The production environment that broadcast infrastructure was built around no longer exists. The model of the studio with a defined perimeter, a predictable signal path, and content that stayed within four walls has been under pressure for years, and the pressure is not easing.
Remote production means contributors are no longer in the building. Corporate clients with broadcast-quality studios want those studios to feed their enterprise networks: the lobby screen, the conference room, the overflow space on the floor above.
Houses of worship running full production environments need the same signal that powers their main auditorium to reach every classroom and corridor in the complex. Education facilities are building multi-camera setups not just for events, but for daily instruction that flows across buildings and campuses.
None of these are edge cases. They represent a significant and growing segment of production activity, and none of them are well-served by a contained baseband workflow. The SDI-to-IP conversation has been going on for a decade. What has not kept pace is the infrastructure layer that determines whether IP actually delivers on its promise.
The Real Barrier is not the Protocol
What often gets lost in the SDI-versus-IP debate is that the barrier to adoption is rarely the protocol itself. NDI, Dante, AES67, SMPTE ST 2110 are proven, mature, and increasingly affordable. The technology works. The problem is the network sitting underneath it.
For years, configuring a switch for a broadcast or professional AV environment required someone with deep IT networking knowledge. Someone who understands multicast routing, IGMP snooping, PTP grandmaster configuration, VLAN segmentation across mixed protocols. Large integrators might have had two or three people in the entire organization capable of doing it correctly.
That bottleneck was a meaningful brake on IP adoption, and operations lacking that specialist knowledge either stayed with SDI, made configuration errors that eroded confidence in IP workflows, or relied on unmanaged switches that worked until they did not.
Complexity Does Not Scale
The problem compounds in hybrid environments, which are now the norm. A typical broadcast or corporate studio operation today might be running NDI for remote contributors, Dante for audio, SMPTE ST 2110 for its core video fabric, and AES67 for audio interop—all simultaneously, all across the same physical infrastructure. Each protocol has its own network requirements, and getting them to coexist without colliding requires switch-level configuration that is specific, precise, and easy to get wrong.
The industry largely accepted this as an unavoidable cost of IP adoption, but I think it’s worth challenging. When NDI 4 was transporting NDI traffic, control traffic, and Dante over the same ports without separation, manufacturers were fielding support calls that were about the network surrounding their product.
The fix was not to change the protocol, but to build the separation into the switch itself. Manufacturers who adopted that approach reported reductions in NDI-related support calls of 80-90%.
The Next Generation of Broadcast Infrastructure
The operations successfully making the transition to IP are not necessarily the biggest or best-resourced. What they share is infrastructure specified for the workflow, not generic IT hardware pressed into service.
As production workflows move toward 25 Gigabit and 100 Gigabit transport—already underway at the leading edge of the market—the margin for infrastructure that is merely adequate shrinks considerably. A switch that handles 1 GB NDI traffic may not handle uncompressed SMPTE ST 2110 at the same level of reliability. The physics change, and so do the timing requirements, the synchronization demands, and the consequences of getting it wrong.
The operations thriving in hybrid IP environments are not the ones that chose the right protocol. They are the ones that built the right foundation first.
Broadcast engineers understand this intuitively from the SDI world. The discipline around termination, signal path, and equipment specification that makes SDI reliable translates to an IP workflow. PTP grandmaster configuration is the timing equivalent of gen-lock. VLAN segmentation is the traffic management equivalent of router output assignment. The execution requires infrastructure purpose-built to handle them.
The Partnership Layer
For most of the last decade, the production vendor and the network vendor operated as separate procurement tracks. The integrator specified a production system from one supplier and a network from another, then reconciled the two on-site, often discovering incompatibilities that had not surfaced during pre-sales.
The market is beginning to move away from that model. When production vendors validate and distribute switching infrastructure alongside their own product, the compatibility question is already answered before the project starts. The integrator is deploying a solution that has been tested end-to-end, not reconciling two separately specified systems.
For integrators handling the full scope of a project without a dedicated network engineering resource, removing that compatibility uncertainty changes the risk profile of the project in a way that makes IP adoption viable for clients who would otherwise stay with what they know.
Foundation Before Conversation
Years of SDI-versus-IP debate have generated panel discussions, white papers, and show floor demonstrations. What the industry has scrutinized less is the infrastructure layer underneath those protocols - and whether it is actually ready to deliver on the flexibility and scalability that IP promises.
The operations thriving in hybrid IP environments are not the ones that chose the right protocol. They are the ones that built the right foundation first. The network is not a commodity input in a broadcast IP workflow. It is the layer that determines whether everything above it works—reliably, repeatedly, and at scale. Until the industry starts treating it that way, the gap between the IP conversation and IP reality will remain.
