The Internet of Things (IoT) describes a networked computing system where applications intelligently monitor and control remote sensors, mobile devices and smart machines, and where devices such as actuators, valves and switches are connected and communicating.
The resulting data can automate processes, eliminate inefficiencies and accelerate business innovation. IoT-driven intelligent systems are actually a ‘system of systems’ deployed in cloud, datacentre and field environments. By gathering and analysing massive amounts of data, this wave of increasingly intelligent systems can help organisations make more accurate and well-informed decisions. And by intelligently controlling remote devices, next-generation applications will help improve automation and optimise business processes.
Until now much of the discussion about the Internet of Things (IoT) has centred on consumer products like wearable devices, smart cars and appliances. These devices, aimed at delivering better control over consumers’ lives, are provoking thoughtful discussions in boardrooms and strong investment from venture capital firms. This is unsurprising given the natural fascination with ways to use technology to improve our lives. The rapid development of futuristic technology gives hope that solutions to complex problems and challenges can be found.
The Possibilities Are Endless
When it comes to an IoT approach in the enterprise, the full range of possibilities continues to grow and the impact of these technologies is likely to fundamentally revolutionise the way enterprises do business. Some of the resulting advancements in productivity and flexibility can almost compare with the advancements seen thanks to the advent of computing itself. The process has been prohibitively expensive for most organisations—until now—but the evolving economics of the IoT architecture make it possible for every business to benefit.
But enterprises can only benefit from the IoT if their architecture is fast and reliable enough for instant information and decision-making.
Distributed devices can gather invaluable data, which can be analysed for better decision-making and revenue-generation. For example, rail transport can be transformed by an IoT approach. In the traditional model, a train driver is in control of the train’s speed and manually adjusts it if another train is too close or there is a reason to slow down or speed up. These decisions rest with the individual drivers and there is room for errors and inefficiency. Using an IoT approach, the rail company can operate a centralised system that knows exactly where all the trains are at any given time. It can automatically adjust the speed for safety and to improve things like timetable efficiency and resource consumption.
Because of the mission-critical nature of IoT in the enterprise, there need to be fundamental differences in the architecture.
For consumer-driven IoT use cases a two-tier architecture is typically used. The device itself (the first tier) connects directly back to a cloud- or datacentre-based service (the second tier). The analysis occurs at the datacentre and, if action is required, then this is transmitted back to the device. Because consumer devices do not usually require time-critical decisions, this process does not negatively impact the application’s performance. For example, if a consumer uses a smartphone to instruct the home thermostat to turn up the air conditioning, it is usually not a problem if it takes a minute or so to complete the action.
This is not true in an enterprise environment where decisions can be measured in fractions of a second. Taking the example of the automated train control system, a minute’s delay in reducing the speed of a train can result in a catastrophic accident. Or if the under-voltage sensor in an electrical grid took minutes to bring additional capacity online when voltage started to drop, entire power grids could go down and the costs to industry could be massive.
Another consideration is the bandwidth required for IoT applications. When the consumer is paying for the bandwidth it doesn’t matter how much bandwidth the application requires.
By contrast, in an enterprise every byte counts and shaving just a single byte from a message can save the organisation millions of dollars in transmission costs in industrial IoT use cases. So it is crucial for enterprises to consider bandwidth implications when designing IoT applications.
With these considerations in mind it becomes clear that the traditional two-tier architecture is simply too slow for important data and too expensive for unimportant data. A better approach is to build a three-tier architecture that includes a new, functionally capable middle or controller tier.
Adding this third tier between the device and the datacentre effectively brings datacentre functionality closer to the device. It can collect, analyse and take action based on data from its connected sensors and devices. The middle tier must be smart enough to take the required action quickly while sending only the most important summary data back to the main datacentre.
This concept is known as Near Field Processing. It lets decisions be made as close as possible to the edge of the network, so less data has to travel all the way back to the main datacentre. This minimises transmission costs and lets enterprises go from data to decision faster. The reduced time and bandwidth means that the IoT can become a business reality for enterprises.
In the past, organisations required very deep pockets to benefit from the process of gathering data from distributed devices, analysing it and using it to drive better decision-making and increased revenue. The economics of the IoT architecture are bringing these benefits into reach for more and more organisations thanks to cheaper hardware, the ubiquitous nature of connectivity, the increasing acceptance of big data and its analysis, and the ability to develop new architectures. With the right approach, every enterprise can benefit from an IoT approach.