Art-Net and sACN
When a rig grows beyond what physical DMX runs can comfortably carry, lighting control moves onto the network, tunnelling many DMX512 universes at once over ordinary Ethernet. Two protocols dominate that world, and the difference between them is mostly about how they scale. Art-Net is the long-established one, and traditionally it uses broadcast — sending every universe to every device on the network. That is simple, but it does not scale well, because broadcast traffic reaches everything whether it needs it or not. sACN, properly Streaming ACN and standardised as ANSI E1.31, uses multicast instead, so a node only receives the universes it has actually subscribed to, and it adds genuinely useful features like per-source priority, which lets a backup console take over cleanly. Multicast is the normal mode, not the only one: E1.31 also allows universes to be sent unicast, and a receiver must answer data sent to its unicast address — there is simply no discovery mechanism for those addresses.
For a large system the difference becomes decisive. sACN with multicast, running on a managed switch that is doing IGMP snooping, keeps the traffic targeted and the network calm. Broadcast Art-Net at the same scale can saturate links and bring on lag or outright data loss.
Whichever protocol you use, the universes themselves are planned deliberately — assigned, distributed to the right nodes, and kept on a clean network that is ideally separate from other traffic. The short version is that big rigs lean on sACN, and broadcast-based protocols are better kept to the smaller ones where their simplicity is an asset rather than a liability.
Large rigs use sACN — not broadcast protocols.