The Internet of Things (IoT) is a massive area of growth and investment in IT currently (as noted by Gartner), but what do businesses need to do to take advantage? In this blog, I describe what IoT is, and how organisations can take advantage of it as for many, gaining the benefits of IoT will not be derived from directly implementing all aspects of an IoT architecture but from integration with their “Things” providers.
So what is IoT? IoT is having devices in situ and being able to connect these to a network and exchange information about their status and receive updates. “Things” is a deliberately general word – it can be anything from rays of sensors in cars or under desks (as used in Keytree’s Matrix Booking product), factory machinery, vehicles and even in one solution developed by Keytree – monitoring a herd of dairy cows in a field. The power in IoT is it’s two way – it’s not just gathering data from sensors on “Things”, it also can send commands back to react.
The power in IoT is it’s two way – it’s not just gathering data from sensors on “Things”, it also can send commands back to react.
Understanding the “Things” that pull it all together
“Things” have smart sensors and control boards, and the data from these are (often) collected on an edge device, e.g. sensors across a floor of a building to a central hub – specialist low energy communications protocols are commonly used here. These devices then connect to a cloud IoT platform but with large data sets, an approach called ‘lambda’ architecture can be applied – so data is processed in batch and stream modes.
In stream mode, real-time decisions are made against the data as it arrives, and this triggers alerts or real-time feedback. Batch mode is used to analyse data over time, as data is aggregated and stored for offline analytics. This is where IoT is enabled by other key IT industry trends such as big data and machine learning.
IoT specialist areas range from low energy networking, big data, AI, and stream processing but does it mean an organisation needs to start building up skills in these areas to get the benefits of IoT? Not necessarily. Firstly, platforms and tools are maturing (I’ve previously blogged about SAP Leonardo, and other cloud providers such as AWS also have offerings). Secondly, many organisations will be consuming IoT data, but invariably outsource the collection and management of this information.
An obvious example is a car manufacturer fitting sensors to the vehicles it builds and collecting data from them. Others would include real estate managers or machinery providers. It makes sense as the more closely the IoT sensors and controls are embedded in the product the more effective both can be. Also, for areas such as predictive maintenance – based on long-term data trends – the more significant the data set, the better as equipment manufacturers can gather the largest data set.
It doesn’t mean that other companies don’t have to engage with IoT, they do, but concerns become focused on service level agreements, data ownership and integration with the IoT systems of their “Things” providers. For example, do you have access to the telemetrics from your equipment from the manufacturers of that equipment? Who owns the data, and what can they do with it? If you have access what is the level of aggregation and how do you receive it? Do you merely have a manufacturer-provided portal or can you get raw data via an API? Does the data conform to standards and does your provider keep the information secure? What happens when equipment moves between countries with different data protection regimes? How is the data the used in the companies systems of record?
There are obvious similarities here with the approach to Software as a Service systems such as cloud-based HR apps – critical elements of a company’s data are held by a third party, and there is integration with that, supported by SLAs and data protection agreements. So when acquiring IoT systems procurement – legal, enterprise integration and service management teams will be more critical than experts in low power networks for many, if not most, organisations.