Hi Lixian,
I’m following up to check whether the issue has been resolved. Feel free to reply if you need further information. If the information provided was helpful, please click "Accept Answer" to help others in the community. Thank you!
This browser is no longer supported.
Upgrade to Microsoft Edge to take advantage of the latest features, security updates, and technical support.
We operate a highly available virtualization infrastructure running Hyper-V, which currently consists of a dozen physical nodes that each pack 56 processor cores. Across this entire server pool, we are hosting exactly 50 guest VMs. I am completely lost on how to properly calculate the exact number of Windows Server core entitlments we actually need to buy to stay compliant, especially since these workloads can dynamically migrate to any node at any time. Could someone break down the exact math formula for a clustered environment like this? Mircosoft's official documentation is just making my head spin, and I really don't want to mess up what will likely be a massive purchase order.
Hi Lixian,
I’m following up to check whether the issue has been resolved. Feel free to reply if you need further information. If the information provided was helpful, please click "Accept Answer" to help others in the community. Thank you!
Hi Lixian,
Basically, Windows Server core licenses are assigned to the physical hardware, not to the number of VMs. Because your workloads can migrate across any node, you must license every physical core in the cluster.
Here’s the math: each of your 12 nodes has 56 cores, which means 12 × 56 = 672 physical cores in total. Windows Server is sold in packs of 2 cores, with a minimum of 16 cores required per server. For your environment, you would need 672 ÷ 2 = 336 two‑core packs to cover the entire cluster. Once all physical cores are licensed, you can run as many VMs as you like, provided you’re using the Datacenter edition. If you’re on Standard edition, each set of core licenses only covers two VMs per node, so Datacenter is usually the more cost‑effective choice for highly virtualized clusters like yours.
I hope the response provided some helpful insight. If it clarified the issue for you, please consider marking it as Accept Answer so others with the same issue can find the solution.
Jason.