Azure IoT hub, cloud to device connector

But not every business is ready to handle the complexity of IoT on their own, especially when it comes to taking advantage of edge computing, digital twins and AI to make the most of IoT investments.

Microsoft’s recent IoT Signals report reveals that the most common reasons for IoT projects failing to make it from pilot to production systems are the cost of scaling up and not having the right infrastructure or expertise in house. It’s one thing to design a workload and push it to a handful of devices; it’s another to manage a solution at scale across many, many devices which have very limited local resources.

Once you know what you want IoT devices to do and plan how you’re going to use the data the sensors deliver, you need an easy way to deploy, provision and manage those devices. Azure IoT Hub is a managed service that simplifies connecting, monitoring and managing IoT devices at scale, from provisioning and updates to sending and receiving data (and routing messages to other Azure services like Event Grid, Functions and Power BI), controlling devices remotely and integrating with Azure IoT Edge modules for running code and analytics at the edge.

A database of sensors

Think of IoT Hub as a database of sensors (both the physical devices and any digital twins you create for devices and systems) with a  cloud gateway that handles the bi-directional connectivity to those devices for command and control as well as data ingestion.

Each IoT hub has an identity registry for all the devices that connect to it and the service handles authentication and credential management, as well as collecting telemetry (like sensor readings, alerts, error messages and device status), sending commands (to reboot a device or control actuators) and reading or setting properties. When you collect data from IoT devices, you can use IoT Hub for message enrichments: adding extra information to the data to make it easier for the endpoint to process.

Some IoT devices you can connect to directly; for those, IoT Hub routes messages to connected devices and receives responses. When you need more context, IoT Hub uses device twins: JSON documents that store metadata like the make, model, battery life and software revision of a device, as well as context like where it’s located and what it’s used for, with both the desired configuration that the device is supposed to have and the current reported connection.

Managing and provisioning devices

You can both provision and discover device capabilities, update firmware and configurations and retire devices that fail or have been replaced; if you’re rolling out devices at very large scale, you can pre-provision basic information during manufacturing that will allow new devices to connect to the Device Provisioning Service, download the full configuration and start communicating with the right IoT Hub instance. This is even simpler with devices that support IoT Plug and Play.

Device Update for IoT Hub lets you group devices and schedule updates, so you can upgrade all the devices on the fifth floor of a particular building over a weekend, make sure that smart lighting systems are only updated when there’s no-one in the office who needs the lights on or deploy regional-specific updates just to the devices that are physically installed in that region. Think of it as Windows Update, but for any kind of device, deploying updates as images or packages.

IoT Hub can also help you manage devices with intermittent connections by using a cloud-hosted software agent that represents the physical device: you can send the update to the cloud proxy and when the device connects, it will pass on the update and collect an event from the device so you know it’s been applied.

IoT Hub can provision and manage not just ‘traditional’ IoT sensors and microcontrollers (and the gateways that route communications for groups of those devices) but Windows IoT Core devices and Windows IoT systems (the replacement for Windows Embedded) like kiosks and manufacturing systems, or edge compute options like Microsoft’s Enclave Device Blueprint for confidential computing at the edge.

IoT Hub use case

You probably won’t use Azure IoT Hub on its own, but as the cloud connectivity gateway for an entire solution. Project Eclipse is a good example; built by Microsoft Research to do air quality sensing for entire city neighbourhoods, using sensors on bus shelters, it records data every one to five minutes and transmits the data three or four times an hour. That data comes into IoT Hub and is sent to Azure Stream Analytics, as well as to a SQL database in Azure Blob storage and an API that applications can call to get current air quality data. Project Eclipse uses that API to drive visualisations in Power BI, although Power BI’s new streaming dataflows for analysing real-time data can take streaming data directly from IoT Hub, to help you visualise your IoT devices can tell you about what’s happening in your environment.

IoT connectivity platforms

You might also find you’re using Azure IoT Hub as part of another IoT service, because it’s designed for building IoT solutions on. Many third-party, industry-specific IoT services and connectivity platforms are using IoT Hub for connectivity and management, especially with LoRaWAN, because it can handle the scale of the many sensors you can deploy on low-power wide area networks. And if you need IoT devices with cellular connectivity, you can put a Twilio Trust Onboard SIM in them and have them authenticate against Azure IoT Hub automatically.