Getty Images/iStockphoto
Windows Server 2025 Hyper-V updates promise speed boost
Will the new and improved features in the upcoming Hyper-V release come at the perfect time for Microsoft to capitalize on a series of unpopular changes from VMware?
Timing is everything, and Microsoft's plans for a major update to Hyper-V, coupled with recent unpopular changes to VMware, could work to Microsoft's advantage.
By most accounts, VMware remains the leading server virtualization vendor in terms of market share with some industry research surveys showing numbers as high as 60%. However, Broadcom's recent acquisition of VMware and consequent moves, such as ending perpetual licenses and eliminating the free version of the ESXi hypervisor, have brought unease to its customers. The timing of VMware's upheaval is fortuitous for other vendors in the virtualization space, particularly Microsoft. The company plans to release Windows Server 2025 later this year with new features and updates in its Hyper-V virtualization platform that have been on the wish list for server admins for quite some time. Will Hyper-V 2025 deliver on these expectations and sway more enterprises to Microsoft's side?
Hyper-V 2025 will handle more resource-hungry workloads
One of the areas where Microsoft invested a lot of development in Windows Server 2025 is Hyper-V, specifically with its scalability improvements.
As it stands today, only the largest enterprises will be able to fully take advantage of these enhancements, simply because a majority of Windows workloads do not require this level of scalability. But, for companies that need more comprehensive VMs to handle resource-intensive workloads related to AI applications or big data processing, the changes coming in Windows Server 2025 could persuade existing VMware customers to look more closely at Hyper-V. As time goes on, these optimizations should extend their benefits to other organizations as well.
Hyper-V's scalability improvements in Windows Server 2025 include the following:
- VMs can support up to 2,048 virtual CPUs (vCPUs).
- Maximum RAM of 240 TB for generation 2 VMs.
- 256 SCSI disks per VM with four SCSI controllers.
- Maximum virtual disk size of 64 TB when using VHDX format.
- Maximum of 68 virtual network interface cards per VM.
- Maximum RAM of 4 PB for Hyper-V hosts.
- Maximum of 64 nodes and 8,000 VMs in a failover cluster.
According to Microsoft documentation, the numbers are similar to Windows Server 2022 except in the area of vCPUs. Windows Server 2025 supports up to 2,048 vCPUs for generation 2 VMs, an increase from 1,024 in Windows Server 2022.
Microsoft updates GPU partitioning in Windows Server 2025
A Hyper-V feature Microsoft intends to update in Windows Server 2025 is GPU partitioning, or GPU-P as the company sometimes calls it, which gives multiple VMs access to the server host's GPU to boost performance of virtualized applications or services. GPU vendors design their hardware for parallel processing to make it ideal for AI workloads, such as training machine learning models.
In earlier Windows Server versions, Hyper-V used Discrete Device Assignment (DDA) to dedicate a physical GPU to a VM. However, there were some significant drawbacks. Using DDA to attach a GPU to a VM limits certain features during that connection, such as the ability to use dynamic memory, VM Save/Restore and live migration. Also, the VM cannot participate in a high availability (HA) cluster. Another issue is the GPU cannot be shared among multiple VMs, requiring a separate GPU for each VM to run GPU-intensive workloads.
The revamped GPU-P feature in Windows Server 2025 gives the functionality to share a GPU among multiple VMs. Each VM receives dedicated resources within a physical GPU. Admins can configure and manage these VMs using Windows Admin Center.
Additionally, Microsoft will support live migration and HA for VMs that use GPU-P in Windows Server 2025. If IT needs to perform maintenance, then critical workloads can continue to run uninterrupted thanks to live migration. If an outage or some other unexpected event occurs, then the HA update can restart a VM that is using GPU-P on another node in the cluster.
Workgroup clusters get more flexibility
In Windows Server 2016, Microsoft introduced workgroup clusters, designed for smaller organizations that wanted to build failover clusters without an AD domain.
These workgroup clusters were primarily meant for use with Microsoft SQL Server. While they could be used with Hyper-V, it went against Microsoft recommendations because migration was not supported.
In Windows Server 2025, Microsoft adjusted the workgroup clusters feature to permit the live migration of Hyper-V VMs by using certificates in the absence of an AD environment.
Windows 2025 workgroup clusters may prove to be useful for an enterprise that needs to operate Hyper-V failover clusters in an isolated edge environment without incurring the expense and overhead associated with deploying an AD domain.
Microsoft revises dynamic processor compatibility
Another revamped feature coming in Windows Server 2025 Hyper-V is processor compatibility mode. Microsoft debuted PCM in Windows Server 2008 R2. Prior to this, if you wanted to live migrate a VM or if you wanted to restore a production checkpoint onto a different Hyper-V host, then the source host and the destination host needed identical CPUs.
PCM bypasses this issue by turning off manufacturer-specific or generation-specific CPU features to force the VM to run on what amounts to a bare-bones CPU. The VM can then run without issue on the destination host during a live migration because any compatibility issues have been stripped away.
The disadvantage of PCM is degraded VM performance when shutting off speed-optimizing CPU functionality.
In Windows Server 2025, Microsoft will introduce the dynamic processor compatibility (DPC) mode. This is an evolution of PCM and attempts to improve overall performance during these Hyper-V operations. With DPC, Hyper-V compares the CPU features on the source host and the destination host to determine what to leave enabled for optimal speeds.
Generation 2 VMs will be Windows Server 2025 Hyper-V default
Microsoft introduced generation 2 VMs in Windows Server 2012 R2. At the time, generation 2 VMs were designed to provide enhanced hardware capabilities to VMs running newer Windows OSes. These capabilities included features such as Preboot Execution Environment boot, SCSI virtual disk and DVD drive support, Secure Boot and Unified Extensible Firmware Interface support.
Starting with Windows Server 2025, Microsoft will make generation 2 VMs the default option when creating a new VM. The benefits for the enterprise include the default use of Secure Boot to avoid malicious action on firmware at bootup and virtualization-based security.
Brien Posey is a former 22-time Microsoft MVP and a commercial astronaut candidate. In his more than 30 years in IT, he has served as a lead network engineer for the U.S. Department of Defense and a network administrator for some of the largest insurance companies in America.
