When we talk about cloud-based production, we often focus on technology, scalability, and flexibility—but none of that matters unless everything works together, seamlessly, in real-world conditions. That’s why events like the March 2025 Interoperability Workshop at Amazon Prime Studios are so important.
This was more than just a technical test. It was a chance to sit in the trenches with engineers from AWS, Evertz, Ross, TAG, TVU, Vizrt, Spalk, Telos, and others, put our technology under stress, and see not just what works, but what could work better. Cloud-based media production is growing fast, but interoperability is the key to making it viable at scale. Without an open, collaborative approach, everyone is just solving their own problems in isolation.
AMPP OS in Action
Grass Valley’s AMPP OS-powered solutions played a key role in the workshop, demonstrating how they integrate into cloud-based production environments. Maverik-X was used for multi-ME production switching, Audio Mixer-X handled cloud-based audio processing, and Kaleido IP-X provided real-time multiviewing. Routing was managed via the ATP panel, while Densite-X functioned as a test signal generator. AMPP Health Monitor and Control ensured real-time system diagnostics, while the NDI I/O and SRT Output AMPP applications facilitated seamless media transport between cloud, remote, and on-prem sources.
This wasn’t just about validating these tools—it was about understanding how they behave under pressure, how they interact with third-party systems, and how we can continue optimizing them for real-world cloud production.
Deployment Speed Matters—But So Does Flexibility
One of the biggest takeaways was how quickly and efficiently AMPP OS can be deployed and scaled in a cloud environment. Refining CloudFormation templates reduced deployment time to just around 10 minutes, allowing production teams to scale their operations in minutes rather than hours. Transitioning CloudFormation integration templates from JSON to Python provided more control over instance discovery and deployment, making it easier to manage cloud-based resources dynamically.
Observability Is Critical When You Can’t See the Hardware
In traditional broadcast environments, engineers can spot an issue by seeing a warning light or hearing a change in fan noise. In the cloud, those physical signals don’t exist, so having deep observability built into the system is essential.
By integrating custom telemetry logging from AMPP OS into CloudWatch, we enabled real-time monitoring of CPU, GPU, and memory usage. This proved invaluable when an NDI feed wasn’t reaching the system—AMPP’s observability tools flagged the issue before it had been detected anywhere else.
Stress-testing AMPP using AWS’ Fault Injection Simulator reinforced how well AMPP handles failures. During one test, AMPP respun 69 news workloads in just over four minutes and 206 sports workloads in just over seven minutes. Those numbers matter because they demonstrate that even in worst-case scenarios, AMPP can get productions back on air fast.
Industry Standards and Open Collaboration
One of the most valuable discussions centered around open-source licensing servers. Licensing in the cloud remains a challenge for many media companies. Grass Valley has already taken a leadership role with AMPP OS and its open API, ensuring that our approach to licensing aligns with the industry’s future needs.
Why This Work Matters
These workshops aren’t just about verifying that systems work—they’re about making them work better.
Cloud-based production is still evolving, and every opportunity to collaborate with other technology leaders helps push the industry forward. Seeing different teams come together, sharing insights, and refining workflows is what makes this industry exciting.
A huge thank you to AWS for organizing the event and to all the companies that participated. We’re proud of the role AMPP OS played in these tests—and even more excited about what’s next.
