It's a tempting idea: my desktop gets capped along with everything else on the user network, so why not give gaming its own VLAN with its own rules and let it breathe? I went down this road and came back. A gaming VLAN doesn't fix the problem people think it fixes, and it charges real rent to not fix it.
The Instinct, and Why It's Wrong#
The assumption underneath "gaming VLAN" is that gaming is hungry for bandwidth and the speed cap is starving it. It isn't. Active gameplay uses very little throughput — single-digit megabits. A download cap of a few hundred megabits isn't even in the same area code as the bottleneck. If frames feel bad, raw bandwidth is almost never the reason.
Latency vs Bandwidth#
What actually wrecks a game session is latency, and specifically latency under load — the jitter that shows up when something else on the line saturates your upstream and packets start queuing. The constraint that matters isn't how many megabits down you're allowed; it's how much upload headroom you have and whether your gateway keeps the queue from ballooning when the pipe fills. That's a completely different lever than a per-VLAN speed limit.
What Smart Queues Actually Do#
The thing that protects game latency is active queue management at the WAN edge — Smart Queues, fq_codel, call it what your gateway calls it. It keeps the buffer from filling up under load so a big upload can't park your game's packets behind it. And it works at the internet edge, regardless of which VLAN the traffic came from. A new segment doesn't add a single drop of that protection; you already have all of it.
The Cost of a New VLAN#
Now the rent. A new VLAN isn't a setting, it's a zone — and a zone has to be reasoned about against every other zone. DNS and NTP redirects, gateway-access rules, resolver reachability, the catch-all blocks at the bottom of each policy, a DHCP scope, a blocking-group assignment. That's a pile of new firewall rows and new edge cases, all to relocate one desktop. On a network already carrying a couple hundred policies, you do not add a zone casually, and certainly not to solve a problem the zone doesn't solve.
Do This Instead#
If a cap is genuinely biting — and the only thing that ever really bites is a big game-store download hitting the ceiling — target the one machine, not the whole segment. A client-specific speed override on the desktop gives it more room without touching anyone else and without a single new firewall rule. Then measure instead of guessing: run a bufferbloat test idle and again while a large download runs. If the grade holds with low added latency, your queues are doing their job and there's nothing to tune. Check the per-client bandwidth history too — if the desktop never actually approaches the cap, the cap was never the problem.
The QoS Footnote#
One trap worth flagging from when I reworked all of this: on a lot of gateways, "Unlimited" is the off switch, not a setting you turn on. For a latency-sensitive client you don't want to assign it a big number — you want to prioritize it and let it run uncapped, while you reserve the actual limits for the noisy stuff like IoT and streaming boxes. Prioritize the things that care about latency; cap the things that only care about throughput. That ordering does more for your games than any amount of new VLANs ever will.