RemoteFX vGPU (unexpected) End-of-Life

RIP cheap GPU virtualization

Door Mikhail Zuskov

The year of 2020 was a very odd one, full of unexpected events and surprises. The COVID-19 pandemic was definitely one huge “black swan”, but it was not the only one. Another event which we might classify as “black swan” was brought by Microsoft during July 2020 patch round. It is the decision to disable RemoteFX vGPU functionality in all Windows versions from Windows 7 SP1 and up to latest and fully supported versions of Windows 10 and Windows Server and pull out the code completely during February 2021 patch cycle (read more). Microsoft announced it back in march 2018 that RemoteFX vGPU feature is no longer being actively developed (read more), yet this decision fell completely out of the blue because it involves products which haven’t even reached the end of standard support cycle.

A lot of organization which are using Microsoft Remote Desktop Services in conjunction with Microsoft Hyper-V platform were unhappily surprised with this development. These (mostly Small to Medium Sized) organizations rely on affordable GPU-backed desktop virtualization (aka VDI) for their daily operations, but starting from February 2021 Patch Tuesday this has become quite challenging and each organization has to make a choice between:

1. Staying up-to-date and secure, but not being able to use RemoteFX vGPU. In this case organization has 2 options:

  • Stay on Microsoft virtualization stack and switch to Discrete Device Assignment (DDA)
  • Use an alternative VDI solution

2. Keep using RemoteFX vGPU, but accept vulnerabilities in the infrastructure, because:

  • The Hyper-V servers can’t be patched any more starting from February 2021 as long as you need to use RemoteFX vGPU 
  • Using RemoteFX vGPU increases the risks for the CVE-2020-1036 to be actually exploited in your network 

All these choices have their own pro’s and con’s.

All these choices have their own pro’s and con’s. Is your company overwhelmed by this sudden surprise or ‘black swan’ and do you need additional information to decide how to provide your users with GPU-enabled VDI? Do not hesitate to contact us.