To read the full article, learn about JOS Recommendation and receive updates, please register.
IoT is a Data Problem
IoT puts enormous strain on data infrastructure. Small data, produced quickly and streamlined continuously, can strain unprepared infrastructure.
Sensors also come with various protocols and emit different data types. And they produce two distinct types of data: data from the sensor about the location/machinery/application and data about the sensor itself. The latter is equally important to understand whether the sensor is working correctly.
So, before jumping on the IoT bandwagon, it is prudent for companies to first see whether their data management architecture is ready.
Bridging IT and OT
The role to manage data architecture often lies within the information technology (IT) department. In many companies, IT and operational technology (OT)—technologies to support operations, like air-conditioning units in building management or medical devices in hospital—stand apart as two separate units. But in IoT, they need to work together.
While IT offers knowledge in data and network architecture design, OT is required for knowledge in sensor data accuracy and sensors management. The challenge often lies in getting IT and OT to work together for integrating sensor data to the enterprise systems and business logic.
When there is a missing link of IT and OT, data collected from the sensors may not be able to feed into the enterprise systems for analysis. It could be simple problem like sensor placement or more complicated issue with sensors protocol management.
CIOs need to ensure that both departments start collaborating to manage IoT systems end-to-end. But a strong middleware that brings interoperability between sensors with the underlying IT system is a good place to start.
The IoT industry is still evolving. New devices are always coming online, and companies are experimenting with specific IoT use cases. Betting your entire IoT future on a single sensor or network technology may solve your immediate problems, but can constrain your ability to grow, adapt and use new devices. Adopting a best-of-breed approach by deploying a modular IoT middleware allows you to avoid this situation. It also allows you to build an IoT blueprint that is aligned with your company’s ambitions.
Get an Edge
Many view Edge IT—which proposes moving some compute power to the edge of a corporate network and nearer to the operational location—an evolutionary step for IT. For IoT, it is becoming a mandatory one.
“Is your culture ready for IoT-based real-time monitoring? The answer will determine your approach,” said Michael Lee, Senior Consulting Manager, JOS.
The reason is simple. IoT data, especially in B2B environments like manufacturing plants, hospitals and smart cities, requires almost instantaneous analysis of data. While new networking technologies like NB-IoT, LoRa and 5G can speed up the data delivery, you still need compute power to process them into insights. Moving compute power nearer to the sensor can reduce the volume of data traffic, speed up the application performance.
IoT is not a standalone solution. It needs to part of one. For example, autonomous vehicles need IoT to ensure that the onboard artificial intelligence communicates with other sensors and navigate the best path. Predictive maintenance use machine learning to analyse sensor data and other data to proactively upkeep machinery. Smart buildings rely on sensor data to understand behaviours for enhancing security and convenience.
These applications all another layer of complexity: instantaneous response. This makes the deployment of analytics capabilities necessary at the sensor edge.
Companies looking to maximise their IoT initiatives will need to start re-architecting their internal IT architecture for Edge IT. It will allow the sensors to deliver instantaneous responses that future innovations will depend on.
Rewire your Security
In IoT, data flows in an expanded enterprise network across multiple sensors. If your system is connecting with sensors that are managed by external parties, data will flow between networks. Essentially, it expands your risk surface.
These sensors can be used to spy, create havoc or even control of the enterprise systems like online traffic lights becoming compromised or brakes of autonomous vehicles becoming hacked. A clear data security policy that extends beyond the own corporate network is required in an IoT architecture. Encrypting the data is just the first step; also needed is a security by design approach at the application development stage.
Sensor data can also impinge on privacy regulations. In some countries, for example, regulators view electricity data as customer information, applying the same laws to govern their use. So, companies need to pay close attention to potential abuse and their privacy risk exposure.
Lastly, sensors require maintenance and upkeep. If sensors go into disuse or are tampered with, the data collected would be inaccurate. This is an area of physical security that many companies overlook.
IoT is means to an end
IoT answers today’s demands for instantaneous, dynamic market landscape and raises the bar in operational efficiency by collecting vital data that improves your situational awareness.
“Customers are not looking for technology but solutions to their business challenges,” said Lee.
So, the best way to start your IoT journey is to understand the problem you are trying to address. Then, you can see how IoT can help – not the other way around. Having a strong partner who has the local support, understands your business and part of the IoT industry evolution to create the best-of-breed solution will be a huge plus.
How do you feel about implement IoT in 2019?
How do you feel about managing data from IoT sensors?