The Dynamic Media Facility (DMF)
Efficiently managing the lifecycle of your media infrastructure in contribution, production & distribution environments
As the media and broadcast industry continues to evolve, a fundamental shift is underway: from SDI to IP, from hardware to software, and from static infrastructure to dynamic, usage-based environments.
This transformation opens the door to greater flexibility, efficiency, and sustainability. But to fully capitalize on it, broadcasters must rethink how they design, build, and manage their infrastructure, not just in terms of technology, but in how operations are planned, automated, and optimized.
The Dynamic Media Facility (DMF), a reference architecture defined by the European Broadcasting Union (EBU), provides a blueprint for building facilities that can adapt dynamically to changing production demands.
What are Dynamic Media Facilities (DMF)?
Dynamic Media Facilities (DMF) represent a new approach to media infrastructure, as defined by the European Broadcasting Union (EBU). Unlike traditional broadcast facilities with fixed hardware configurations, DMF leverages containerized, software-based media functions that can be dynamically provisioned, scaled, and managed across distributed computing resources.
Key characteristics of Dynamic Media Facilities:
- Containerized media functions: All media processing tasks run as software containers
- Cloud-Native architecture: Built for on-premises, hybrid, or full cloud deployment
- Dynamic resource allocation: Compute, storage, and network resources scale automatically based on demand
- Vendor-agnostic design: Mix and match media functions from multiple vendors
- High-performance media exchange: Ultra-low latency media sharing between functions
Static infrastructure: a bottleneck in a dynamic world
Traditionally, broadcast facilities have been designed for peak loads and worst-case scenarios. This has led to significant overprovisioning and underutilization, for example with ingest servers and media processing resources.
Let’s consider a real-world example.
A broadcaster invests in 100 ingest channels for contribution feed recording: 80 HD and 20 UHD. This delivers a theoretical capacity of 876,000 hours of recording per year. However, only a small fraction of that capacity is actually used. In many cases, utilization does not even reach 10 percent.
In reality, 10 ingest channels may be sufficient to handle the average daily load. The challenge is covering occasional peaks, without maintaining 90 idle channels the rest of the year.
The Hybrid approach: static base, elastic extension
The solution is to design your facility with a hybrid infrastructure model:
- Run your base load on-premises, covering average demand with high reliability and low cost.
- Provision extra resources dynamically in the private or public cloud for live events, peak times, or UHD-heavy workflows, only when needed.
This model aligns perfectly with pay-as-you-go commercial strategies, offering scalability and cost control.
What’s holding broadcasters back?
In most facilities today, operational workflows are highly manual and siloed.
Here’s what a UHD ingest request might still look like:
- The scheduling team prints a request
- The MCR and production teams check resource availability in an Excel sheet
- Resources are manually reserved and communicated via email
- An operator routes signals and starts configuring all on-premises devices manually
This process may have worked for years, but it is inefficient, expensive, error-prone, and not scalable in a dynamic, cloud-integrated world.
Even simple tasks like reconfiguring an ingest channel from HD to UHD require manual setup, testing, and coordination across teams.
The goal: resource orchestration across the entire media production lifecycle
To unlock the full potential of dynamic media infrastructure, broadcasters must digitize and orchestrate every part of their operations.
Here is what needs to be done to manage and orchestrate the lifecycle of a media production:
Infrastructure repository
Maintain a complete inventory of all available resources. This includes on-prem, virtual, and cloud-based resources, as well as people, rooms, SNGs, cloud credits, licenses, and more. Eventually, everything becomes a resource.
Bookable resource pools
Classify resources by type, capacity, capability, and availability.
Plan ahead
Define the resource needs for each production. For example: 5 ingest channels, UHD support, 50 Mbps XDCAM, and specific routing requirements.
Automated allocation
An orchestration engine automatically reserves and allocates the appropriate resources, considering availability, capacity, and capabilities.
Just-in-time deployment and configuration
Resources are deployed where needed, configured, and signal paths are routed just before production begins.
Dynamic operation
Operators use tailored UIs to control the active resources assigned to them during live operation. Resource monitoring adapts dynamically and follow the production lifecycle.
Automated teardown
Resources are released and reset post-production, making them ready for the next task.
Cost tracking
Usage is calculated using predefined rate cards for internal or external billing.
Post-service optimization
Usage data is retained for long-term optimization and investment planning. It also supports automated ESG (Environmental, Social, and Governance) reporting and goal-setting.
Managing the lifecycle of dynamic media resources and their dependencies
The Dynamic Media Facility requires more than traditional, and often separate, scheduling and automation tools, unified orchestration solutions are required to manage resources that do not yet exist at the time of booking.
These could include:
- A video server microservice spun up in the cloud
- A virtual machine provisioned on-prem
- A reconfigurable FPGA-based media processing platform
To handle such resources, the orchestration engine must be able to:
- Deploy, reconfigure, or spin up media functions
- Check for available compute, network capacity, and licenses
- Integrate all dependencies into the scheduling logic
A Dynamic media facility is not only about dynamic infrastructure, but also about dynamics in operations, it must be able to respond to unplanned or unforeseen situations on-the-fly, including:
- Breaking news
- Resource changes during live productions
- Swap faulty resources
- Adapt to dynamic resource pricing
- and much more
Managing this dynamic infrastructure and the dynamics of operations effectively requires advanced orchestration across the entire technology stack combined with all your operational resources.
Reality Check: where are we today?
Major vendors have already taken first steps, releasing the Media Exchange Layer (MXL) SDK, providing the data-plane to achieve interoperability across platforms. A production-ready release is targeted for Q4 2025.
However, the orchestration and control layer are still in an early stage. Simply combining any media functions on the fly would introduce complexity that’s impossible to manage at scale and doesn’t align with how cloud-native functions are developed and deployed. Automated version control, security and CI/CD-pipelines are required to build, test, verify, release, deploy, operate and monitor media productions, i.e. media productions must rely on pre-tested templates.
For example, in a DMF-based live production:
- Media functions in the production studio, their configuration and connectivity can be spun up using predefined templates
- Contribution feeds can be managed more dynamically based on venue or signal format
- Playout and channel operations remain more static and templated
Skyline in action: orchestrating the future
At Skyline Communications, we are working closely with leading vendors such as Grass Valley (GV AMPP) or LAWO (HOME APPs) to help broadcasters benefit from software-based media functions and new commercial models such as pay-as-you-go or subscription.
To make this this a success, we focus on small, manageable steps: using templated workflows and productions wherever possible and the associated resources throughout the production lifecycle.
🔗 Watch the GV AMPP demo video
🔗 Explore the LAWO HOME use case
🔗 Explore the GV AMPP use case
Ready to build your own Dynamic Media Facility?
With Skyline’s DataMiner orchestration and deep ecosystem integration, you can:
- ✅ Automate and template complex operational workflows
- ✅ Provision infrastructure when needed
- ✅ Optimize resource usage and costs
Contact us to explore how Skyline can support your journey toward a Dynamic Media Facility already today!