Microsoft Brings Plug and Play to the Internet of Things

SEATTLE—We've been hearing a lot about the Internet of Things (IoT) these days, but how can developers connect their code with IoT devices, which often use proprietary, tightly coupled inputs and outputs? There's no standard way to know which kinds of data a particular device can send. Until now.

Back in the days of Windows 95, Microsoft had a similar problem: connecting the operating system with printers and other hardware peripherals. Ultimately, the company came up with the Plug and Play model, which made printer capabilities clear and certified the devices.

We're going through a similar phase with IoT devices right now, and once again, Microsoft's response is to create a plug-and-play solution for IoT, called—unsurprisingly—IoT Plug and Play.

At the center of Microsoft's IoT initiatives are Azure IoT Central and Azure IoT Hub. Azure IoT Hub, which debuted in 2015, is a Microsoft Azure cloud service that "offers reliable and secure device-to-cloud and cloud-to-device messaging that scales to millions of devices." Microsoft IoT Central, which debuted in spring 2017, is a dashboard that pulls together data from all of your IoT devices "without requiring cloud solution expertise."

At an IoT session here at the Build developer conference, Peter Provost, Principal PM Manager for Azure IoT Central, outlined a vision for IoT in which digital feedback loops help companies track, optimize, simulate, and predict future scenarios by tying together things, places, and people.

Provost ran through some numbers to drive home his case: 50GB of data generated every day by a smart home, 5TB by an autonomous vehicle, 200TB per sporting event at a stadium, and 1 petabyte daily at a connected factory. A million new devices are coming online per hour.

Why not harvest all that data, analyze it, and draw some insights or drive processes with it? That's the whole point of Azure IoT Central. Two slides from the presentation demonstrate just how much coding Azure IoT Central takes out of the project: provisioning, user interface, identity, device state, stream processing, storage, and analtytics and machine learning are all provided by Azure IoT Central.

The device can send time-based telemetry as well as settings, properties, and commands. And IoT Central can generate charts from the data and specify rules that trigger actions. For example, if the temperature goes over a certain level, send me an email.

Here's a sample IoT Central dashboard for IoT-equipped vending machines:

Challenges abound for those developing an IoT system: knowledge of embedded systems, transforming low-level device messages into meaningful data, and hardware sourcing, integration, and production, to name a few. All of these revolve around the physical devices. Enter IoT Plug and Play.

IoT Plug and Play, which will be in public preview later this year, allows device makers to certify their hardware's capabilities, making them far more easily usable for customers building IoT systems. The devices will expose uniquely identifiable capabilities, including properties, telemetry, and commands. They'll adhere to a device-capability model, and Microsoft is making a description language available as Digital Twin Description Language (DTDL). The latter provides a virtual model of the device, which allows modeling scenarios using the device without needing the actual device itself. Its open-source code is available at aka.ms/DTDL.

This diagram shows the system overview:

The device can either send its capability model to your cloud app itself, the Azure Capability Model Repository, or refer to a URI (which is good for on-premises deployments). Of course, the repository integrates best with Azure cloud services and developer tools—but it's not required by IoT Plug and Play. Dev tools include device and service SDKs, a Workbench extension to Visual Studio Code, a CLI extension, and an Azure IoT Device Explorer (for finding suitable devices).

Towards the end of the session, Provost and Briton Zurcher, Group Program Manager for Azure IoT, showed an example of an IoT Plug and Play device, the Askey Connected Car Cam+. Hooking the device into Azure IoT Central was quick, and data started streaming into IoT Central visibly in graph lines.

You can browse the complete catalog of certified IoT Plug and Play devices at catalog.azureiotsolutions.com. The team will be at Computex next month in its continued quest to make the world of IoT Plug and Play ready.

Microsoft’s HoloLens 2 Is a Huge Leap Forward for Augmented Reality