Gavin Payne is a principal architect for Coeo, and a Microsoft Certified Architect and Microsoft Certified Master. His role is to design, deploy and fix Microsoft data platforms.
Sensors and devices that connect to the Internet of Things (IoT), and the services to manage them, are now appearing left, right and centre. Real-time activity and environment data is being captured anywhere, from factory machinery to domestic kitchen appliances, and stored in cloud services. Yet the value from these great networks of data generating devices – managed by global command and control networks – comes from analytics as much as their infrastructure. We have to join the thousands of interconnected dots together to make a picture that shows what’s going on.
It started with a racing car team
Systems that collect and analyse data are nothing new, we used to call it telemetry and sometimes still do. When racing car teams drive test laps, computers monitor and analyse everything the car does. Captured values can be compared to theoretical maximums and clever maths used to calculate unknown values – like how much Russian rain slows an F1 car down. While this example has been used many times over the years, it’s probably still true.
And often ends with modern analytics
It’s also a good template about how to use the Internet of Things. When terabytes of data are being collected from around the world using IoT networks, it needs to be stored near an analytical toolbox. An individual sensor reading on its own probably isn’t of much value. Combine it with millions of other sensor readings to spot historical trends and you have one of the principles of big data. Pass the same data through computational models to predict the future and you have one of the principles of machine learning. It’s perhaps no surprise then that the home of most future IoT solutions will be with cloud service providers that offer IoT sensor management, mass data storage services andadvanced analytical capabilities.
But sometimes misses the point
Connecting sensors and devices to the internet on its own isn’t a new capability, neither is remotely controlling a device over the internet. My home’s central heating system might be connected to the internet, it also collects some real-time data, but despite its claims it’s not futuristic. I can’t compare historic outside temperatures to inside temperatures even though both values are being captured. It can’t predict how long my house will take to warm up tomorrow even though it monitored it warming up yesterday.
Technologists need to use the Internet of Things to support a period of innovation, not use it as a way to re-brand yesterday’s wares.