For years, the direction of VDI seemed straightforward: move desktops to the cloud.
But enterprise infrastructure rarely follows a straight line.
Organizations still have physical workstations in offices and data centers. They have virtual machines running on existing hypervisors. They have applications and data that need to remain on-premises. At the same time, they want the centralized management and accessibility associated with cloud services.
Microsoft’s latest expansion of Azure Virtual Desktop reflects that reality.
With Azure Virtual Desktop Hybrid, organizations can now use the Azure Virtual Desktop service while hosting session hosts on existing on-premises hypervisors or supported physical Windows systems. Azure Arc connects those resources back to Azure, allowing them to participate in AVD host pools without requiring the desktops themselves to run in Azure.
It’s an important development, but perhaps the more interesting story is what it says about the direction of EUC.
Hybrid Is Becoming the Architecture, Not the Transition
Hybrid infrastructure was once treated as an intermediate stage.
An organization might keep some workloads on-premises while it gradually migrated everything else to the cloud. Eventually, the thinking went, most desktops and applications would live there.
That hasn’t necessarily happened.
Instead, organizations are discovering that different workloads belong in different places.
A standard knowledge worker desktop may make sense in the cloud. A GPU workstation may need to remain close to large datasets. An organization may have existing virtualization infrastructure that continues to deliver value. Regulatory, latency, cost, and data sovereignty requirements may keep other resources on-premises.
The result isn’t necessarily a temporary hybrid environment waiting to become cloud-only. For many organizations, hybrid is the architecture.
Microsoft’s decision to extend AVD to existing on-premises infrastructure reinforces that idea.
The Control Plane Doesn’t Have to Live With the Desktop
There’s another important architectural lesson here.
Where a desktop runs and how IT manages access to that desktop do not have to be the same decision.
With AVD Hybrid, Microsoft’s service remains in Azure while the session hosts can remain on-premises. Azure Arc connects those resources into the Azure management model.
That separation is significant.
It reflects a broader shift from thinking about VDI as a tightly integrated stack toward thinking about the access and management layer independently from the infrastructure underneath it.
The desktop can live in one place. The control plane can live somewhere else.
Change the infrastructure, and you shouldn’t necessarily have to change how users access their workspaces.
But Hybrid Infrastructure Is Only One Kind of Flexibility
This is where organizations should look beyond the question of whether a platform supports cloud and on-premises resources.
Infrastructure location is only one decision in an EUC architecture.
There are others:
- Which hypervisor should host a particular workload?
- Should the resource be physical or virtual?
- Does the workload require Windows, Linux, or macOS?
- Which display protocol delivers the right experience?
- Should the resource be persistent, pooled, or dynamically provisioned?
- How should resources be started, stopped, assigned, and reused?
- What happens when one of those technologies changes?
AVD Hybrid expands where Azure Virtual Desktop session hosts can reside. But the architecture remains centered on the Microsoft ecosystem.
That may be exactly what some organizations need.
Others may need a broader definition of hybrid.
Hybrid Should Mean More Than Cloud + On-Prem
Consider an engineering organization.
It might have physical GPU workstations in one location, virtual desktops running on an existing hypervisor in another, Windows desktops in Azure, and specialized Linux resources supporting high-performance workloads.
Different users may also require different display protocols depending on graphics performance, network conditions, endpoint devices, and workflow requirements.
From the user’s perspective, none of that should matter.
They should authenticate, see the resources they’re authorized to use, and connect.
For IT, however, the ability to make those underlying choices matters a great deal.
A truly flexible EUC architecture should allow the organization to choose the right infrastructure and connection technology for each workload without creating a separate access silo every time.
Management Matters, Too
There is also an important distinction between connecting to a resource and orchestrating that resource.
Microsoft currently notes that AVD Hybrid does not provision or manage the state of on-premises virtual machines. Power management, autoscale, Start VM on Connect, and Session Host Configuration aren’t supported for AVD Hybrid. Organizations need to manage those capabilities through their hypervisor tools, scripts, partner solutions, or other systems.
That becomes particularly relevant as environments grow.
Knowing where a desktop lives is one thing.
Knowing when it should start, who should receive it, which protocol they should use, when it can be powered down, and what should happen after the session ends is another.
This is where the control plane becomes more than an access portal. It becomes an orchestration layer.
Design Around the Workspace, Not the Vendor
None of this makes Azure Virtual Desktop Hybrid a step in the wrong direction.
Quite the opposite.
It validates an important idea: organizations shouldn’t have to move every workload to a particular infrastructure simply to create a consistent digital workspace experience.
But it should also prompt IT teams to ask how far they want that flexibility to extend.
If your organization is primarily standardized around Microsoft technologies, an Azure-centered architecture may make sense.
If your environment spans multiple clouds, hypervisors, operating systems, workstation types, and display protocols, the question becomes broader.
How do you create consistency without forcing all of those resources into the same technology stack?
The answer may be to design the access layer around the workspace rather than around the vendor providing the infrastructure underneath it.
Where Leostream Fits
Leostream has approached EUC from this perspective for more than two decades.
The Leostream Platform provides a vendor-neutral control plane for orchestrating access and connections to desktops, workstations, applications, and high-performance resources across on-premises, cloud, and hybrid environments.
Rather than requiring organizations to standardize on a particular cloud, hypervisor, operating system, or display protocol, Leostream allows IT to select technologies based on the requirements of each workload while maintaining centralized access, policy, brokering, and resource management.
That includes the ability to orchestrate resource lifecycle actions such as provisioning and power control where supported, an increasingly important consideration when expensive cloud and GPU resources are part of the environment.
The goal isn’t to move everything somewhere new.
It’s to make the infrastructure you already have, and the infrastructure you adopt next, part of a consistent access strategy.
Cloud-Managed Doesn’t Have to Mean Cloud-Hosted
Azure Virtual Desktop Hybrid is another indication that the industry is moving beyond the idea that the future of EUC means putting every desktop in the public cloud.
The more interesting future is one where organizations can place each workload where it makes the most sense while giving users a consistent way to access it.
That’s a meaningful step forward.
The next question is how much choice your architecture gives you when the cloud, hypervisor, protocol, operating system, or vendor changes again.
Because the goal of hybrid EUC shouldn’t simply be supporting more places to run a desktop.
It should be giving IT the freedom to decide where that desktop belongs.