Leverage Windows Edge Devices for Video Analytics
Published Aug 31 2021 09:01 AM 3,243 Views



With hundreds of stores worldwide, Contoso Ltd., an example company representing the AVA customer base, requires a substantial amount of computing resources to run its vast network of IoT-enabled video cameras. Its Vision AI models furthermore require powerful IoT Edge devices that can make hundreds of computations per minute and stream inferences to the Cloud. Contoso’s Digital Advisory Team already has an existing fleet of Windows devices and seeks to leverage those devices to run Azure Video Analyzer Edge. By leveraging Windows IoT and the growing trend of Linux-based containerized microservices, Contoso can use its existing assets to derive real time business intelligence by applying AI to video. 


Instead of purchasing new Linux Edge devices, users like Contoso can instead leverage EFLOW. EFLOW (Azure IoT Edge for Linux on Windows) is an Azure product that allows users to run Linux workloads in Windows IoT deployments. Linux processes can work simultaneously with Windows processes, bringing interoperability between the two systems so that IT administrators can use their existing Windows apps and tools to manage these devices. With EFLOW, users can run a full Linux virtual machine that comes pre-installed with its own OS and Azure’s IoT Edge runtime, making it easier and faster to deploy Edge devices that work with AVA—all the while leveraging existing computing resources.


This week, we refreshed our documentation to showcase a simplified integration with EFLOW. In line with our focus on ensuring a consistent experience for video analytics solutions developers, irrespective of the OS and of underlying hardware acceleration platform, we now guide users through the PowerShell experience of running EFLOW with AVA. This integration simplifies the user journey of spinning up a Linux VM and installing the AVA Edge modules from a single terminal. All it takes is five simple steps.


Overview of steps to deploy AVA Edge modules on EFLOW:

  1. Install IoT Edge for Linux on Windows on your Windows device
  2. Check if your Windows device is now on IoT Edge
  3. Run script to set up local folders, input folders and user account for Azure Video Analyzer on IoT Edge device
  4. Run script to set up cloud resources for Azure Video Analyzer
  5. Run pipeline and observe results



Step-by-step instructions in the form of a how-to guide can be found here. Follow the guide to deploy AVA Edge modules on EFLOW, and get up and running with AVA.

Version history
Last update:
‎Sep 01 2021 08:58 AM
Updated by: