In live production, switching happens in the moment. In sports, for example, no one knows what’s coming next—and if the operator misses the shot, it’s gone. Even in news, where rundowns provide structure, switching still has to react dynamically to what’s happening on air.
The pressure hasn’t gone away. But the systems behind it have evolved.
In my previous column, I said that FPGA changed the trajectory of production tools by letting engineers evolve a switcher instead of rebuilding it. Functions that once required fixed hardware can now be designed in software and loaded onto programmable silicon. If something needs fixing, it can be reprogrammed. If requirements change, it can often be reworked, provided the FPGA has been sized correctly.
A production switcher can be built to meet today's requirements. When a customer asks whether the system could also do something it wasn’t originally designed to do, the question becomes whether that capability can be added within the existing architecture.
Cloud With Constraints
Across the industry, business models are under pressure. More content is being delivered over the internet and OTT platforms. Long-established broadcasters are rethinking how they operate. Sports organizations stream directly to audiences, corporate teams run significant production infrastructure, and independent producers put content online every day. The production landscape has grown, and with it, the infrastructure decisions.
Five years or so ago, the industry buzzed about moving everything to the cloud. With income and turnover reducing, many organizations needed to find ways to cut both operating costs and CapEx budgets—and the cloud seemed like the answer. Providers like AWS and Azure offer massive amounts of compute power, so the whole industry began working on virtualizing what we've typically done in hardware and FPGAs.
The hybrid approach will be the real game changer.
Some parts of media production made that shift easily. Media playout, for example, has always suited the cloud, Netflix being the obvious example.
But live production is different. The challenge isn't insurmountable, but it's technically complex. Getting multiple feeds into the cloud, producing there, and getting the output back requires compromises, including a lot more compression. Still, if it can save money, people want to make it work.
More recently, early enthusiasm about the cloud has given way to a shift in thinking. Without major advances in technology, putting everything in the cloud remains, for live production at least, something closer to a utopian dream than a practical reality, except at relatively small scales.
Bursting into the Cloud
In live production, systems have always been built for the biggest day—the worst-case size and production need. That means a lot of capacity sits idle most of the time. Every studio in the world is overbuilt on purpose, sized for the maximum need, even though that maximum might only happen a few times a year.
Bursting offers a different approach. Rather than investing in permanent infrastructure to handle rare peaks, the cloud can supplement an existing system when demand briefly exceeds the base setup.
Elections are a good example. A broadcaster may suddenly need additional camera feeds for a few hours or a few days, more than the on-prem system can support. In those cases, extra capability can be added temporarily by bursting into the cloud.
Used this way, the cloud doesn’t replace on-prem systems; it complements them. For many use cases, putting everything in the cloud can end up costing more than running systems on-prem. As a supplement, however, there can be enough affordable horsepower to make it worthwhile.
That’s why the conversation has shifted. It’s no longer about on-prem or cloud, but how to combine the two in a way that fits organizations, workflows, and business models. The hybrid approach will be the real game changer. And as it matures, so will the tools and services that support it.
For most, the future looks hybrid.
