This article is more than 1 year old
Microsoft bulks up Linux on Windows at the IoT edge in Azure
EFLOW makes the open source OS a full partner of Windows at the far reaches of the cloud
Microsoft is rolling out the latest iteration of a tool that lets anyone using a Windows IoT device in Azure to also run containerized Linux workloads.
The Azure IoT Edge for Linux on Windows (EFLOW) 1.4 LTS (long-term servicing) branch, now generally available, includes the latest features that came with version 1.4 of Azure IoT Edge, which was released in August, as well as exposing hardware capabilities on the host device to the Linux workloads and storage for sharing files and folders across both Windows and Linux.
Microsoft will support EFLOW 1.4 LTS until November 12, 2024, and those now running version 1.1 LTS can update without having to reinstall the tool, Martin Tuip, senior product marketing manager for Azure edge devices, Internet of Things security, and robotics, wrote in a blog post this week.
Azure IoT Edge for Linux on Windows makes the open-source OS more of a full partner with Windows in Microsoft's cloud. It enables users to run Windows and Linux together on a Windows IoT device, with "businesses that rely on Windows IoT to power their edge devices [now able to] take advantage of the cloud-native analytics solutions being built in Linux," Microsoft wrote in a note in June.
With EFLOW, a Linux virtual machine is run on a Windows device and is pre-installed with the IoT Edge runtime. Linux IoT Edge modules that run inside the virtual machine while Windows runs on the host device and Windows applications can communicate with the modules in the Linux VM.
With the bidirectional communication, users can leverage Windows processes for user interfaces or hardware proxies for the Linux container workloads.
The Linux VM is based on the CBL-Mariner Linux distribution and is built for hosting IoT Edge workloads, according to the giant software maker. At the same time, organizations can use Windows management tools like PowerShell modules for scripting deployments and sample scripts for provisioning EFLOW installations via Intune and Arc for Server.
- Microsoft emits more Win 11 fixes for AMD speed issues and death by PowerShell bug
- Windows 10 to hang on for five more years with 21H2 update
- Ubuntu desktop team teases 'proof of concept' systemd on Windows Subsystem for Linux
- Oracle Linux appears somewhere unexpected: The Windows Store
There also are hardware capabilities that include Trusted Platform Module passthrough to automatically provisioning Azure IoT Edge for Linux devices, bring GPU acceleration to the Linux VM module, and camera and serial passthroughs.
The new features in EFLOW 1.4 LTS includes an update to version 2.0 of CBL-Mariner Linux, support for more networking configurations such as static IP, multiple NICs, and DNS configurations, and GPU passthrough technologies like Direct Device Assignment for allocating GPUs to either the Linux VM or Windows and GPU-Paravirtualization for sharing the GPU between the two.
EFLOW supports Nvidia's T4 Tensor Core GPU for AI workloads and A2 Tensor Core GPU for video analytics as well as its GeForce, Quadro, and RTX accelerators. It also supports Intel's integrated GPU.
There also is support for deploying EFLOW on such virtual infrastructures as VMware Windows VMs. ®