IoT devices send data to IoT Hub lub IoT Central, which are the centre of the IoT architecture. They are used to manage sensors and gather data.
Next telemetry is sent to other Azure services such as Stream Analytics to analyse and process live data. In the end, the data is gathered in the Azure SQL
database and visualized in Power BI
Telemetry data can be sent directly to the cloud, or it can be processed on the local Edge device. This solution is used to limit the amount of
data sent to the cloud. In the Microsoft environment, this service is called Azure IoT Edge.
IoT is a set of connected technologies that could be characterised by those three elements:
Things – physical sensors that are sending telemetry data to the Internet.
Insights – processed data sent by sensors; many services can be used to process and visualize data, such as Stream Analytics,
Power BI, Azure Functions, and many others.
Action – the outcome; decisions can be made not only by humans but also through the automated processes of IoT service - IoT Hub can communicate
with devices and sensors bidirectionally.
The simplified IoT architecture is shown below:
Things are the IoT devices - sensors or machines equipped with sensors that can be connected to the cloud or an Edge device that processes data locally and then sends it to the cloud.
Insights includes various services such as processing services (Stream Analytics, Azure Functions), storage services (Azure Blob, SQL Database), and data visualization (Power Apps, Power BI, Azure Time Series Insights).
Action is where decisions are made.
Let us focus on describing Azure IoT services used by Antdata. The most important part of the entire infrastructure are those services that integrate the physical device with the cloud. In the Microsoft environment, they are called IoT Hub and IoT Central. Both systems perform the same tasks, but there are fundamental differences between them.
IoT Hub – it is a PaaS environment, acting as the centre of the IoT infrastructure, which allows for two-way communication between the Hub and IoT devices. It ensures a secure connection with millions of IoT devices. It enables handling large amounts of data and monitoring connected sensors.
IoT Central – performs the same task as IoT Hub, but it is a SaaS environment, which means it is a cloud application. It still allows for the integration of multiple sensors, two-way communication, and the creation of device patterns, but it does not have that flexibility in personalizing solutions based on specific needs. It facilitates the application of the IoT solution and shortens the implementation time.
A very important part of the IoT implementation is making sure that the transmitted data is secure. Azure IoT solution was created with security in mind. The Microsoft infrastructure used by Antdata protects the entire architecture - sensors, data, IoT Hub, but also services responsible for data processing, storage, and visualization.
The following products are responsible for data security:
Azure Sphere – services and products that increase security by connecting sensors with Azure Sphere OS and the cloud. Azure Sphere also lets you use new IoT devices on certified Azure Sphere microcontrollers.
Azure Device Provisioning Service – ensures the flawless connection between devices and the cloud. IoT devices are automatically connected to the IoT Hub with a secured connection.
Azure Security Center for IoT – provides insight about the solution security for IoT implementation managers.
Another solution we mentioned earlier is an edge device. In the Microsoft environment, it is called Azure IoT Edge. It is part of the IoT Hub, functioning as a physical device
that collects data from sensors locally, processes them, and then sends them to the cloud. You can say it is an IoT Hub deployed on a local device. It can also consist of multiple
modules, like Stream Analytics. This solution allows you to save up some storage space in the cloud and perform all the necessary data processing operations on the local device.
The following Azure services are most often used to perform data operations:
Stream Analytics for real-time event analysis and processing; performs operations and tasks on large amounts of data streamed from multiple sources simultaneously. It can run on the IoT Edge device.
Azure Time Series Insights – provides data handling and visualization, allows for anomaly detection and analysis.
Azure Functions - an event-driven serverless computing platform that can also serve for data processing purposes
Power BI can show both live and historical data in the same report. In the Power BI service, the real-life telemetry data can be displayed on the dashboard in the form of tiles and visualizations.
A diagram showing an example of Azure IoT architecture:
The benefits of implementing IoT solutions
The implementation of IoT allows organizations to increase operational efficiency and safety in the workplace. For a production company, IoT brings
increased production efficiency and may become the next stage in the development of Lean Manufacturing strategies.
When talking about Lean Manufacturing, it is impossible not to mention Kaizen - the concept of continuous improvement. If you think about it, IoT is a constant strive
for improvement and optimization of processes. The implementation of Lean Manufacturing involves optimization based on human experience and manually collected data.
As a result, huge amounts of data generated during the production process are neither captured nor analysed and potentially valuable information is wasted.
Lean Manufacturing is about minimizing waste and minimizing data and information waste is part of it. Gathering data from sensors allocated at different stages
of production may lead to some relevant observations and can help to take the concept of waste elimination to the next level. Digitization of machines and controlling
their work supports the analysis and decision-making process at each level of the organisation.
Centralised Business Intelligence system transforms data into useful information, which when appropriately analysed, directly contributes to making better decisions.
Integrating live data coming from the production line with historical data and insights from other departments of the company creates one efficient cooperating environment
and enables profitable improvements.
An example of Azure IoT architecture in a company in the plastics industry
IoT solutions can be implemented in a wide range of areas. An interesting example is the implementation of sensors for intelligent monitoring of the production
line that consists of injection moulding machines.
In the case of a batch non-rhythmic production system, the entire technological process is performed on one machine - an injection moulding machine and the products are
manufactured in batches. The most important parameters in monitoring this kind of production are the number of manufactured products and the cycle time of producing one
product - counted from closing the mould, through injecting plastic, to opening the mould.
The implementation of the IoT architecture begins with the installation of a sensor (IoT device) on the injection mould that will be responsible for counting the mould
opening and closing cycles. Thanks to this solution it is possible to capture how many times the mould was closed, i.e., how many products were produced, and to calculate
the time between closings - cycle time. In addition to that, inside of the container with the moulding material, there is an intelligent scale installed that sends telemetry
data about the amount of plastic that is left. Knowing the weight of the product allows for calculating the indicator which shows how many products can be produced with the
In the presented form of production, one operator very often manages several machines. Sending him appropriate alerts at the right time can significantly improve his workflow.
The data sent by the sensors is integrated within the Azure cloud.
Thanks to the use of Stream Analytics and AI it is possible to detect anomalies in the cycle times, e.g., a one-time jump during the cycle or a very short cycle
time. IoT Hub, on the other hand, is used to monitor the data and to control the operation of the sensors. Notification Hub is used to create an alert system. In the case of
any anomaly, the system sends a notification to the operator with the information that the process is unstable and it is necessary to check what is happening. Another alert is
informing the operator about the amount of the material left in the container and it is being sent when the plastic amount reaches a critical level. The last rule of the
notification system is responsible for sending an e-mail with the batch completion protocol (data obtained from a sensor counting the number of cycles). The message is sent
directly to the warehouse (informing that the batch is ready for transportation), the production coordinator (informing whether the production plan is running correctly),
and to the team responsible for retooling the machines (shortening the changeovers time - SMED).
Of course, further development of this IoT solution is possible. The extension can start with the implementation of more sensors, but in this case, an interesting case would be
integrating the entire system with the Power BI reporting, Power Apps
and Antdata’s BI Kiosk. . The data collected from each injection moulding machine could be visualised in the Power BI report to which
an operator and a production manager would have an access to. On the screen of Antdata’s BI Kiosk, they would be able to select the machine or the production order to check the
production progress of the particular batch or the state of the machine. The Power BI report would display historical data and live data about the level of plastic material
left in the container and about the downtime. To take the IoT system even further, we could implement the Power Apps application, in which the production manager could log
information about the next order and assign a specific machine to it.
Why choose Microsoft Azure IoT Infrastructure?
The Azure platform is a ready-to-use environment created by Microsoft. The multitude of services available in the Azure cloud allows for an individual approach to each
project and a truly customised end solution.
The biggest advantage of the Microsoft-based solution is that the environment is ready, so the implementation cost is lower than in the case of solutions programmed from scratch.
The programming approach also takes much
more time and requires more complex processes and high commitment of the company's employees.
It should also be noted that thanks to the use of the Azure IoT platform, it is very easy to integrate other elements of Business Intelligence, such as Microsoft Power BI,
Power Apps, Azure data architecture or Power Automate.
This holistic approach eliminates compatibility issues and ensures a fully operating robust ecosystem.