Challenge
Live production is moving onto software media functions, which vendors deliver as containers and facilities run in dozens on their own servers. Each function shares a Linux host with other workloads and exchanges media with other hosts over RDMA. When a function falls behind real time, the cause can lie in the host or the network between hosts. DMF and MXL describe how such a facility fits together. MXL reached version 1.0 in March 2026 but has not yet been shown in round-the-clock production. DMF is still a living document and leaves the hosts and networks beneath it to standard IT, with open questions such as how several vendors’ functions share one host safely and at full speed.
Approach
Our engineers have built and run production container platforms since 2015, and work with your broadcast team and vendors on the hosts and networks beneath your media functions. After an intro call, and if we can help, we run a workshop on one infrastructure decision with no commitment beyond it, then you choose when to start proofs of concept.
We tune each host and the network between hosts so media functions keep up with real time. Because a container cannot restart inside a frame, we set up hot standby, which also lets you patch hosts without going off air.
We help you set a trust model which keeps vendors and productions apart in one facility, without taking away the privileges low-latency tuning needs. We also define each production’s hosts and workloads as code, so you can test a setup, repeat it, and move it between your facility and the cloud.
If your organisation buys through AWS, you can also engage us through AWS Marketplace, with billing on your AWS account. If you are ready to run TAMS, TAMOSS, our open-source TAMS store, has its own three-week pilot.
- 11+ Years
- running production container platforms
- 3 of 6 DMF layers
- are hosts, containers, and networks
- 1 hour
- Intro call
Cloud Native Broadcast Engineers FAQ
Priya Shah, Head of Broadcast Engineering
Is DMF ready for production?
LiveWyer
Not yet, so we start with proofs of concept
The EBU and AMWA's JT-DMF only began formal work in late 2025, and no vendor has yet shown MXL in round-the-clock production. Our work on DMF and MXL is trials and test workloads which get your team ready to adopt MXL once vendors run it in production, and DMF once the task force finishes its work. The host, security, and network skills behind that work are the ones our engineers use on production platforms today.
Daniel Okafor, Chief Technology Officer
Do you replace our vendors or our broadcast engineers?
LiveWyer
No, we work alongside both
Your vendors build the media functions and your team knows the facility. Our engineers cover the Linux hosts, container runtime, and networks underneath.
Hannah Clarke, Director of Technology
Do we have to run on AWS?
LiveWyer
No, AWS is one option
Proofs of concept run wherever you want to test, on your own servers or on AWS. On AWS, you can prove a setup before buying hardware, then use it for overflow productions or as a backup facility. AWS provides RDMA between instances through Elastic Fabric Adapter, and precise clocks on supported EC2 instances.
Talk to a specialist
Moving media functions onto standard IT infrastructure? Book a quick chat with one of our senior engineers, and we will tell you whether we can help.
Talk to a specialistCase Study: API control plane for broadcast file delivery
Red Bee Media now runs broadcast file delivery through an API-driven serverless control plane, with its legacy systems unchanged.
- 0 changes to legacy software
- 4 existing systems integrated
- 3 years supported in production
- Go agents wrap each existing system, so no vendor software needed modifying
- AWS managed services carry the workflow and its state
- Infrastructure as code, an admin panel, and monitoring for the operations team
- Handed to the in-house team with the documentation to run and extend it