- ✓Arduino and Raspberry Pi are popular platforms for IoT prototyping, with Arduino suited to simple, real-time control tasks and Raspberry Pi to more complex, software-heavy applications.
- ✓Cloud IoT platforms including AWS IoT Core, Google Cloud IoT, and Azure IoT Hub provide managed services for device connectivity, data ingestion, and analytics.
- ✓MQTT's publish-subscribe model decouples producers and consumers of data, making it highly scalable and suitable for large numbers of connected devices.
- ✓REST APIs are commonly used to expose IoT data to web and mobile applications, providing a familiar and widely supported interface.
- ✓Iterative prototyping, building and testing in small increments rather than attempting to build everything at once, is particularly effective for IoT development.
Listen to the full episode inside the course. Enrol to access all 80 episodes, plus assignments, tutor support and Student Finance funding.
Start learning →Alex: Today we're looking at how to develop an IoT application. Sam, let's start with the hardware side. What platforms do developers typically use?
Sam: For prototyping and student projects, Arduino and Raspberry Pi are by far the most accessible platforms. Arduino is a family of microcontroller boards based on Atmel AVR or ARM processors. It's programmed in a simplified version of C++ and is ideal for simple, real-time control tasks: reading sensors, driving actuators, managing timing. It has an enormous ecosystem of libraries and shields, add-on boards that add specific functionality.
Alex: And Raspberry Pi?
Sam: Raspberry Pi is a full single-board computer running Linux, which makes it far more capable than Arduino but also more power-hungry and more complex. It can run Python applications, connect to the internet easily, host web servers, and run machine learning inference at the edge. For IoT applications that need significant processing power, a rich operating system environment, or complex networking, Raspberry Pi is an excellent choice. For simple sensor reading and control tasks, Arduino is often more appropriate.
Alex: How do these devices connect to cloud platforms?
Sam: Most cloud IoT platforms provide device SDKs for common platforms. AWS IoT Core, Google Cloud IoT Core, and Azure IoT Hub all have SDKs that handle the complexities of secure MQTT communication, device registration, and certificate management. The device authenticates to the cloud using X.509 certificates, which is more secure than username and password. Once connected, the device can publish messages to topics, and cloud-side rules engines or stream processing services can react to those messages in real time.
Alex: Can you walk through a simple example?
Sam: Imagine an IoT application that monitors greenhouse temperature. A temperature sensor connects to an Arduino or Raspberry Pi. The device reads the temperature every minute and publishes it as an MQTT message to a topic on an IoT cloud platform. A cloud-side rule triggers when the temperature exceeds a threshold and sends a notification to the greenhouse operator. Historical temperature readings are stored in a time-series database and displayed on a dashboard. This is a simple but complete IoT system covering sensors, connectivity, cloud processing, storage, and visualisation.
Alex: What are the key software development challenges?
Sam: Error handling is critical: what does the device do if the sensor fails, if connectivity is lost, or if the cloud platform is unavailable? Firmware update management: how do you push updates to deployed devices securely and reliably? Power management: for battery-powered devices, putting the processor to sleep when not actively measuring and communicating can dramatically extend battery life. And testing: testing IoT software is more complex than testing a web application because you have hardware dependencies and real-world environmental factors to contend with.
Alex: Any advice on an iterative approach to development?
Sam: Absolutely. Start with the simplest possible version that demonstrates the core concept: just the sensor, the device, and a connection to a cloud platform, with data displayed in a simple way. Then add features incrementally: alerting, historical data, actuation, user authentication, additional sensors. This iterative approach lets you validate your technical assumptions early and avoids the trap of building everything before discovering that a fundamental assumption was wrong.
Alex: Thanks Sam. Next we look at evaluating IoT applications.