Virtualization Meets Composability: A New Era for Media & Entertainment

When virtualization first entered the enterprise data center around 2008, it did so for some very practical and powerful reasons: 

1. Faster provisioning 

Deploying new physical servers used to take weeks. If you needed to run a new application or change your workload, you were stuck waiting. Virtual machines (VMs) changed that—enabling near-instant deployment and rapid scaling. 

2. Better resource utilization 

Many servers had far more CPU, memory, and storage capacity than their workloads actually required. Virtualization allowed organizations to maximize those resources, significantly improving cost efficiency. 

3. Increased agility 

Virtual machines made it easy to move workloads across different server types, as long as the resources matched. This flexibility meant you could choose how and when to run applications, adapting quickly to changing business needs. 

Virtualization meets media & entertainment 

The benefits of virtualization—speed, resource optimization, and agility—translate directly into the Media and Entertainment (M&E) world. In post-production environments, workloads like editing, color grading, and visual effects can be highly variable. It’s often unpredictable when compute, GPU, memory, and storage resources will be needed, and how intense the need will be. Virtualization helps to better optimize those resources, ensuring systems are being used with utmost efficiency and effectiveness. 

Traditionally, we’ve been limited by the physical architecture of the systems—specifically, locking GPUs directly into the server itself. The server becomes the point of reference for resource allocation.  

Since GPUs are often very expensive, being able to maximize their usage—by running more virtual machines on a single server—offered some value. But it didn’t fully solve key challenges around agility, scalability, or the ability to rapidly use different types of GPUs across various workloads. We were still limited by a one-to-one pairing of server and GPU. 

Considering the cost of these server systems and the cost of GPUs, leveraging GPUs dynamically when needed, to be able to spin up bigger or smaller virtual machines became very useful. Even with resources locked into the server, by running more VMs on that server, the GPUs and other expensive resources could be better utilized. While that provides some efficiency, it still doesn’t solve the problems of agility, cost optimization and the ability to mix different types of GPUs as needed. 

That’s where composability changes the game. Composability adds the next dimension of optimization, agility, scale, diversity of resources, and application matching that’s very critical to efficient post-production and editing of content in M&E. 

Why do M&E companies use virtualization? 

Virtualization is often used in media and entertainment because the graphic artists doing post-production editing and visual special effects aren’t always in data center environments. They work in typical offices using a laptop, and the actual GPUs that they need to access may be located elsewhere. 

Depending on the load and requirements of the graphic artists, with virtual machines they have access to resources not local to where they are sitting, but perhaps in some proximity. That allows those GPUs to then be allocated to different virtual machines. While virtualization delivers those resources to the applications, it still requires adding more infrastructure. 

So, we ask ourselves: 

1. How can we be more agile?  

2. How can we reduce our dependence on additional infrastructure?  

3. How can we do it without losing the flexibility to deliver resources exactly when and where they’re needed? 

Combining virtualization with composability is the answer to these questions.  

How does composability augment existing VM environments? 

As you run more applications in a virtualized environment, it becomes critical to allocate resources efficiently—without disrupting the existing workloads already running inside those virtual machines. 

For example, if you’re running multiple virtual machines and want to deploy GPUs for some of them, you shouldn’t have to reboot the entire server to make that happen. The ability to assign resources like GPUs in real time—without disruption—becomes a critical operational requirement in virtual environments. 

How can Cerio help?  

What makes Cerio’s approach so powerful is its ability to add resources—like GPUs—to a running hypervisor in real time, even while guest virtual machines are actively running. This means you can spin up new applications or enhance existing ones with additional resources, all without disrupting any of the other virtual machines on the host. 

This is why bringing composability to virtualization is so groundbreaking.  

While others have attempted this using standard off-the-shelf PCIe infrastructure, the reality is that PCIe was never designed for dynamic, real-time resource allocation. Cerio’s platform extends and enhances the PCIe specification—while maintaining full compatibility—enabling truly dynamic resource provisioning without compromising consistency or performance. Cerio is the only platform capable of deploying assets—such as GPUs—to a host running virtual machines without disrupting the operation of the VMs already in place. That’s a real game-changer for organizations with virtual machine deployments—they can now add new capabilities without altering the way they manage or operate their existing virtual environments. 

Composable Disaggregated Infrastructure (CDI) for M&E   

As both tier-one media and entertainment companies, and smaller organizations that rely on virtualization, seek greater agility, efficiency, and real-time resource allocation, the demand for flexible infrastructure grows. With GPU technology evolving rapidly, CDI and virtualization are becoming essential tools for adopting new innovations without disrupting existing infrastructure. As CDI becomes more widely adopted, many of the features traditionally handled by hypervisors are now also available through composability. This gives users the flexibility to choose between running virtualized environments and what can be referred to as bare metal environments—depending on their specific workload needs. 

Composability and virtualization share many of the same agility benefits, offering not just hybrid flexibility, but also the ability to choose the right technology for the current deployment needs. Virtualization expands customers’ options, and when combined with composability, the benefits are significantly amplified. The combination of composability and virtualization also introduces groundbreaking possibilities for virtualizing GPU resource deployment, providing even more choices for performance and workload optimizations.