobile communications such as GSM, GPRS, 3G, LTE, 4G, WiMAX or even satellite connections. Low bandwidth protocols like MQTT, AMQP, CoAP, DASH7 and the like are critical to ensure lowering the cost and increasing the effectiveness of communications – the effective deployment and use of these will most certainly add to data everywhere.
How you network these sensors is entirely up to you. In a small office you may link them to your on-premise server or even a NAS device in your environment. Enterprise customers will more than likely feed the information directly into the data centre to provide information rich resources for use by myriad enterprise applications. Then there is also the cloud, which is fast becoming the preferred means of communication for IoT applications. There are also short and long-range uses of sensors to be considered. RFID scanning devices in a warehouse may use a short wave, typically 868Mhz, communication medium back to central control, whereas mobile connected devices on cellphone masts in remote areas will use GSM technologies.
Data, data everywhere
The IoT is data hungry. So if you are considering its application in your business, you need to look at the state of your existing data centre. Do you have a system that offers and enables scalable, storage and processing capacity? This is where the cloud does become a real option. The application of data is entirely dependent on your business. Data can be stored for analysis later or it can be fed, in real-time, into enterprise applications that will then consume the data for a designated purpose. In order for M2M and IoT to work standards need to be adhered to. Today there are still a number of heterogeneous established and emerging competing technologies and standards that may require you to “stick to single vendor technologies”. However, as a result of moves in the software-defined arena we are starting to see the intelligence move from the sensor to the software, which is enabling for multiple devices, from multiple vendors to run in a product vendor agnostic environment and feed data in central repositories for use.
There is no doubt that IoT/M2M is technology is here to stay, but the industries that have adopted it to its full potential locally are still relatively small. Back in 2004, the magazine BusinessWeek predicted that IoT/M2M would be a $180-billion market by 2008, however in 2007 The Economist estimated it was only around $35-billion. In 2010 Machina Research released a report that stated it was at that stage a $120-billion market. Machina also predicted that the market will grow from $200-billion in 2011 to $1,2-trillion in 2022. Whether IoT/M2M had a false start, or like all technologies, simply followed Moore’s Law, there is no doubt that the technology is taking off. Juniper Research has stated that by 2017 (yes this year) we should see 400-million IoT sensors hooked into the internet. In 2015, Gartner Research stated: “4,9-billion connected things in use in 2015 … and will reach 20,8-billion by 2020. The point is not to obsess with the numbers as these change no matter who you speak to – but rather to observe the trend that IoT M2M is growing and it is pervasive. The local markets where we are starting to see its use becoming more and more pervasive include mining, manufacturing (in particular car manufacturing) and security.
While as a technology driven industry we all aim to benefit from IoT/M2M, there needs to still be a meeting of minds when it comes to standards behind the sensors and the systems. Furthermore, we need to ensure interoperability between systems, vendors need to stop trying to play the “lock-in” game and in South Africa we need more competitive bandwidth pricing in order to really scale IoT’s application. This is a market that will develop – so be sure to carefully navigate your next purchasing decisions to ensure that you have the systems to monitor your devices, the sensors to connect to your network, and the skills to support it.
Contact Netshield, Tel 012 841-0320, email@example.com