In previous posts, I have discussed using a Raspberry Pi device connected to Microsoft Azure IoT. While this gives you a good example of how smart devices can communicate with the cloud and deliver useful information, it is really just treating the devices as communication tools, and doesn’t take advantage of any of the processing power that exists within the Raspberry Pi. When you think of a complete “Internet of Things” (IoT), you can picture the “thing” as being a device that bridges the gap between on premises activity and the cloud. At Microsoft, we call this the “Intelligent Edge”, and we see this as a major opportunity to bring the power of the cloud to the world of IoT.
The Intelligent Edge
The heart of the Intelligent Edge is the new Azure IoT Edge service. In short, this service is built for those who want to push some of the cloud processing power down to the edge devices, so that they can independently analyze data (think about deploying an AI model to the edge and allowing the device to perform a portion of the work instead of having to rely on sending everything to the cloud and waiting for the results).
The IoT Edge service consists of cloud components, such as Azure IoT Hub, Azure Stream Analytics, and other Azure Services. The IoT Edge also includes a runtime service that is deployed to the device. This service is based on Docker, and is designed to simplify the deployment and management of code modules that are deployed to the device, as well as facilitate communication both with the cloud, and with any downstream devices that might be part of the environment (Think of the Edge device as being a controller that monitors the local environment, performs analysis there, and then reports status upstream to the cloud). The overall architecture of a connected IoT Edge device looks like this:
The IoT Edge runtime supports both Windows and Linux devices, and simplifies the deployment and management of code modules.
Next Steps
This post has been an introduction to the Intelligent Edge and is really meant to provide an overview of the term and the services necessary to implement it. In follow-on posts to this topic, I will walk through how to configure a Raspberry Pi device to become an Intelligent Edge device, and will walk through deploying intelligent modules to the device.